..
  -------------------------------------------------------------------
  NOTE: This file is automatically generated by running clang-tblgen
  -gen-builtin-docs. Do not edit this file by hand!!
  -------------------------------------------------------------------

===============
AMDGPU Builtins
===============

.. contents::
   :local:
   :depth: 2

This document describes the AMDGPU target-specific builtins available in Clang.
Most of these builtins provide direct access to AMDGPU hardware instructions
and intrinsics.

All AMDGPU builtins use the ``__builtin_amdgcn_`` prefix (or ``__builtin_r600_``
for R600 targets). Some arguments must be compile-time constant expressions;
this is noted in the descriptions where applicable.

.. warning::

   These builtins, including their names, arguments, and target requirements,
   are all subject to change without warning across LLVM releases.

.. note::

   This document is a work in progress. Not all builtins are fully documented
   yet. The initial descriptions were generated with AI assistance,
   cross-referencing the following sources:

   - ``clang/include/clang/Basic/BuiltinsAMDGPU.td`` (builtin definitions)
   - ``llvm/include/llvm/IR/IntrinsicsAMDGPU.td`` (intrinsic definitions)
   - ``clang/lib/Sema/SemaAMDGPU.cpp`` (argument validation and constraints)
   - ``clang/lib/CodeGen/TargetBuiltins/AMDGPU.cpp`` (lowering logic)
   - `GPUOpen Machine-Readable ISA <https://gpuopen.com/machine-readable-isa/>`_
     (ISA documents)


ABI / Special Register Builtins
===============================

These builtins provide access to kernel dispatch metadata, work-item and
workgroup identification, and other ABI-level information. They are available
on all AMDGCN targets unless otherwise noted.

``__builtin_amdgcn_cluster_id_x``
---------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_id_x()

**Target Features:** gfx1250-insts

Returns the cluster ID in the X dimension.


``__builtin_amdgcn_cluster_id_y``
---------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_id_y()

**Target Features:** gfx1250-insts

Returns the cluster ID in the Y dimension.


``__builtin_amdgcn_cluster_id_z``
---------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_id_z()

**Target Features:** gfx1250-insts

Returns the cluster ID in the Z dimension.


``__builtin_amdgcn_cluster_workgroup_flat_id``
----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_flat_id()

**Target Features:** gfx1250-insts

Returns the flat (linearized) workgroup ID within the cluster.


``__builtin_amdgcn_cluster_workgroup_id_x``
-------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_id_x()

**Target Features:** gfx1250-insts

Returns the workgroup ID within the cluster in the X dimension.


``__builtin_amdgcn_cluster_workgroup_id_y``
-------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_id_y()

**Target Features:** gfx1250-insts

Returns the workgroup ID within the cluster in the Y dimension.


``__builtin_amdgcn_cluster_workgroup_id_z``
-------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_id_z()

**Target Features:** gfx1250-insts

Returns the workgroup ID within the cluster in the Z dimension.


``__builtin_amdgcn_cluster_workgroup_max_flat_id``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_max_flat_id()

**Target Features:** gfx1250-insts

Returns the maximum flat (linearized) workgroup ID within the cluster.


``__builtin_amdgcn_cluster_workgroup_max_id_x``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_max_id_x()

**Target Features:** gfx1250-insts

Returns the maximum workgroup ID within the cluster in the X dimension.


``__builtin_amdgcn_cluster_workgroup_max_id_y``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_max_id_y()

**Target Features:** gfx1250-insts

Returns the maximum workgroup ID within the cluster in the Y dimension.


``__builtin_amdgcn_cluster_workgroup_max_id_z``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cluster_workgroup_max_id_z()

**Target Features:** gfx1250-insts

Returns the maximum workgroup ID within the cluster in the Z dimension.


``__builtin_amdgcn_dispatch_ptr``
---------------------------------

**Prototype:**

.. code-block:: c

  void address_space<4> * __builtin_amdgcn_dispatch_ptr()

Returns a read-only pointer to the dispatch packet, which contains
workgroup size, grid size, and other dispatch parameters.


``__builtin_amdgcn_grid_size_x``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_grid_size_x()

Returns the total grid size in the X dimension.


``__builtin_amdgcn_grid_size_y``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_grid_size_y()

Returns the total grid size in the Y dimension.


``__builtin_amdgcn_grid_size_z``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_grid_size_z()

Returns the total grid size in the Z dimension.


``__builtin_amdgcn_implicitarg_ptr``
------------------------------------

**Prototype:**

.. code-block:: c

  void address_space<4> * __builtin_amdgcn_implicitarg_ptr()

Returns a pointer to the implicit arguments appended after explicit
kernel arguments. Layout depends on the code object version.


``__builtin_amdgcn_kernarg_segment_ptr``
----------------------------------------

**Prototype:**

.. code-block:: c

  void address_space<4> * __builtin_amdgcn_kernarg_segment_ptr()

Returns a pointer to the beginning of the kernel argument segment.


``__builtin_amdgcn_queue_ptr``
------------------------------

**Prototype:**

.. code-block:: c

  void address_space<4> * __builtin_amdgcn_queue_ptr()

Returns a pointer to the ``queue_t`` object for the queue executing the
current kernel.


``__builtin_amdgcn_workgroup_id_x``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workgroup_id_x()

Returns the workgroup ID in the X dimension.


``__builtin_amdgcn_workgroup_id_y``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workgroup_id_y()

Returns the workgroup ID in the Y dimension.


``__builtin_amdgcn_workgroup_id_z``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workgroup_id_z()

Returns the workgroup ID in the Z dimension.


``__builtin_amdgcn_workgroup_size_x``
-------------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_workgroup_size_x()

Returns the workgroup size in the X dimension.


``__builtin_amdgcn_workgroup_size_y``
-------------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_workgroup_size_y()

Returns the workgroup size in the Y dimension.


``__builtin_amdgcn_workgroup_size_z``
-------------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_workgroup_size_z()

Returns the workgroup size in the Z dimension.


``__builtin_amdgcn_workitem_id_x``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workitem_id_x()

Returns the work-item (thread) ID within the workgroup in the X dimension.


``__builtin_amdgcn_workitem_id_y``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workitem_id_y()

Returns the work-item (thread) ID within the workgroup in the Y dimension.


``__builtin_amdgcn_workitem_id_z``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_workitem_id_z()

Returns the work-item (thread) ID within the workgroup in the Z dimension.



Available/Visible Memory Access Builtins
========================================

These builtins perform 128-bit global or flat memory loads and stores with
available/visible (AV) semantics.

``__builtin_amdgcn_av_load_b128``
---------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_av_load_b128(_ExtVector<4, unsigned int> * ptr, int scope)

**Target Features:** flat-global-insts

Loads 128 bits (4 x i32) from the pointer ``ptr``. The pointer must be in
the global or generic address space. The ``scope`` argument specifies the
synchronization scope using a ``__MEMORY_SCOPE_*`` constant.


``__builtin_amdgcn_av_store_b128``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_av_store_b128(_ExtVector<4, unsigned int> * ptr, _ExtVector<4, unsigned int> data, int scope)

**Target Features:** flat-global-insts

Stores 128 bits (4 x i32) of ``data`` to the pointer ``ptr``. The pointer
must be in the global or generic address space. The ``scope`` argument
specifies the synchronization scope using a ``__MEMORY_SCOPE_*`` constant.



Bit Manipulation Builtins
=========================

These builtins perform bit-manipulation operations within a wavefront.

``__builtin_amdgcn_s_bitreplicate``
-----------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_s_bitreplicate(unsigned int src)

**Target Features:** gfx9-insts

Replicates each bit of the 32-bit ``src`` operand into two adjacent bits of a
64-bit result. Bit ``i`` of the input is copied to bits ``2*i`` and ``2*i+1``
of the output. The operand must be uniform across the wavefront. A divergent
value gives an undefined result.



S-Buffer Load Builtins
======================

These builtins lower to ``llvm.amdgcn.s.buffer.load`` and issue an
``s_buffer_load`` when the byte offset is uniform across the wavefront.
When the offset is divergent, the backend may lower to a ``buffer_load``.

.. warning::

   These builtins should generally be avoided. They must only be used to emit
   ``s_buffer_load`` instructions that access memory not modified during kernel
   execution. ``s_buffer_load`` uses a separate cache that is not coherent with
   the caches used by other load and store operations. The compiler does not
   verify the alignment requirements imposed by ``s_buffer_load``. However,
   because automatic promotion of ``buffer_load`` to ``s_buffer_load`` where
   legal is not yet fully implemented, these builtins are provided so that
   programmers can use such load operations directly.

These builtins take the buffer resource as ``__amdgpu_buffer_rsrc_t``, which
can be created with ``__builtin_amdgcn_make_buffer_rsrc``. The backend
lowers the resource to the 4-dword SGPR descriptor required by
``llvm.amdgcn.s.buffer.load``.

The return type selects the load width. Separate builtins are provided for
each supported scalar and vector element type.

``__builtin_amdgcn_s_buffer_load_f16``
--------------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_s_buffer_load_f16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_f32``
--------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_s_buffer_load_f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_i16``
--------------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_s_buffer_load_i16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_i32``
--------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_s_buffer_load_i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_i8``
-------------------------------------

**Prototype:**

.. code-block:: c

  char __builtin_amdgcn_s_buffer_load_i8(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_u16``
--------------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_s_buffer_load_u16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_u8``
-------------------------------------

**Prototype:**

.. code-block:: c

  unsigned char __builtin_amdgcn_s_buffer_load_u8(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v16f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<16, float> __builtin_amdgcn_s_buffer_load_v16f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v16i32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<16, int> __builtin_amdgcn_s_buffer_load_v16i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v2f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<2, _Float16> __builtin_amdgcn_s_buffer_load_v2f16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v2f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<2, float> __builtin_amdgcn_s_buffer_load_v2f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v2i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<2, int> __builtin_amdgcn_s_buffer_load_v2i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v2i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<2, char> __builtin_amdgcn_s_buffer_load_v2i8(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v3f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<3, _Float16> __builtin_amdgcn_s_buffer_load_v3f16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v3f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<3, float> __builtin_amdgcn_s_buffer_load_v3f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v3i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<3, int> __builtin_amdgcn_s_buffer_load_v3i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v3i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<3, char> __builtin_amdgcn_s_buffer_load_v3i8(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v4f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<4, _Float16> __builtin_amdgcn_s_buffer_load_v4f16(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v4f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<4, float> __builtin_amdgcn_s_buffer_load_v4f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v4i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<4, int> __builtin_amdgcn_s_buffer_load_v4i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v4i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<4, char> __builtin_amdgcn_s_buffer_load_v4i8(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v8f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<8, float> __builtin_amdgcn_s_buffer_load_v8f32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.


``__builtin_amdgcn_s_buffer_load_v8i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _Vector<8, int> __builtin_amdgcn_s_buffer_load_v8i32(__amdgpu_buffer_rsrc_t rsrc, int offset, _Constant int aux)

Loads data from a buffer using an SGPR buffer descriptor.

- ``rsrc``: buffer resource descriptor.
- ``offset``: byte offset from the base of the buffer. May be variable; a
  uniform offset enables ``s_buffer_load`` selection.
- ``aux``: cache-policy and control flags. Must be a compile-time constant.
  The encoding is target-dependent. Common fields include:

  - Pre-GFX12: bit 0 = glc, bit 1 = slc, bit 2 = dlc (GFX10/GFX11), bit 3 =
    swz, bit 4 = scc (GFX90a).
  - GFX942: bit 0 = sc0, bit 1 = nt, bit 3 = swz, bit 4 = sc1.
  - GFX12+: bits [0-2] = th, bits [3-4] = scope, bit 6 = swz.

  The volatile bit is not permitted for this intrinsic.



Tensor DMA Builtins
===================

Asynchronous tensor DMA transfers between global memory and LDS. Track
completion using ``TENSOR_CNT`` or *asyncmarks* (see the ``AMDGPU
Asynchronous Operations`` document).

``D0``..``D4`` are the five chunks of the hardware tensor descriptor. The
``_d2`` machine instruction is selected when ``D2`` and ``D3`` are zero-
initialized; otherwise ``_d4`` is used. ``D4`` is reserved and silently
ignored on gfx1250. ``cpol`` is a compile-time cache-policy bitfield
(``th`` in bits [0:2], scope in bits [3:4]).

``__builtin_amdgcn_tensor_load_to_lds``
---------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_tensor_load_to_lds(_ExtVector<4, unsigned int> D0, _ExtVector<8, int> D1, _ExtVector<4, int> D2, _ExtVector<4, int> D3, _ExtVector<8, int> D4, _Constant int cpol)

**Target Features:** gfx1250-insts

Asynchronously copies a tensor from global memory into LDS.


``__builtin_amdgcn_tensor_store_from_lds``
------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_tensor_store_from_lds(_ExtVector<4, unsigned int> D0, _ExtVector<8, int> D1, _ExtVector<4, int> D2, _ExtVector<4, int> D3, _ExtVector<8, int> D4, _Constant int cpol)

**Target Features:** gfx1250-insts

Asynchronously copies a tensor from LDS into global memory.



WMMA (Wave Matrix Multiply-Accumulate) Builtins
===============================================

WMMA builtins perform wave-cooperative matrix multiply-accumulate of the form
``D = A * B + C``. The work is distributed across all lanes in the wavefront.
The ``_w32`` suffix indicates a wave32 variant, ``_w64`` for wave64.

``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_tied_w32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_tied_w32(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<16, short> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize32

Same as the non-tied f16/bf16 WMMA variant, but the destination register
is tied to the input accumulator: the unselected 16-bit half is preserved
from the input rather than being undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_tied_w64``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_tied_w64(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<8, short> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize64

Same as the non-tied f16/bf16 WMMA variant, but the destination register
is tied to the input accumulator: the unselected 16-bit half is preserved
from the input rather than being undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_w32(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<16, short> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.

The ``opsel`` argument selects which 16-bit half of the accumulator
registers to read from and write to. The content of the other 16-bit
half is undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_w32_gfx12``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_w32_gfx12(_ExtVector<8, short> a, _ExtVector<8, short> b, _ExtVector<8, short> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.
Returns the result matrix.

On GFX12, the output always uses the low 16-bit half of the accumulator
registers (the ``opsel`` bit is implicitly false). In these builtins,
no explicit A/B matrix replication is required in the arguments.


``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_w64(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<8, short> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.

The ``opsel`` argument selects which 16-bit half of the accumulator
registers to read from and write to. The content of the other 16-bit
half is undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_bf16_16x16x16_bf16_w64_gfx12``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, short> __builtin_amdgcn_wmma_bf16_16x16x16_bf16_w64_gfx12(_ExtVector<4, short> a, _ExtVector<4, short> b, _ExtVector<4, short> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.
Returns the result matrix.

On GFX12, the output always uses the low 16-bit half of the accumulator
registers (the ``opsel`` bit is implicitly false). In these builtins,
no explicit A/B matrix replication is required in the arguments.


``__builtin_amdgcn_wmma_bf16_16x16x32_bf16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_wmma_bf16_16x16x32_bf16(_Constant bool a_neg, _ExtVector<16, __bf16> a, _Constant bool b_neg, _ExtVector<16, __bf16> b, _Constant short c_mod, _ExtVector<8, __bf16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with f16 or bf16 inputs and matching
f16 or bf16 output, using a 32-deep K dimension. Multiplies a 16x32
matrix ``A`` by a 32x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_bf16f32_16x16x32_bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_wmma_bf16f32_16x16x32_bf16(_Constant bool a_neg, _ExtVector<16, __bf16> a, _Constant bool b_neg, _ExtVector<16, __bf16> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with bf16 inputs, f32 accumulator, and bf16
output. Uses a 32-deep K dimension. Multiplies a 16x32 matrix ``A`` by a
32x16 matrix ``B``, adds the 16x16 f32 accumulator ``C``, and converts the
result to bf16.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x128_bf8_bf8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x128_bf8_bf8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x128_bf8_fp8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x128_bf8_fp8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x128_fp8_bf8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x128_fp8_bf8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x128_fp8_fp8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x128_fp8_fp8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x16_f16_tied_w32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_tied_w32(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<16, _Float16> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize32

Same as the non-tied f16/bf16 WMMA variant, but the destination register
is tied to the input accumulator: the unselected 16-bit half is preserved
from the input rather than being undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_f16_16x16x16_f16_tied_w64``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_tied_w64(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<8, _Float16> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize64

Same as the non-tied f16/bf16 WMMA variant, but the destination register
is tied to the input accumulator: the unselected 16-bit half is preserved
from the input rather than being undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_f16_16x16x16_f16_w32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_w32(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<16, _Float16> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.

The ``opsel`` argument selects which 16-bit half of the accumulator
registers to read from and write to. The content of the other 16-bit
half is undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_f16_16x16x16_f16_w32_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_w32_gfx12(_ExtVector<8, _Float16> a, _ExtVector<8, _Float16> b, _ExtVector<8, _Float16> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.
Returns the result matrix.

On GFX12, the output always uses the low 16-bit half of the accumulator
registers (the ``opsel`` bit is implicitly false). In these builtins,
no explicit A/B matrix replication is required in the arguments.


``__builtin_amdgcn_wmma_f16_16x16x16_f16_w64``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_w64(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<8, _Float16> c, _Constant bool opsel)

**Target Features:** wmma-256b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.

The ``opsel`` argument selects which 16-bit half of the accumulator
registers to read from and write to. The content of the other 16-bit
half is undefined.

On GFX11, the A and B input matrices require replication (full 256-bit
operands).


``__builtin_amdgcn_wmma_f16_16x16x16_f16_w64_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_wmma_f16_16x16x16_f16_w64_gfx12(_ExtVector<4, _Float16> a, _ExtVector<4, _Float16> b, _ExtVector<4, _Float16> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the
accumulator ``C``, where all operands and the result are f16 or bf16.
Returns the result matrix.

On GFX12, the output always uses the low 16-bit half of the accumulator
registers (the ``opsel`` bit is implicitly false). In these builtins,
no explicit A/B matrix replication is required in the arguments.


``__builtin_amdgcn_wmma_f16_16x16x32_f16``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x32_f16(_Constant bool a_neg, _ExtVector<16, _Float16> a, _Constant bool b_neg, _ExtVector<16, _Float16> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with f16 or bf16 inputs and matching
f16 or bf16 output, using a 32-deep K dimension. Multiplies a 16x32
matrix ``A`` by a 32x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x64_bf8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x64_bf8_bf8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x64_bf8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x64_bf8_fp8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x64_fp8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x64_fp8_bf8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f16_16x16x64_fp8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_wmma_f16_16x16x64_fp8_fp8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, _Float16> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x128_bf8_bf8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x128_bf8_bf8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x128_bf8_fp8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x128_bf8_fp8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x128_f8f6f4``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x128_f8f6f4(_Constant int matrix_a_fmt, _ExtVector<16, int> a, _Constant int matrix_b_fmt, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c)

**Target Features:** gfx1250-insts,wavefrontsize32

Mixed-format wave matrix multiply-accumulate with a 128-deep K dimension.
Multiplies a 16x128 matrix ``A`` by a 128x16 matrix ``B`` and adds the
16x16 f32 accumulator ``C``. The input element types are selected
independently per operand.

- ``matrix_a_fmt`` / ``matrix_b_fmt``: format selectors for A and B
  (FP8, BF8, FP6, BF6, or FP4). Must be compile-time constants.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).


``__builtin_amdgcn_wmma_f32_16x16x128_fp8_bf8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x128_fp8_bf8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x128_fp8_fp8``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x128_fp8_fp8(_ExtVector<16, int> a, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 128-deep K dimension. Multiplies a 16x128 matrix ``A``
by a 128x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x16_bf16_w32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_bf16_w32(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<8, float> c)

**Target Features:** wmma-256b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

On GFX11, the A and B input matrices require replication (full 256-bit
operands): 2 copies for wave32, 4 copies for wave64. The total VGPR
footprint is the same for both wave sizes.


``__builtin_amdgcn_wmma_f32_16x16x16_bf16_w32_gfx12``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_bf16_w32_gfx12(_ExtVector<8, short> a, _ExtVector<8, short> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

Unlike the GFX11 variants, these GFX12 builtins do not require A/B matrix
replication in their input arguments, so A/B vectors are smaller (128-bit
for wave32 and 64-bit for wave64).


``__builtin_amdgcn_wmma_f32_16x16x16_bf16_w64``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_bf16_w64(_ExtVector<16, short> a, _ExtVector<16, short> b, _ExtVector<4, float> c)

**Target Features:** wmma-256b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

On GFX11, the A and B input matrices require replication (full 256-bit
operands): 2 copies for wave32, 4 copies for wave64. The total VGPR
footprint is the same for both wave sizes.


``__builtin_amdgcn_wmma_f32_16x16x16_bf16_w64_gfx12``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_bf16_w64_gfx12(_ExtVector<4, short> a, _ExtVector<4, short> b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

Unlike the GFX11 variants, these GFX12 builtins do not require A/B matrix
replication in their input arguments, so A/B vectors are smaller (128-bit
for wave32 and 64-bit for wave64).


``__builtin_amdgcn_wmma_f32_16x16x16_bf8_bf8_w32_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_bf8_bf8_w32_gfx12(_ExtVector<2, int> a, _ExtVector<2, int> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_bf8_bf8_w64_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_bf8_bf8_w64_gfx12(int a, int b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_bf8_fp8_w32_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_bf8_fp8_w32_gfx12(_ExtVector<2, int> a, _ExtVector<2, int> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_bf8_fp8_w64_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_bf8_fp8_w64_gfx12(int a, int b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_f16_w32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_f16_w32(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<8, float> c)

**Target Features:** wmma-256b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

On GFX11, the A and B input matrices require replication (full 256-bit
operands): 2 copies for wave32, 4 copies for wave64. The total VGPR
footprint is the same for both wave sizes.


``__builtin_amdgcn_wmma_f32_16x16x16_f16_w32_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_f16_w32_gfx12(_ExtVector<8, _Float16> a, _ExtVector<8, _Float16> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

Unlike the GFX11 variants, these GFX12 builtins do not require A/B matrix
replication in their input arguments, so A/B vectors are smaller (128-bit
for wave32 and 64-bit for wave64).


``__builtin_amdgcn_wmma_f32_16x16x16_f16_w64``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_f16_w64(_ExtVector<16, _Float16> a, _ExtVector<16, _Float16> b, _ExtVector<4, float> c)

**Target Features:** wmma-256b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

On GFX11, the A and B input matrices require replication (full 256-bit
operands): 2 copies for wave32, 4 copies for wave64. The total VGPR
footprint is the same for both wave sizes.


``__builtin_amdgcn_wmma_f32_16x16x16_f16_w64_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_f16_w64_gfx12(_ExtVector<4, _Float16> a, _ExtVector<4, _Float16> b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` (both f16 or bf16)
and adds the 16x16 f32 accumulator ``C`` using fused multiply-add. Returns
the f32 result matrix.

Unlike the GFX11 variants, these GFX12 builtins do not require A/B matrix
replication in their input arguments, so A/B vectors are smaller (128-bit
for wave32 and 64-bit for wave64).


``__builtin_amdgcn_wmma_f32_16x16x16_fp8_bf8_w32_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_fp8_bf8_w32_gfx12(_ExtVector<2, int> a, _ExtVector<2, int> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_fp8_bf8_w64_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_fp8_bf8_w64_gfx12(int a, int b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_fp8_fp8_w32_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x16_fp8_fp8_w32_gfx12(_ExtVector<2, int> a, _ExtVector<2, int> b, _ExtVector<8, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize32

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x16_fp8_fp8_w64_gfx12``
--------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_wmma_f32_16x16x16_fp8_fp8_w64_gfx12(int a, int b, _ExtVector<4, float> c)

**Target Features:** wmma-128b-insts,wavefrontsize64

WMMA with 8-bit floating-point inputs (FP8 or BF8) and f32 result.
Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds the f32
accumulator ``C``.

A and B contain packed 8-bit floats (4 values per 32-bit word). Since FP8
and BF8 are not first-class LLVM types, the inputs are passed as integer
vectors.


``__builtin_amdgcn_wmma_f32_16x16x32_bf16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x32_bf16(_Constant bool a_neg, _ExtVector<16, __bf16> a, _Constant bool b_neg, _ExtVector<16, __bf16> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with f16 or bf16 inputs and f32 output,
using a 32-deep K dimension. Multiplies a 16x32 matrix ``A`` by a 32x16
matrix ``B`` and adds the 16x16 f32 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x32_f16``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x32_f16(_Constant bool a_neg, _ExtVector<16, _Float16> a, _Constant bool b_neg, _ExtVector<16, _Float16> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with f16 or bf16 inputs and f32 output,
using a 32-deep K dimension. Multiplies a 16x32 matrix ``A`` by a 32x16
matrix ``B`` and adds the 16x16 f32 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x4_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x4_f32(_Constant bool a_neg, _ExtVector<2, float> a, _Constant bool b_neg, _ExtVector<2, float> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with f32 inputs and f32 output, using a
4-deep K dimension. Multiplies a 16x4 matrix ``A`` by a 4x16 matrix ``B``
and adds the 16x16 f32 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x64_bf8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x64_bf8_bf8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x64_bf8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x64_bf8_fp8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x64_fp8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x64_fp8_bf8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_16x16x64_fp8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_f32_16x16x64_fp8_fp8(_ExtVector<8, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Wave matrix multiply-accumulate with FP8 or BF8 inputs and f32 or f16
output, using a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a
64x16 matrix ``B`` and adds the 16x16 accumulator ``C``.

A and B contain packed 8-bit floats passed as integer vectors.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_f32_32x16x128_f4``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_wmma_f32_32x16x128_f4(_ExtVector<16, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<16, float> c)

**Target Features:** gfx1250-insts,wavefrontsize32

FP4 wave matrix multiply-accumulate with a 128-deep K dimension and a
32x16 output. Multiplies a 32x128 matrix ``A`` by a 128x16 matrix ``B``
and adds the 32x16 f32 accumulator ``C``.

- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).


``__builtin_amdgcn_wmma_f64_16x16x4_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, double> __builtin_amdgcn_wmma_f64_16x16x4_f64(_Constant bool a_neg, _ExtVector<2, double> a, _Constant bool b_neg, _ExtVector<2, double> b, _Constant short c_mod, _ExtVector<8, double> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1251-gemm-insts,wavefrontsize32

Wave matrix multiply-accumulate with f64 inputs and f64 output, using a
4-deep K dimension. Multiplies a 16x4 matrix ``A`` by a 4x16 matrix ``B``
and adds the 16x16 f64 accumulator ``C``.

- ``a_neg`` / ``b_neg``: if true, negate the corresponding input matrix.
- ``c_mod``: accumulator modifier (0 = none, 1 = neg, 2 = abs,
  3 = neg(abs)).
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_i32_16x16x16_iu4_w32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x16_iu4_w32(_Constant bool a_sign, _ExtVector<2, int> a, _Constant bool b_sign, _ExtVector<2, int> b, _ExtVector<8, int> c, _Constant bool clamp)

**Target Features:** wmma-256b-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

On GFX11, the A and B input matrices require replication: 2 copies for
wave32 and 4 copies for wave64. In the builtin API, IU8 A/B operands are
128-bit and IU4 A/B operands are 64-bit.


``__builtin_amdgcn_wmma_i32_16x16x16_iu4_w32_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x16_iu4_w32_gfx12(_Constant bool a_sign, int a, _Constant bool b_sign, int b, _ExtVector<8, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

In these builtins, no explicit A/B matrix replication is required in the
arguments.


``__builtin_amdgcn_wmma_i32_16x16x16_iu4_w64``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_wmma_i32_16x16x16_iu4_w64(_Constant bool a_sign, _ExtVector<2, int> a, _Constant bool b_sign, _ExtVector<2, int> b, _ExtVector<4, int> c, _Constant bool clamp)

**Target Features:** wmma-256b-insts,wavefrontsize64

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

On GFX11, the A and B input matrices require replication: 2 copies for
wave32 and 4 copies for wave64. In the builtin API, IU8 A/B operands are
128-bit and IU4 A/B operands are 64-bit.


``__builtin_amdgcn_wmma_i32_16x16x16_iu4_w64_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_wmma_i32_16x16x16_iu4_w64_gfx12(_Constant bool a_sign, int a, _Constant bool b_sign, int b, _ExtVector<4, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize64

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

In these builtins, no explicit A/B matrix replication is required in the
arguments.


``__builtin_amdgcn_wmma_i32_16x16x16_iu8_w32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x16_iu8_w32(_Constant bool a_sign, _ExtVector<4, int> a, _Constant bool b_sign, _ExtVector<4, int> b, _ExtVector<8, int> c, _Constant bool clamp)

**Target Features:** wmma-256b-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

On GFX11, the A and B input matrices require replication: 2 copies for
wave32 and 4 copies for wave64. In the builtin API, IU8 A/B operands are
128-bit and IU4 A/B operands are 64-bit.


``__builtin_amdgcn_wmma_i32_16x16x16_iu8_w32_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x16_iu8_w32_gfx12(_Constant bool a_sign, _ExtVector<2, int> a, _Constant bool b_sign, _ExtVector<2, int> b, _ExtVector<8, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

In these builtins, no explicit A/B matrix replication is required in the
arguments.


``__builtin_amdgcn_wmma_i32_16x16x16_iu8_w64``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_wmma_i32_16x16x16_iu8_w64(_Constant bool a_sign, _ExtVector<4, int> a, _Constant bool b_sign, _ExtVector<4, int> b, _ExtVector<4, int> c, _Constant bool clamp)

**Target Features:** wmma-256b-insts,wavefrontsize64

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

On GFX11, the A and B input matrices require replication: 2 copies for
wave32 and 4 copies for wave64. In the builtin API, IU8 A/B operands are
128-bit and IU4 A/B operands are 64-bit.


``__builtin_amdgcn_wmma_i32_16x16x16_iu8_w64_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_wmma_i32_16x16x16_iu8_w64_gfx12(_Constant bool a_sign, int a, _Constant bool b_sign, int b, _ExtVector<4, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize64

Integer wave matrix multiply-accumulate on packed 8-bit or 4-bit integer
inputs. Multiplies a 16x16 matrix ``A`` by a 16x16 matrix ``B`` and adds
the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.

In these builtins, no explicit A/B matrix replication is required in the
arguments.


``__builtin_amdgcn_wmma_i32_16x16x32_iu4_w32_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x32_iu4_w32_gfx12(_Constant bool a_sign, _ExtVector<2, int> a, _Constant bool b_sign, _ExtVector<2, int> b, _ExtVector<8, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 4-bit integer inputs with
a 32-deep K dimension (16x16x32). Multiplies a 16x32 matrix ``A`` by a
32x16 matrix ``B`` and adds the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.


``__builtin_amdgcn_wmma_i32_16x16x32_iu4_w64_gfx12``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_wmma_i32_16x16x32_iu4_w64_gfx12(_Constant bool a_sign, int a, _Constant bool b_sign, int b, _ExtVector<4, int> c, _Constant bool clamp)

**Target Features:** wmma-128b-insts,wavefrontsize64

Integer wave matrix multiply-accumulate on packed 4-bit integer inputs with
a 32-deep K dimension (16x16x32). Multiplies a 16x32 matrix ``A`` by a
32x16 matrix ``B`` and adds the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``clamp``: if true, the result saturates instead of wrapping on overflow.


``__builtin_amdgcn_wmma_i32_16x16x64_iu8``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_wmma_i32_16x16x64_iu8(_Constant bool a_sign, _ExtVector<8, int> a, _Constant bool b_sign, _ExtVector<8, int> b, _ExtVector<8, int> c, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse, ...)

**Target Features:** gfx1250-insts,wavefrontsize32

Integer wave matrix multiply-accumulate on packed 8-bit integer inputs
with a 64-deep K dimension. Multiplies a 16x64 matrix ``A`` by a 64x16
matrix ``B`` and adds the i32 accumulator ``C``.

- ``a_sign`` / ``b_sign``: if true, the corresponding operand is treated
  as signed; if false, as unsigned.
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.
- ``clamp`` (optional): an optional 8th argument. When present, it must be
  a compile-time constant integer expression convertible to bool. If true,
  the result saturates instead of wrapping on overflow. If omitted, defaults
  to false.


``__builtin_amdgcn_wmma_scale16_f32_16x16x128_f8f6f4``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_scale16_f32_16x16x128_f8f6f4(_Constant int matrix_a_fmt, _ExtVector<16, int> a, _Constant int matrix_b_fmt, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant int matrix_a_scale, _Constant int matrix_a_scale_fmt, long int matrix_a_scale_exp, _Constant int matrix_b_scale, _Constant int matrix_b_scale_fmt, long int matrix_b_scale_exp, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Scaled wave matrix multiply-accumulate with 64-bit (16-bit per element)
scale exponents, as opposed to the 32-bit exponents used by the non-16
scale variant.

- For ``..._f8f6f4`` variants, ``matrix_a_fmt`` / ``matrix_b_fmt`` are format
  selectors for A and B. They must be compile-time constants.
- ``c_mod``: accumulator modifier.
- ``matrix_a_scale`` / ``matrix_b_scale``: scale factor selectors.
  Must be compile-time constants.
- ``matrix_a_scale_fmt`` / ``matrix_b_scale_fmt``: scale format selectors.
  Must be compile-time constants.
- ``matrix_a_scale_exp`` / ``matrix_b_scale_exp``: 64-bit scale exponents.
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_scale16_f32_32x16x128_f4``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_wmma_scale16_f32_32x16x128_f4(_ExtVector<16, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<16, float> c, _Constant int matrix_a_scale, _Constant int matrix_a_scale_fmt, long int matrix_a_scale_exp, _Constant int matrix_b_scale, _Constant int matrix_b_scale_fmt, long int matrix_b_scale_exp, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Scaled wave matrix multiply-accumulate with 64-bit (16-bit per element)
scale exponents, as opposed to the 32-bit exponents used by the non-16
scale variant.

- For ``..._f8f6f4`` variants, ``matrix_a_fmt`` / ``matrix_b_fmt`` are format
  selectors for A and B. They must be compile-time constants.
- ``c_mod``: accumulator modifier.
- ``matrix_a_scale`` / ``matrix_b_scale``: scale factor selectors.
  Must be compile-time constants.
- ``matrix_a_scale_fmt`` / ``matrix_b_scale_fmt``: scale format selectors.
  Must be compile-time constants.
- ``matrix_a_scale_exp`` / ``matrix_b_scale_exp``: 64-bit scale exponents.
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_scale_f32_16x16x128_f8f6f4``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_wmma_scale_f32_16x16x128_f8f6f4(_Constant int matrix_a_fmt, _ExtVector<16, int> a, _Constant int matrix_b_fmt, _ExtVector<16, int> b, _Constant short c_mod, _ExtVector<8, float> c, _Constant int matrix_a_scale, _Constant int matrix_a_scale_fmt, int matrix_a_scale_exp, _Constant int matrix_b_scale, _Constant int matrix_b_scale_fmt, int matrix_b_scale_exp, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Scaled wave matrix multiply-accumulate. Extends the ``f8f6f4`` or ``f4``
WMMA with per-operand scale factors applied during the computation.

- For ``..._f8f6f4`` variants, ``matrix_a_fmt`` / ``matrix_b_fmt`` are format
  selectors for A and B. They must be compile-time constants.
- ``c_mod``: accumulator modifier.
- ``matrix_a_scale`` / ``matrix_b_scale``: scale factor selectors.
  Must be compile-time constants.
- ``matrix_a_scale_fmt`` / ``matrix_b_scale_fmt``: scale format selectors.
  Must be compile-time constants.
- ``matrix_a_scale_exp`` / ``matrix_b_scale_exp``: 32-bit scale exponents.
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.


``__builtin_amdgcn_wmma_scale_f32_32x16x128_f4``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_wmma_scale_f32_32x16x128_f4(_ExtVector<16, int> a, _ExtVector<8, int> b, _Constant short c_mod, _ExtVector<16, float> c, _Constant int matrix_a_scale, _Constant int matrix_a_scale_fmt, int matrix_a_scale_exp, _Constant int matrix_b_scale, _Constant int matrix_b_scale_fmt, int matrix_b_scale_exp, _Constant bool matrix_a_reuse, _Constant bool matrix_b_reuse)

**Target Features:** gfx1250-insts,wavefrontsize32

Scaled wave matrix multiply-accumulate. Extends the ``f8f6f4`` or ``f4``
WMMA with per-operand scale factors applied during the computation.

- For ``..._f8f6f4`` variants, ``matrix_a_fmt`` / ``matrix_b_fmt`` are format
  selectors for A and B. They must be compile-time constants.
- ``c_mod``: accumulator modifier.
- ``matrix_a_scale`` / ``matrix_b_scale``: scale factor selectors.
  Must be compile-time constants.
- ``matrix_a_scale_fmt`` / ``matrix_b_scale_fmt``: scale format selectors.
  Must be compile-time constants.
- ``matrix_a_scale_exp`` / ``matrix_b_scale_exp``: 32-bit scale exponents.
- ``matrix_a_reuse`` / ``matrix_b_reuse``: hints to the hardware that
  matrix A or B data can be reused from a previous WMMA instruction.



Wave Data Exchange Builtins
===========================

These builtins provide cross-lane data exchange within a wavefront.

``__builtin_amdgcn_wave_shuffle``
---------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_wave_shuffle(int src, int idx)

Returns the value of ``src`` held by the lane identified by ``idx``. This is
a high-level wave shuffle operation that is correctly lowered for both wave32
and wave64 without requiring the caller to compute byte offsets.

If the index is outside the valid range ``[0, wavefront_size)``, it wraps
around modulo the wavefront size. Reading from an inactive lane returns a
nondeterministic value.



Undocumented
============

This section lists builtins which are recognized by Clang, but which are
currently missing documentation.

``__builtin_amdgcn_add_max_i32``
--------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_add_max_i32(int, int, int, _Constant bool)

**Target Features:** add-min-max-insts

No documentation.


``__builtin_amdgcn_add_max_u32``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_add_max_u32(unsigned int, unsigned int, unsigned int, _Constant bool)

**Target Features:** add-min-max-insts

No documentation.


``__builtin_amdgcn_add_min_i32``
--------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_add_min_i32(int, int, int, _Constant bool)

**Target Features:** add-min-max-insts

No documentation.


``__builtin_amdgcn_add_min_u32``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_add_min_u32(unsigned int, unsigned int, unsigned int, _Constant bool)

**Target Features:** add-min-max-insts

No documentation.


``__builtin_amdgcn_alignbit``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_alignbit(unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_alignbyte``
------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_alignbyte(unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_ashr_pk_i8_i32``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_ashr_pk_i8_i32(unsigned int, unsigned int, unsigned int)

**Target Features:** ashr-pk-insts

No documentation.


``__builtin_amdgcn_ashr_pk_u8_i32``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_ashr_pk_u8_i32(unsigned int, unsigned int, unsigned int)

**Target Features:** ashr-pk-insts

No documentation.


``__builtin_amdgcn_asyncmark``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_asyncmark()

**Target Features:** vmem-to-lds-load-insts|asynccnt

No documentation.


``__builtin_amdgcn_atomic_dec32``
---------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_atomic_dec32(uint32_t volatile *, uint32_t, unsigned int, char const *)

No documentation.


``__builtin_amdgcn_atomic_dec64``
---------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_atomic_dec64(uint64_t volatile *, uint64_t, unsigned int, char const *)

No documentation.


``__builtin_amdgcn_atomic_inc32``
---------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_atomic_inc32(uint32_t volatile *, uint32_t, unsigned int, char const *)

No documentation.


``__builtin_amdgcn_atomic_inc64``
---------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_atomic_inc64(uint64_t volatile *, uint64_t, unsigned int, char const *)

No documentation.


``__builtin_amdgcn_ballot_w32``
-------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_ballot_w32(bool)

**Target Features:** wavefrontsize32

No documentation.


``__builtin_amdgcn_ballot_w64``
-------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_ballot_w64(bool)

No documentation.


``__builtin_amdgcn_bitop3_b16``
-------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_bitop3_b16(short, short, short, _Constant unsigned int)

**Target Features:** bitop3-insts

No documentation.


``__builtin_amdgcn_bitop3_b32``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_bitop3_b32(int, int, int, _Constant unsigned int)

**Target Features:** bitop3-insts

No documentation.


``__builtin_amdgcn_buffer_wbinvl1``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_buffer_wbinvl1()

No documentation.


``__builtin_amdgcn_buffer_wbinvl1_vol``
---------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_buffer_wbinvl1_vol()

**Target Features:** ci-insts

No documentation.


``__builtin_amdgcn_class``
--------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_class(double, int)

No documentation.


``__builtin_amdgcn_classf``
---------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_classf(float, int)

No documentation.


``__builtin_amdgcn_classh``
---------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_classh(_Float16, int)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_cluster_load_async_to_lds_b128``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cluster_load_async_to_lds_b128(_ExtVector<4, int> address_space<1> *, _ExtVector<4, int> address_space<3> *, _Constant int, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_async_to_lds_b32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cluster_load_async_to_lds_b32(int address_space<1> *, int address_space<3> *, _Constant int, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_async_to_lds_b64``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cluster_load_async_to_lds_b64(_ExtVector<2, int> address_space<1> *, _ExtVector<2, int> address_space<3> *, _Constant int, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_async_to_lds_b8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cluster_load_async_to_lds_b8(char address_space<1> *, char address_space<3> *, _Constant int, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_b128``
--------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_cluster_load_b128(_ExtVector<4, int> address_space<1> *, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_b32``
-------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cluster_load_b32(int address_space<1> *, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cluster_load_b64``
-------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_cluster_load_b64(_ExtVector<2, int> address_space<1> *, _Constant int, int)

**Target Features:** mcast-load-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_load_16x8B``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_cooperative_atomic_load_16x8B(_ExtVector<2, int> *, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_load_32x4B``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cooperative_atomic_load_32x4B(int *, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_load_8x16B``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_cooperative_atomic_load_8x16B(_ExtVector<4, int> *, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_store_16x8B``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cooperative_atomic_store_16x8B(_ExtVector<2, int> *, _ExtVector<2, int>, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_store_32x4B``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cooperative_atomic_store_32x4B(int *, int, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cooperative_atomic_store_8x16B``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_cooperative_atomic_store_8x16B(_ExtVector<4, int> *, _ExtVector<4, int>, _Constant int, char const *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_cos_bf16``
-----------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_cos_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_cosf``
-------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cosf(float)

No documentation.


``__builtin_amdgcn_cosh``
-------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_cosh(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_cubeid``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cubeid(float, float, float)

**Target Features:** cube-insts

No documentation.


``__builtin_amdgcn_cubema``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cubema(float, float, float)

**Target Features:** cube-insts

No documentation.


``__builtin_amdgcn_cubesc``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cubesc(float, float, float)

**Target Features:** cube-insts

No documentation.


``__builtin_amdgcn_cubetc``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cubetc(float, float, float)

**Target Features:** cube-insts

No documentation.


``__builtin_amdgcn_cvt_f16_bf8``
--------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_cvt_f16_bf8(int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_f16_fp8``
--------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_cvt_f16_fp8(int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_f32_bf8``
--------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_f32_bf8(int, _Constant int)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_f32_fp8``
--------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_f32_fp8(int, _Constant int)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_f32_fp8_e5m3``
-------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_f32_fp8_e5m3(int, _Constant int)

**Target Features:** fp8e5m3-insts

No documentation.


``__builtin_amdgcn_cvt_off_f32_i4``
-----------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_off_f32_i4(int)

No documentation.


``__builtin_amdgcn_cvt_pk_bf8_f16``
-----------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_cvt_pk_bf8_f16(_ExtVector<2, _Float16>)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_pk_bf8_f32``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_pk_bf8_f32(float, float, int, _Constant bool)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_pk_f16_bf8``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_pk_f16_bf8(short)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_pk_f16_fp8``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_pk_f16_fp8(short)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_pk_f32_bf8``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, float> __builtin_amdgcn_cvt_pk_f32_bf8(int, _Constant bool)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_pk_f32_fp8``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, float> __builtin_amdgcn_cvt_pk_f32_fp8(int, _Constant bool)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_pk_fp8_f16``
-----------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_cvt_pk_fp8_f16(_ExtVector<2, _Float16>)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_pk_fp8_f32``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_pk_fp8_f32(float, float, int, _Constant bool)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_pk_fp8_f32_e5m3``
----------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_pk_fp8_f32_e5m3(float, float, int, _Constant bool)

**Target Features:** fp8e5m3-insts

No documentation.


``__builtin_amdgcn_cvt_pk_i16``
-------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_pk_i16(int, int)

No documentation.


``__builtin_amdgcn_cvt_pk_u16``
-------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned short> __builtin_amdgcn_cvt_pk_u16(unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_cvt_pk_u8_f32``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_pk_u8_f32(float, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_cvt_pknorm_i16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_pknorm_i16(float, float)

**Target Features:** cvt-pknorm-vop2-insts

No documentation.


``__builtin_amdgcn_cvt_pknorm_u16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned short> __builtin_amdgcn_cvt_pknorm_u16(float, float)

**Target Features:** cvt-pknorm-vop2-insts

No documentation.


``__builtin_amdgcn_cvt_pkrtz``
------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __fp16> __builtin_amdgcn_cvt_pkrtz(float, float)

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_bf16_bf6``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, __bf16> __builtin_amdgcn_cvt_scale_pk16_bf16_bf6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_bf16_fp6``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, __bf16> __builtin_amdgcn_cvt_scale_pk16_bf16_fp6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_f16_bf6``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, _Float16> __builtin_amdgcn_cvt_scale_pk16_f16_bf6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_f16_fp6``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, _Float16> __builtin_amdgcn_cvt_scale_pk16_f16_fp6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_f32_bf6``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_cvt_scale_pk16_f32_bf6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk16_f32_fp6``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_cvt_scale_pk16_f32_fp6(_ExtVector<3, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_bf16_bf8``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_cvt_scale_pk8_bf16_bf8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_bf16_fp4``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_cvt_scale_pk8_bf16_fp4(unsigned int, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_bf16_fp8``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_cvt_scale_pk8_bf16_fp8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f16_bf8``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_cvt_scale_pk8_f16_bf8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f16_fp4``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_cvt_scale_pk8_f16_fp4(unsigned int, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f16_fp8``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_cvt_scale_pk8_f16_fp8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f32_bf8``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_cvt_scale_pk8_f32_bf8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f32_fp4``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_cvt_scale_pk8_f32_fp4(unsigned int, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scale_pk8_f32_fp8``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_cvt_scale_pk8_f32_fp8(_ExtVector<2, unsigned int>, unsigned int, _Constant unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_2xpk16_bf6_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_2xpk16_bf6_f32(_ExtVector<16, float>, _ExtVector<16, float>, float)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_2xpk16_fp6_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_2xpk16_fp6_f32(_ExtVector<16, float>, _ExtVector<16, float>, float)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_f16_bf8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_scalef32_f16_bf8(_ExtVector<2, _Float16>, int, float, _Constant int, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_f16_fp8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_scalef32_f16_fp8(_ExtVector<2, _Float16>, int, float, _Constant int, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_f32_bf8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_scalef32_f32_bf8(int, float, _Constant int)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_f32_fp8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_cvt_scalef32_f32_fp8(int, float, _Constant int)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_bf6_bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_bf6_bf16(_ExtVector<16, __bf16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_bf6_f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_bf6_f16(_ExtVector<16, _Float16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_bf6_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_bf6_f32(_ExtVector<16, float>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_fp6_bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_fp6_bf16(_ExtVector<16, __bf16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_fp6_f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_fp6_f16(_ExtVector<16, _Float16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk16_fp6_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_pk16_fp6_f32(_ExtVector<16, float>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_bf16_bf6``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, __bf16> __builtin_amdgcn_cvt_scalef32_pk32_bf16_bf6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_bf16_fp6``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, __bf16> __builtin_amdgcn_cvt_scalef32_pk32_bf16_fp6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_bf6_bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_pk32_bf6_bf16(_ExtVector<32, __bf16>, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_bf6_f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_pk32_bf6_f16(_ExtVector<32, _Float16>, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_f16_bf6``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, _Float16> __builtin_amdgcn_cvt_scalef32_pk32_f16_bf6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_f16_fp6``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, _Float16> __builtin_amdgcn_cvt_scalef32_pk32_f16_fp6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_f32_bf6``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_cvt_scalef32_pk32_f32_bf6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_f32_fp6``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_cvt_scalef32_pk32_f32_fp6(_ExtVector<6, unsigned int>, float)

**Target Features:** fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_fp6_bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_pk32_fp6_bf16(_ExtVector<32, __bf16>, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk32_fp6_f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_pk32_fp6_f16(_ExtVector<32, _Float16>, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_bf8_bf16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_bf8_bf16(_ExtVector<8, __bf16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_bf8_f16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_bf8_f16(_ExtVector<8, _Float16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_bf8_f32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_bf8_f32(_ExtVector<8, float>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp4_bf16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk8_fp4_bf16(_ExtVector<8, __bf16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp4_f16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk8_fp4_f16(_ExtVector<8, _Float16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp4_f32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk8_fp4_f32(_ExtVector<8, float>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp8_bf16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_fp8_bf16(_ExtVector<8, __bf16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp8_f16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_fp8_f16(_ExtVector<8, _Float16>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk8_fp8_f32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_pk8_fp8_f32(_ExtVector<8, float>, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf16_bf8``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __bf16> __builtin_amdgcn_cvt_scalef32_pk_bf16_bf8(unsigned int, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf16_fp4``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __bf16> __builtin_amdgcn_cvt_scalef32_pk_bf16_fp4(unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf16_fp8``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __bf16> __builtin_amdgcn_cvt_scalef32_pk_bf16_fp8(unsigned int, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf8_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_bf8_bf16(_ExtVector<2, short>, _ExtVector<2, __bf16>, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf8_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_bf8_f16(_ExtVector<2, short>, _ExtVector<2, _Float16>, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_bf8_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_bf8_f32(_ExtVector<2, short>, float, float, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f16_bf8``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_scalef32_pk_f16_bf8(unsigned int, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f16_fp4``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_scalef32_pk_f16_fp4(unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f16_fp8``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_scalef32_pk_f16_fp8(unsigned int, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f32_bf8``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, float> __builtin_amdgcn_cvt_scalef32_pk_f32_bf8(unsigned int, float, _Constant bool)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f32_fp4``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, float> __builtin_amdgcn_cvt_scalef32_pk_f32_fp4(unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_f32_fp8``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, float> __builtin_amdgcn_cvt_scalef32_pk_f32_fp8(unsigned int, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp4_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk_fp4_bf16(unsigned int, _ExtVector<2, __bf16>, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp4_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk_fp4_f16(unsigned int, _ExtVector<2, _Float16>, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp4_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_pk_fp4_f32(unsigned int, float, float, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp8_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_fp8_bf16(_ExtVector<2, short>, _ExtVector<2, __bf16>, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp8_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_fp8_f16(_ExtVector<2, short>, _ExtVector<2, _Float16>, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_pk_fp8_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_cvt_scalef32_pk_fp8_f32(_ExtVector<2, short>, float, float, float, _Constant bool)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_bf8_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_bf8_bf16(int, __bf16, unsigned int, float, _Constant int)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_bf8_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_bf8_f16(int, _Float16, unsigned int, float, _Constant int)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_bf8_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_bf8_f32(int, float, unsigned int, float, _Constant int)

**Target Features:** bf8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_fp8_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_fp8_bf16(int, __bf16, unsigned int, float, _Constant int)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_fp8_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_fp8_f16(int, _Float16, unsigned int, float, _Constant int)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_fp8_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_scalef32_sr_fp8_f32(int, float, unsigned int, float, _Constant int)

**Target Features:** fp8-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_bf16``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_bf16(_ExtVector<16, __bf16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_f16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_f16(_ExtVector<16, _Float16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_bf6_f32(_ExtVector<16, float>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_bf16``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_bf16(_ExtVector<16, __bf16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_f16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_f16(_ExtVector<16, _Float16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk16_fp6_f32(_ExtVector<16, float>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_bf16``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_bf16(_ExtVector<32, __bf16>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_f16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_f16(_ExtVector<32, _Float16>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_bf6_f32(_ExtVector<32, float>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_bf16``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_bf16(_ExtVector<32, __bf16>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_f16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_f16(_ExtVector<32, _Float16>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<6, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk32_fp6_f32(_ExtVector<32, float>, unsigned int, float)

**Target Features:** f16bf16-to-fp6bf6-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_bf16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_bf16(_ExtVector<8, __bf16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_f16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_f16(_ExtVector<8, _Float16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_bf8_f32(_ExtVector<8, float>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_bf16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_bf16(_ExtVector<8, __bf16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_f16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_f16(_ExtVector<8, _Float16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk8_fp4_f32(_ExtVector<8, float>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_bf16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_bf16(_ExtVector<8, __bf16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_f16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_f16(_ExtVector<8, _Float16>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_cvt_scalef32_sr_pk8_fp8_f32(_ExtVector<8, float>, unsigned int, float)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk_fp4_bf16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk_fp4_bf16(unsigned int, _ExtVector<2, __bf16>, unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk_fp4_f16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk_fp4_f16(unsigned int, _ExtVector<2, _Float16>, unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_scalef32_sr_pk_fp4_f32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_cvt_scalef32_sr_pk_fp4_f32(unsigned int, _ExtVector<2, float>, unsigned int, float, _Constant int)

**Target Features:** fp4-cvt-scale-insts

No documentation.


``__builtin_amdgcn_cvt_sr_bf16_f32``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __bf16> __builtin_amdgcn_cvt_sr_bf16_f32(_ExtVector<2, __bf16>, float, unsigned int, _Constant bool)

**Target Features:** f32-to-f16bf16-cvt-sr-insts

No documentation.


``__builtin_amdgcn_cvt_sr_bf8_f16``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_sr_bf8_f16(_Float16, int, unsigned int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_sr_bf8_f32``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_sr_bf8_f32(float, int, int, _Constant int)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_sr_f16_f32``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_sr_f16_f32(_ExtVector<2, _Float16>, float, unsigned int, _Constant bool)

**Target Features:** f32-to-f16bf16-cvt-sr-insts

No documentation.


``__builtin_amdgcn_cvt_sr_fp8_f16``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_sr_fp8_f16(_Float16, int, unsigned int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_cvt_sr_fp8_f32``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_sr_fp8_f32(float, int, int, _Constant int)

**Target Features:** fp8-conversion-insts

No documentation.


``__builtin_amdgcn_cvt_sr_fp8_f32_e5m3``
----------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_cvt_sr_fp8_f32_e5m3(float, int, int, _Constant int)

**Target Features:** fp8e5m3-insts

No documentation.


``__builtin_amdgcn_cvt_sr_pk_bf16_f32``
---------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, __bf16> __builtin_amdgcn_cvt_sr_pk_bf16_f32(float, float, int)

**Target Features:** bf16-cvt-insts

No documentation.


``__builtin_amdgcn_cvt_sr_pk_f16_f32``
--------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_cvt_sr_pk_f16_f32(float, float, int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_div_fixup``
------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_div_fixup(double, double, double)

No documentation.


``__builtin_amdgcn_div_fixupf``
-------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_div_fixupf(float, float, float)

No documentation.


``__builtin_amdgcn_div_fixuph``
-------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_div_fixuph(_Float16, _Float16, _Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_div_fmas``
-----------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_div_fmas(double, double, double, bool)

No documentation.


``__builtin_amdgcn_div_fmasf``
------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_div_fmasf(float, float, float, bool)

No documentation.


``__builtin_amdgcn_div_scale``
------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_div_scale(double, double, bool, bool *)

No documentation.


``__builtin_amdgcn_div_scalef``
-------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_div_scalef(float, float, bool, bool *)

No documentation.


``__builtin_amdgcn_dot4_f32_bf8_bf8``
-------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_dot4_f32_bf8_bf8(unsigned int, unsigned int, float)

**Target Features:** dot11-insts

No documentation.


``__builtin_amdgcn_dot4_f32_bf8_fp8``
-------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_dot4_f32_bf8_fp8(unsigned int, unsigned int, float)

**Target Features:** dot11-insts

No documentation.


``__builtin_amdgcn_dot4_f32_fp8_bf8``
-------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_dot4_f32_fp8_bf8(unsigned int, unsigned int, float)

**Target Features:** dot11-insts

No documentation.


``__builtin_amdgcn_dot4_f32_fp8_fp8``
-------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_dot4_f32_fp8_fp8(unsigned int, unsigned int, float)

**Target Features:** dot11-insts

No documentation.


``__builtin_amdgcn_ds_append``
------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_append(int address_space<3> *)

No documentation.


``__builtin_amdgcn_ds_atomic_async_barrier_arrive_b64``
-------------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_atomic_async_barrier_arrive_b64(long int address_space<3> *)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_ds_atomic_barrier_arrive_rtn_b64``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  long int __builtin_amdgcn_ds_atomic_barrier_arrive_rtn_b64(long int address_space<3> *, long int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_ds_atomic_fadd_f32``
---------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_ds_atomic_fadd_f32(float address_space<3> *, float)

**Target Features:** gfx8-insts

No documentation.


``__builtin_amdgcn_ds_atomic_fadd_f64``
---------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_ds_atomic_fadd_f64(double address_space<3> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_ds_atomic_fadd_v2bf16``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_ds_atomic_fadd_v2bf16(_ExtVector<2, short> address_space<3> *, _ExtVector<2, short>)

**Target Features:** atomic-ds-pk-add-16-insts

No documentation.


``__builtin_amdgcn_ds_atomic_fadd_v2f16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_ds_atomic_fadd_v2f16(_ExtVector<2, _Float16> address_space<3> *, _ExtVector<2, _Float16>)

**Target Features:** atomic-ds-pk-add-16-insts

No documentation.


``__builtin_amdgcn_ds_bpermute``
--------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_bpermute(int, int)

No documentation.


``__builtin_amdgcn_ds_bpermute_fi_b32``
---------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_bpermute_fi_b32(int, int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_ds_bvh_stack_push4_pop1_rtn``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_ds_bvh_stack_push4_pop1_rtn(unsigned int, unsigned int, _ExtVector<4, unsigned int>, _Constant int)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_ds_bvh_stack_push8_pop1_rtn``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_ds_bvh_stack_push8_pop1_rtn(unsigned int, unsigned int, _ExtVector<8, unsigned int>, _Constant int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_ds_bvh_stack_push8_pop2_rtn``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, uint64_t> __builtin_amdgcn_ds_bvh_stack_push8_pop2_rtn(unsigned int, unsigned int, _ExtVector<8, unsigned int>, _Constant int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_ds_bvh_stack_rtn``
-------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_ds_bvh_stack_rtn(unsigned int, unsigned int, _ExtVector<4, unsigned int>, _Constant int)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_ds_consume``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_consume(int address_space<3> *)

No documentation.


``__builtin_amdgcn_ds_faddf``
-----------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_ds_faddf(float address_space<3> *, float, _Constant int, _Constant int, _Constant bool)

No documentation.


``__builtin_amdgcn_ds_fmaxf``
-----------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_ds_fmaxf(float address_space<3> *, float, _Constant int, _Constant int, _Constant bool)

No documentation.


``__builtin_amdgcn_ds_fminf``
-----------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_ds_fminf(float address_space<3> *, float, _Constant int, _Constant int, _Constant bool)

No documentation.


``__builtin_amdgcn_ds_gws_barrier``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_barrier(unsigned int, unsigned int)

**Target Features:** gws

No documentation.


``__builtin_amdgcn_ds_gws_init``
--------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_init(unsigned int, unsigned int)

**Target Features:** gws

No documentation.


``__builtin_amdgcn_ds_gws_sema_br``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_sema_br(unsigned int, unsigned int)

**Target Features:** gws

No documentation.


``__builtin_amdgcn_ds_gws_sema_p``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_sema_p(unsigned int)

**Target Features:** gws

No documentation.


``__builtin_amdgcn_ds_gws_sema_release_all``
--------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_sema_release_all(unsigned int)

**Target Features:** ci-insts

No documentation.


``__builtin_amdgcn_ds_gws_sema_v``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_ds_gws_sema_v(unsigned int)

**Target Features:** gws

No documentation.


``__builtin_amdgcn_ds_load_tr16_b128_v8bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_ds_load_tr16_b128_v8bf16(_ExtVector<8, __bf16> address_space<3> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_load_tr16_b128_v8f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __fp16> __builtin_amdgcn_ds_load_tr16_b128_v8f16(_ExtVector<8, __fp16> address_space<3> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_load_tr16_b128_v8i16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_ds_load_tr16_b128_v8i16(_ExtVector<8, short> address_space<3> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_load_tr4_b64_v2i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_ds_load_tr4_b64_v2i32(_ExtVector<2, int> address_space<3> *)

**Target Features:** transpose-load-f4f6-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_load_tr6_b96_v3i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, int> __builtin_amdgcn_ds_load_tr6_b96_v3i32(_ExtVector<3, int> address_space<3> *)

**Target Features:** transpose-load-f4f6-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_load_tr8_b64_v2i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_ds_load_tr8_b64_v2i32(_ExtVector<2, int> address_space<3> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_ds_permute``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_permute(int, int)

No documentation.


``__builtin_amdgcn_ds_read_tr16_b64_v4bf16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, __bf16> __builtin_amdgcn_ds_read_tr16_b64_v4bf16(_ExtVector<4, __bf16> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_read_tr16_b64_v4f16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, __fp16> __builtin_amdgcn_ds_read_tr16_b64_v4f16(_ExtVector<4, __fp16> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_read_tr16_b64_v4i16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, short> __builtin_amdgcn_ds_read_tr16_b64_v4i16(_ExtVector<4, short> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_read_tr4_b64_v2i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_ds_read_tr4_b64_v2i32(_ExtVector<2, int> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_read_tr6_b96_v3i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, int> __builtin_amdgcn_ds_read_tr6_b96_v3i32(_ExtVector<3, int> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_read_tr8_b64_v2i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_ds_read_tr8_b64_v2i32(_ExtVector<2, int> address_space<3> *)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_ds_swizzle``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_ds_swizzle(int, _Constant int)

No documentation.


``__builtin_amdgcn_endpgm``
---------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_endpgm()

No documentation.


``__builtin_amdgcn_exp2_bf16``
------------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_exp2_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_exp2f``
--------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_exp2f(float)

No documentation.


``__builtin_amdgcn_fcmp``
-------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_fcmp(double, double, _Constant int)

No documentation.


``__builtin_amdgcn_fcmpf``
--------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_fcmpf(float, float, _Constant int)

No documentation.


``__builtin_amdgcn_fdot2``
--------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_fdot2(_ExtVector<2, _Float16>, _ExtVector<2, _Float16>, float, _Constant bool)

**Target Features:** dot10-insts

No documentation.


``__builtin_amdgcn_fdot2_bf16_bf16``
------------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_fdot2_bf16_bf16(_ExtVector<2, short>, _ExtVector<2, short>, short)

**Target Features:** dot9-insts

No documentation.


``__builtin_amdgcn_fdot2_f16_f16``
----------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_fdot2_f16_f16(_ExtVector<2, _Float16>, _ExtVector<2, _Float16>, _Float16)

**Target Features:** dot9-insts

No documentation.


``__builtin_amdgcn_fdot2_f32_bf16``
-----------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_fdot2_f32_bf16(_ExtVector<2, short>, _ExtVector<2, short>, float, _Constant bool)

**Target Features:** dot12-insts

No documentation.


``__builtin_amdgcn_fdot2c_f32_bf16``
------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_fdot2c_f32_bf16(_ExtVector<2, __bf16>, _ExtVector<2, __bf16>, float, _Constant bool)

**Target Features:** dot13-insts

No documentation.


``__builtin_amdgcn_fence``
--------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_fence(unsigned int, char const *, ...)

No documentation.


``__builtin_amdgcn_flat_atomic_fadd_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_flat_atomic_fadd_f32(float address_space<0> *, float)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_flat_atomic_fadd_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_flat_atomic_fadd_f64(double address_space<0> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_flat_atomic_fadd_v2bf16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_flat_atomic_fadd_v2bf16(_ExtVector<2, short> address_space<0> *, _ExtVector<2, short>)

**Target Features:** atomic-flat-pk-add-16-insts

No documentation.


``__builtin_amdgcn_flat_atomic_fadd_v2f16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_flat_atomic_fadd_v2f16(_ExtVector<2, _Float16> address_space<0> *, _ExtVector<2, _Float16>)

**Target Features:** atomic-flat-pk-add-16-insts

No documentation.


``__builtin_amdgcn_flat_atomic_fmax_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_flat_atomic_fmax_f64(double address_space<0> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_flat_atomic_fmin_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_flat_atomic_fmin_f64(double address_space<0> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_flat_load_monitor_b128``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_flat_load_monitor_b128(_ExtVector<4, int> address_space<0> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_flat_load_monitor_b32``
------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_flat_load_monitor_b32(int address_space<0> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_flat_load_monitor_b64``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_flat_load_monitor_b64(_ExtVector<2, int> address_space<0> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_flat_prefetch``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_flat_prefetch(void const address_space<0> *, _Constant int)

**Target Features:** vmem-pref-insts

No documentation.


``__builtin_amdgcn_fmed3f``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_fmed3f(float, float, float)

No documentation.


``__builtin_amdgcn_fmed3h``
---------------------------

**Prototype:**

.. code-block:: c

  __fp16 __builtin_amdgcn_fmed3h(__fp16, __fp16, __fp16)

**Target Features:** gfx9-insts

No documentation.


``__builtin_amdgcn_fract``
--------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_fract(double)

No documentation.


``__builtin_amdgcn_fractf``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_fractf(float)

No documentation.


``__builtin_amdgcn_fracth``
---------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_fracth(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_frexp_exp``
------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_frexp_exp(double)

No documentation.


``__builtin_amdgcn_frexp_expf``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_frexp_expf(float)

No documentation.


``__builtin_amdgcn_frexp_exph``
-------------------------------

**Prototype:**

.. code-block:: c

  short __builtin_amdgcn_frexp_exph(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_frexp_mant``
-------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_frexp_mant(double)

No documentation.


``__builtin_amdgcn_frexp_mantf``
--------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_frexp_mantf(float)

No documentation.


``__builtin_amdgcn_frexp_manth``
--------------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_frexp_manth(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_get_fpenv``
------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_get_fpenv()

No documentation.


``__builtin_amdgcn_global_atomic_fadd_f32``
-------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_global_atomic_fadd_f32(float address_space<1> *, float)

**Target Features:** atomic-fadd-rtn-insts

No documentation.


``__builtin_amdgcn_global_atomic_fadd_f64``
-------------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_global_atomic_fadd_f64(double address_space<1> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_global_atomic_fadd_v2bf16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_global_atomic_fadd_v2bf16(_ExtVector<2, short> address_space<1> *, _ExtVector<2, short>)

**Target Features:** atomic-global-pk-add-bf16-inst

No documentation.


``__builtin_amdgcn_global_atomic_fadd_v2f16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_global_atomic_fadd_v2f16(_ExtVector<2, _Float16> address_space<1> *, _ExtVector<2, _Float16>)

**Target Features:** atomic-buffer-global-pk-add-f16-insts

No documentation.


``__builtin_amdgcn_global_atomic_fmax_f64``
-------------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_global_atomic_fmax_f64(double address_space<1> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_global_atomic_fmin_f64``
-------------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_global_atomic_fmin_f64(double address_space<1> *, double)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_global_load_async_lds``
------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_async_lds(void address_space<1> *, void address_space<3> *, _Constant unsigned int, _Constant int, _Constant unsigned int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_global_load_async_to_lds_b128``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_async_to_lds_b128(_ExtVector<4, int> address_space<1> *, _ExtVector<4, int> address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_async_to_lds_b32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_async_to_lds_b32(int address_space<1> *, int address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_async_to_lds_b64``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_async_to_lds_b64(_ExtVector<2, int> address_space<1> *, _ExtVector<2, int> address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_async_to_lds_b8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_async_to_lds_b8(char address_space<1> *, char address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_lds``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_load_lds(void address_space<1> *, void address_space<3> *, _Constant unsigned int, _Constant int, _Constant unsigned int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_global_load_monitor_b128``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_global_load_monitor_b128(_ExtVector<4, int> address_space<1> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_monitor_b32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_global_load_monitor_b32(int address_space<1> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_monitor_b64``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_global_load_monitor_b64(_ExtVector<2, int> address_space<1> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_load_tr16_b128_v8bf16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_global_load_tr16_b128_v8bf16(_ExtVector<8, __bf16> address_space<1> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr16_b128_v8f16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __fp16> __builtin_amdgcn_global_load_tr16_b128_v8f16(_ExtVector<8, __fp16> address_space<1> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr16_b128_v8i16``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_global_load_tr16_b128_v8i16(_ExtVector<8, short> address_space<1> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr4_b64_v2i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_global_load_tr4_b64_v2i32(_ExtVector<2, int> address_space<1> *)

**Target Features:** transpose-load-f4f6-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr6_b96_v3i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, int> __builtin_amdgcn_global_load_tr6_b96_v3i32(_ExtVector<3, int> address_space<1> *)

**Target Features:** transpose-load-f4f6-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr8_b64_v2i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_global_load_tr8_b64_v2i32(_ExtVector<2, int> address_space<1> *)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v4bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, __bf16> __builtin_amdgcn_global_load_tr_b128_v4bf16(_ExtVector<4, __bf16> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v4f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, __fp16> __builtin_amdgcn_global_load_tr_b128_v4f16(_ExtVector<4, __fp16> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v4i16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, short> __builtin_amdgcn_global_load_tr_b128_v4i16(_ExtVector<4, short> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v8bf16``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_global_load_tr_b128_v8bf16(_ExtVector<8, __bf16> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v8f16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __fp16> __builtin_amdgcn_global_load_tr_b128_v8f16(_ExtVector<8, __fp16> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr_b128_v8i16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_global_load_tr_b128_v8i16(_ExtVector<8, short> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_load_tr_b64_i32``
-------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_global_load_tr_b64_i32(int address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_global_load_tr_b64_v2i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, int> __builtin_amdgcn_global_load_tr_b64_v2i32(_ExtVector<2, int> address_space<1> *)

**Target Features:** gfx12-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_global_prefetch``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_prefetch(void const address_space<1> *, _Constant int)

**Target Features:** vmem-pref-insts

No documentation.


``__builtin_amdgcn_global_store_async_from_lds_b128``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_store_async_from_lds_b128(_ExtVector<4, int> address_space<1> *, _ExtVector<4, int> address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_store_async_from_lds_b32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_store_async_from_lds_b32(int address_space<1> *, int address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_store_async_from_lds_b64``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_store_async_from_lds_b64(_ExtVector<2, int> address_space<1> *, _ExtVector<2, int> address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_global_store_async_from_lds_b8``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_global_store_async_from_lds_b8(char address_space<1> *, char address_space<3> *, _Constant int, _Constant int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_groupstaticsize``
------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_groupstaticsize()

No documentation.


``__builtin_amdgcn_iglp_opt``
-----------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_iglp_opt(_Constant int)

No documentation.


``__builtin_amdgcn_image_bvh8_intersect_ray``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<10, unsigned int> __builtin_amdgcn_image_bvh8_intersect_ray(uint64_t, float, unsigned char, _ExtVector<3, float>, _ExtVector<3, float>, unsigned int, _ExtVector<4, unsigned int>, _ExtVector<3, float> *, _ExtVector<3, float> *)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_image_bvh_dual_intersect_ray``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<10, unsigned int> __builtin_amdgcn_image_bvh_dual_intersect_ray(uint64_t, float, unsigned char, _ExtVector<3, float>, _ExtVector<3, float>, _ExtVector<2, unsigned int>, _ExtVector<4, unsigned int>, _ExtVector<3, float> *, _ExtVector<3, float> *)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_image_bvh_intersect_ray``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_image_bvh_intersect_ray(unsigned int, float, _ExtVector<4, float>, _ExtVector<4, float>, _ExtVector<4, float>, _ExtVector<4, unsigned int>)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_image_bvh_intersect_ray_h``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_image_bvh_intersect_ray_h(unsigned int, float, _ExtVector<4, float>, _ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<4, unsigned int>)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_image_bvh_intersect_ray_l``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_image_bvh_intersect_ray_l(uint64_t, float, _ExtVector<4, float>, _ExtVector<4, float>, _ExtVector<4, float>, _ExtVector<4, unsigned int>)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_image_bvh_intersect_ray_lh``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_image_bvh_intersect_ray_lh(uint64_t, float, _ExtVector<4, float>, _ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<4, unsigned int>)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_image_gather4_lz_2d_v4f32_f32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_gather4_lz_2d_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_load_1d_v4f16_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_1d_v4f16_i32(int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_1d_v4f32_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_1d_v4f32_i32(int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_1darray_v4f16_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_1darray_v4f16_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_1darray_v4f32_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_1darray_v4f32_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2d_f32_i32``
------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_load_2d_f32_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2d_v4f16_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_2d_v4f16_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2d_v4f32_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_2d_v4f32_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2darray_f32_i32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_load_2darray_f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2darray_v4f16_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_2darray_v4f16_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_2darray_v4f32_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_2darray_v4f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_3d_v4f16_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_3d_v4f16_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_3d_v4f32_i32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_3d_v4f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_cube_v4f16_i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_cube_v4f16_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_cube_v4f32_i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_cube_v4f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_1d_v4f16_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_1d_v4f16_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_1d_v4f32_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_1d_v4f32_i32(int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_1darray_v4f16_i32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_1darray_v4f16_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_1darray_v4f32_i32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_1darray_v4f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2d_f32_i32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_load_mip_2d_f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2d_v4f16_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_2d_v4f16_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2d_v4f32_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_2d_v4f32_i32(int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2darray_f32_i32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_load_mip_2darray_f32_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2darray_v4f16_i32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_2darray_v4f16_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_2darray_v4f32_i32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_2darray_v4f32_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_3d_v4f16_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_3d_v4f16_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_3d_v4f32_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_3d_v4f32_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_cube_v4f16_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_load_mip_cube_v4f16_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_load_mip_cube_v4f32_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_load_mip_cube_v4f32_i32(int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_1d_v4f16_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_1d_v4f16_f32(int, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_1d_v4f32_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_1d_v4f32_f32(int, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_1darray_v4f16_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_1darray_v4f16_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_1darray_v4f32_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_1darray_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2d_f32_f32``
--------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_2d_f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2d_v4f16_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_2d_v4f16_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2d_v4f32_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_2d_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2darray_f32_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_2darray_f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2darray_v4f16_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_2darray_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_2darray_v4f32_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_2darray_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_3d_v4f16_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_3d_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_3d_v4f32_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_3d_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_cube_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_cube_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_cube_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_cube_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_1d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_d_1d_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_1d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_d_1d_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_1darray_v4f16_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_d_1darray_v4f16_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_1darray_v4f32_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_d_1darray_v4f32_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2d_f32_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_d_2d_f32_f32(int, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_d_2d_v4f16_f32(int, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_d_2d_v4f32_f32(int, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2darray_f32_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_d_2darray_f32_f32(int, float, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2darray_v4f16_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_d_2darray_v4f16_f32(int, float, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_2darray_v4f32_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_d_2darray_v4f32_f32(int, float, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_3d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_d_3d_v4f16_f32(int, float, float, float, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_d_3d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_d_3d_v4f32_f32(int, float, float, float, float, float, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_1d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_1d_v4f16_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_1d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_1d_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_1darray_v4f16_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_1darray_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_1darray_v4f32_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_1darray_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2d_f32_f32``
----------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_l_2d_f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_2d_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_2d_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2darray_f32_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_l_2darray_f32_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2darray_v4f16_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_2darray_v4f16_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_2darray_v4f32_f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_2darray_v4f32_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_3d_v4f16_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_3d_v4f16_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_3d_v4f32_f32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_3d_v4f32_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_cube_v4f16_f32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_l_cube_v4f16_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_l_cube_v4f32_f32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_l_cube_v4f32_f32(int, float, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_1d_v4f16_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_1d_v4f16_f32(int, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_1d_v4f32_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_1d_v4f32_f32(int, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_1darray_v4f16_f32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_1darray_v4f16_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_1darray_v4f32_f32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_1darray_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2d_f32_f32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_lz_2d_f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2d_v4f16_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_2d_v4f16_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2d_v4f32_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_2d_v4f32_f32(int, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2darray_f32_f32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_image_sample_lz_2darray_f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2darray_v4f16_f32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_2darray_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_2darray_v4f32_f32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_2darray_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_3d_v4f16_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_3d_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_3d_v4f32_f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_3d_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_cube_v4f16_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_image_sample_lz_cube_v4f16_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_sample_lz_cube_v4f32_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_image_sample_lz_cube_v4f32_f32(int, float, float, float, __amdgpu_texture_t, _ExtVector<4, int>, bool, int, int)

**Target Features:** extended-image-insts

No documentation.


``__builtin_amdgcn_image_store_1d_v4f16_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_1d_v4f16_i32(_ExtVector<4, _Float16>, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_1d_v4f32_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_1d_v4f32_i32(_ExtVector<4, float>, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_1darray_v4f16_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_1darray_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_1darray_v4f32_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_1darray_v4f32_i32(_ExtVector<4, float>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2d_f32_i32``
-------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2d_f32_i32(float, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2d_v4f16_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2d_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2d_v4f32_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2d_v4f32_i32(_ExtVector<4, float>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2darray_f32_i32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2darray_f32_i32(float, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2darray_v4f16_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2darray_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_2darray_v4f32_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_2darray_v4f32_i32(_ExtVector<4, float>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_3d_v4f16_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_3d_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_3d_v4f32_i32``
---------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_3d_v4f32_i32(_ExtVector<4, float>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_cube_v4f16_i32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_cube_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_cube_v4f32_i32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_cube_v4f32_i32(_ExtVector<4, float>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_1d_v4f16_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_1d_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_1d_v4f32_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_1d_v4f32_i32(_ExtVector<4, float>, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_1darray_v4f16_i32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_1darray_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_1darray_v4f32_i32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_1darray_v4f32_i32(_ExtVector<4, float>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2d_f32_i32``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2d_f32_i32(float, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2d_v4f16_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2d_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2d_v4f32_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2d_v4f32_i32(_ExtVector<4, float>, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2darray_f32_i32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2darray_f32_i32(float, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2darray_v4f16_i32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2darray_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_2darray_v4f32_i32``
------------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_2darray_v4f32_i32(_ExtVector<4, float>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_3d_v4f16_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_3d_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_3d_v4f32_i32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_3d_v4f32_i32(_ExtVector<4, float>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_cube_v4f16_i32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_cube_v4f16_i32(_ExtVector<4, _Float16>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_image_store_mip_cube_v4f32_i32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_image_store_mip_cube_v4f32_i32(_ExtVector<4, float>, int, int, int, int, int, __amdgpu_texture_t, int, int)

**Target Features:** image-insts

No documentation.


``__builtin_amdgcn_interp_mov``
-------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_interp_mov(unsigned int, unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_interp_p1``
------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_interp_p1(float, unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_interp_p1_f16``
----------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_interp_p1_f16(float, unsigned int, unsigned int, bool, unsigned int)

No documentation.


``__builtin_amdgcn_interp_p2``
------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_interp_p2(float, float, unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_interp_p2_f16``
----------------------------------

**Prototype:**

.. code-block:: c

  __fp16 __builtin_amdgcn_interp_p2_f16(float, float, unsigned int, unsigned int, bool, unsigned int)

No documentation.


``__builtin_amdgcn_inverse_ballot_w32``
---------------------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_inverse_ballot_w32(uint32_t)

**Target Features:** wavefrontsize32

No documentation.


``__builtin_amdgcn_inverse_ballot_w64``
---------------------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_inverse_ballot_w64(uint64_t)

**Target Features:** wavefrontsize64

No documentation.


``__builtin_amdgcn_is_invocable``
---------------------------------

**Prototype:**

.. code-block:: c

  __amdgpu_feature_predicate_t __builtin_amdgcn_is_invocable()

No documentation.


``__builtin_amdgcn_is_private``
-------------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_is_private(void const address_space<0> *)

No documentation.


``__builtin_amdgcn_is_shared``
------------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_is_shared(void const address_space<0> *)

No documentation.


``__builtin_amdgcn_ldexp``
--------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_ldexp(double, int)

No documentation.


``__builtin_amdgcn_ldexpf``
---------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_ldexpf(float, int)

No documentation.


``__builtin_amdgcn_ldexph``
---------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_ldexph(_Float16, int)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_lerp``
-------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_lerp(unsigned int, unsigned int, unsigned int)

**Target Features:** lerp-inst

No documentation.


``__builtin_amdgcn_load_async_to_lds``
--------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_load_async_to_lds(void *, void address_space<3> *, _Constant unsigned int, _Constant int, _Constant unsigned int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_load_to_lds``
--------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_load_to_lds(void *, void address_space<3> *, _Constant unsigned int, _Constant int, _Constant unsigned int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_log_bf16``
-----------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_log_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_log_clampf``
-------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_log_clampf(float)

No documentation.


``__builtin_amdgcn_logf``
-------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_logf(float)

No documentation.


``__builtin_amdgcn_make_buffer_rsrc``
-------------------------------------

**Prototype:**

.. code-block:: c

  __amdgpu_buffer_rsrc_t __builtin_amdgcn_make_buffer_rsrc(void *, short, int64_t, int)

No documentation.


``__builtin_amdgcn_mbcnt_hi``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_mbcnt_hi(unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_mbcnt_lo``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_mbcnt_lo(unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x16bf16_1k``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x16bf16_1k(_ExtVector<4, short>, _ExtVector<4, short>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x16f16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x16f16(_ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x1f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_16x16x1f32(float, float, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x2bf16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_16x16x2bf16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_bf16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_bf16(_ExtVector<8, __bf16>, _ExtVector<8, __bf16>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_bf8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_bf8_bf8(int64_t, int64_t, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_bf8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_bf8_fp8(int64_t, int64_t, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_f16``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_f16(_ExtVector<8, _Float16>, _ExtVector<8, _Float16>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_fp8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_fp8_bf8(int64_t, int64_t, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x32_fp8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x32_fp8_fp8(int64_t, int64_t, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x4bf16_1k``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_16x16x4bf16_1k(_ExtVector<4, short>, _ExtVector<4, short>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x4f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_16x16x4f16(_ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x4f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x4f32(float, float, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x8_xf32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x8_xf32(_ExtVector<2, float>, _ExtVector<2, float>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_16x16x8bf16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_16x16x8bf16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_bf16``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_bf16(_ExtVector<8, __bf16>, _ExtVector<8, __bf16>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_bf8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_bf8_bf8(int64_t, int64_t, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_bf8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_bf8_fp8(int64_t, int64_t, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_f16``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_f16(_ExtVector<8, _Float16>, _ExtVector<8, _Float16>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_fp8_bf8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_fp8_bf8(int64_t, int64_t, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x16_fp8_fp8``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x16_fp8_fp8(int64_t, int64_t, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x1f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_mfma_f32_32x32x1f32(float, float, _ExtVector<32, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x2bf16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_mfma_f32_32x32x2bf16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<32, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x2f32``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x2f32(float, float, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x4_xf32``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x4_xf32(_ExtVector<2, float>, _ExtVector<2, float>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x4bf16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x4bf16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x4bf16_1k``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_mfma_f32_32x32x4bf16_1k(_ExtVector<4, short>, _ExtVector<4, short>, _ExtVector<32, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x4f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, float> __builtin_amdgcn_mfma_f32_32x32x4f16(_ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<32, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x8bf16_1k``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x8bf16_1k(_ExtVector<4, short>, _ExtVector<4, short>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f32_32x32x8f16``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_f32_32x32x8f16(_ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_4x4x1f32``
--------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_4x4x1f32(float, float, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_4x4x2bf16``
---------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_4x4x2bf16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f32_4x4x4bf16_1k``
------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_4x4x4bf16_1k(_ExtVector<4, short>, _ExtVector<4, short>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f32_4x4x4f16``
--------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_f32_4x4x4f16(_ExtVector<4, _Float16>, _ExtVector<4, _Float16>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_f64_16x16x4f64``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, double> __builtin_amdgcn_mfma_f64_16x16x4f64(double, double, _ExtVector<4, double>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_f64_4x4x4f64``
--------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_mfma_f64_4x4x4f64(double, double, double, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx90a-insts

No documentation.


``__builtin_amdgcn_mfma_i32_16x16x16i8``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_mfma_i32_16x16x16i8(int, int, _ExtVector<4, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_16x16x32_i8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_mfma_i32_16x16x32_i8(int64_t, int64_t, _ExtVector<4, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_16x16x4i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_mfma_i32_16x16x4i8(int, int, _ExtVector<16, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_16x16x64_i8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_mfma_i32_16x16x64_i8(_ExtVector<4, int>, _ExtVector<4, int>, _ExtVector<4, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_i32_32x32x16_i8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_mfma_i32_32x32x16_i8(int64_t, int64_t, _ExtVector<16, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_32x32x32_i8``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_mfma_i32_32x32x32_i8(_ExtVector<4, int>, _ExtVector<4, int>, _ExtVector<16, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_i32_32x32x4i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<32, int> __builtin_amdgcn_mfma_i32_32x32x4i8(int, int, _ExtVector<32, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_32x32x8i8``
---------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_mfma_i32_32x32x8i8(int, int, _ExtVector<16, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_i32_4x4x4i8``
-------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_mfma_i32_4x4x4i8(int, int, _ExtVector<4, int>, _Constant int, _Constant int, _Constant int)

**Target Features:** mai-insts

No documentation.


``__builtin_amdgcn_mfma_scale_f32_16x16x128_f8f6f4``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_mfma_scale_f32_16x16x128_f8f6f4(_ExtVector<8, int32_t>, _ExtVector<8, int32_t>, _ExtVector<4, float>, _Constant int, _Constant int, _Constant int, int, _Constant int, int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mfma_scale_f32_32x32x64_f8f6f4``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_mfma_scale_f32_32x32x64_f8f6f4(_ExtVector<8, int32_t>, _ExtVector<8, int32_t>, _ExtVector<16, float>, _Constant int, _Constant int, _Constant int, int, _Constant int, int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_mov_dpp``
----------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_mov_dpp(int, _Constant int, _Constant int, _Constant int, _Constant bool)

**Target Features:** dpp

No documentation.


``__builtin_amdgcn_mov_dpp8``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_mov_dpp8(unsigned int, _Constant unsigned int)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_mqsad_pk_u16_u8``
------------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_mqsad_pk_u16_u8(uint64_t, unsigned int, uint64_t)

**Target Features:** mqsad-pk-insts

No documentation.


``__builtin_amdgcn_mqsad_u32_u8``
---------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_mqsad_u32_u8(uint64_t, unsigned int, _ExtVector<4, unsigned int>)

**Target Features:** mqsad-insts

No documentation.


``__builtin_amdgcn_msad_u8``
----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_msad_u8(unsigned int, unsigned int, unsigned int)

**Target Features:** msad-insts

No documentation.


``__builtin_amdgcn_perm``
-------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_perm(unsigned int, unsigned int, unsigned int)

**Target Features:** gfx8-insts

No documentation.


``__builtin_amdgcn_perm_pk16_b4_u4``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_perm_pk16_b4_u4(unsigned int, unsigned int, _ExtVector<2, unsigned int>)

**Target Features:** tensor-cvt-lut-insts

No documentation.


``__builtin_amdgcn_perm_pk16_b6_u4``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_perm_pk16_b6_u4(unsigned int, unsigned long int, _ExtVector<2, unsigned int>)

**Target Features:** tensor-cvt-lut-insts

No documentation.


``__builtin_amdgcn_perm_pk16_b8_u4``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_perm_pk16_b8_u4(unsigned long int, unsigned long int, _ExtVector<2, unsigned int>)

**Target Features:** tensor-cvt-lut-insts

No documentation.


``__builtin_amdgcn_permlane16``
-------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_permlane16(unsigned int, unsigned int, unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_permlane16_swap``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_permlane16_swap(unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** permlane16-swap

No documentation.


``__builtin_amdgcn_permlane16_var``
-----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_permlane16_var(unsigned int, unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_permlane32_swap``
------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_permlane32_swap(unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** permlane32-swap

No documentation.


``__builtin_amdgcn_permlane64``
-------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_permlane64(unsigned int)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_permlane_bcast``
-----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_permlane_bcast(int, int, int)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_permlane_down``
----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_permlane_down(int, int, int)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_permlane_idx_gen``
-------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_permlane_idx_gen(int, int)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_permlane_up``
--------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_permlane_up(int, int, int)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_permlane_xor``
---------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_permlane_xor(int, int, int)

**Target Features:** gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_permlanex16``
--------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_permlanex16(unsigned int, unsigned int, unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_permlanex16_var``
------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_permlanex16_var(unsigned int, unsigned int, unsigned int, _Constant bool, _Constant bool)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_pk_add_max_i16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_pk_add_max_i16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<2, short>, _Constant bool)

**Target Features:** pk-add-min-max-insts

No documentation.


``__builtin_amdgcn_pk_add_max_u16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned short> __builtin_amdgcn_pk_add_max_u16(_ExtVector<2, unsigned short>, _ExtVector<2, unsigned short>, _ExtVector<2, unsigned short>, _Constant bool)

**Target Features:** pk-add-min-max-insts

No documentation.


``__builtin_amdgcn_pk_add_min_i16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, short> __builtin_amdgcn_pk_add_min_i16(_ExtVector<2, short>, _ExtVector<2, short>, _ExtVector<2, short>, _Constant bool)

**Target Features:** pk-add-min-max-insts

No documentation.


``__builtin_amdgcn_pk_add_min_u16``
-----------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned short> __builtin_amdgcn_pk_add_min_u16(_ExtVector<2, unsigned short>, _ExtVector<2, unsigned short>, _ExtVector<2, unsigned short>, _Constant bool)

**Target Features:** pk-add-min-max-insts

No documentation.


``__builtin_amdgcn_prng_b32``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_prng_b32(unsigned int)

**Target Features:** prng-inst

No documentation.


``__builtin_amdgcn_processor_is``
---------------------------------

**Prototype:**

.. code-block:: c

  __amdgpu_feature_predicate_t __builtin_amdgcn_processor_is(char const *)

No documentation.


``__builtin_amdgcn_qsad_pk_u16_u8``
-----------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_qsad_pk_u16_u8(uint64_t, unsigned int, uint64_t)

**Target Features:** qsad-insts

No documentation.


``__builtin_amdgcn_raw_buffer_load_b128``
-----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, unsigned int> __builtin_amdgcn_raw_buffer_load_b128(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_b16``
----------------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_raw_buffer_load_b16(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_b32``
----------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_raw_buffer_load_b32(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_b64``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, unsigned int> __builtin_amdgcn_raw_buffer_load_b64(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_b8``
---------------------------------------

**Prototype:**

.. code-block:: c

  unsigned char __builtin_amdgcn_raw_buffer_load_b8(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_b96``
----------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<3, unsigned int> __builtin_amdgcn_raw_buffer_load_b96(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_format_v4f16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_raw_buffer_load_format_v4f16(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_load_format_v4f32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_raw_buffer_load_format_v4f32(__amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b128``
------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b128(_ExtVector<4, unsigned int>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b16``
-----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b16(unsigned short, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b32(unsigned int, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b64(_ExtVector<2, unsigned int>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b8``
----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b8(unsigned char, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_b96``
-----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_b96(_ExtVector<3, unsigned int>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_format_v4f16``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_format_v4f16(_ExtVector<4, _Float16>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_buffer_store_format_v4f32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_buffer_store_format_v4f32(_ExtVector<4, float>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_add_i32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_raw_ptr_buffer_atomic_add_i32(int, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fadd_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_raw_ptr_buffer_atomic_fadd_f32(float, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-fadd-rtn-insts

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fadd_v2f16``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<2, _Float16> __builtin_amdgcn_raw_ptr_buffer_atomic_fadd_v2f16(_ExtVector<2, _Float16>, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-buffer-global-pk-add-f16-insts

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fmax_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_raw_ptr_buffer_atomic_fmax_f32(float, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-fmin-fmax-global-f32

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fmax_f64``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_raw_ptr_buffer_atomic_fmax_f64(double, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-fmin-fmax-global-f64

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fmin_f32``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_raw_ptr_buffer_atomic_fmin_f32(float, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-fmin-fmax-global-f32

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_atomic_fmin_f64``
---------------------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_raw_ptr_buffer_atomic_fmin_f64(double, __amdgpu_buffer_rsrc_t, int, int, _Constant int)

**Target Features:** atomic-fmin-fmax-global-f64

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_load_async_lds``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_ptr_buffer_load_async_lds(__amdgpu_buffer_rsrc_t, void address_space<3> *, _Constant unsigned int, int, int, _Constant int, _Constant int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_raw_ptr_buffer_load_lds``
--------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_raw_ptr_buffer_load_lds(__amdgpu_buffer_rsrc_t, void address_space<3> *, _Constant unsigned int, int, int, _Constant int, _Constant int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_rcp``
------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_rcp(double)

No documentation.


``__builtin_amdgcn_rcp_bf16``
-----------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_rcp_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_rcpf``
-------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_rcpf(float)

No documentation.


``__builtin_amdgcn_rcph``
-------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_rcph(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_read_exec``
------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_read_exec()

No documentation.


``__builtin_amdgcn_read_exec_hi``
---------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_read_exec_hi()

No documentation.


``__builtin_amdgcn_read_exec_lo``
---------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_read_exec_lo()

No documentation.


``__builtin_amdgcn_readfirstlane``
----------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_readfirstlane(int)

No documentation.


``__builtin_amdgcn_readlane``
-----------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_readlane(int, int)

No documentation.


``__builtin_amdgcn_rsq``
------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_rsq(double)

No documentation.


``__builtin_amdgcn_rsq_bf16``
-----------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_rsq_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_rsq_clamp``
------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_rsq_clamp(double)

No documentation.


``__builtin_amdgcn_rsq_clampf``
-------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_rsq_clampf(float)

No documentation.


``__builtin_amdgcn_rsqf``
-------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_rsqf(float)

No documentation.


``__builtin_amdgcn_rsqh``
-------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_rsqh(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_s_barrier``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier()

No documentation.


``__builtin_amdgcn_s_barrier_init``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_init(void *, int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_join``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_join(void *)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_leave``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_leave(_Constant short)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_signal``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_signal(_Constant int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_signal_isfirst``
---------------------------------------------

**Prototype:**

.. code-block:: c

  bool __builtin_amdgcn_s_barrier_signal_isfirst(_Constant int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_signal_var``
-----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_signal_var(void *, int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_barrier_wait``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_barrier_wait(_Constant short)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_buffer_prefetch_data``
-------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_buffer_prefetch_data(__amdgpu_buffer_rsrc_t, _Constant int, unsigned int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_cluster_barrier``
--------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_cluster_barrier()

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_s_dcache_inv``
---------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_dcache_inv()

No documentation.


``__builtin_amdgcn_s_dcache_inv_vol``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_dcache_inv_vol()

**Target Features:** ci-insts

No documentation.


``__builtin_amdgcn_s_dcache_wb``
--------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_dcache_wb()

**Target Features:** gfx8-insts

No documentation.


``__builtin_amdgcn_s_decperflevel``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_decperflevel(_Constant int)

No documentation.


``__builtin_amdgcn_s_get_barrier_state``
----------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_s_get_barrier_state(int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_get_named_barrier_state``
----------------------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_s_get_named_barrier_state(void *)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_getpc``
----------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_s_getpc()

No documentation.


``__builtin_amdgcn_s_getreg``
-----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_s_getreg(_Constant int)

No documentation.


``__builtin_amdgcn_s_incperflevel``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_incperflevel(_Constant int)

No documentation.


``__builtin_amdgcn_s_memrealtime``
----------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_s_memrealtime()

**Target Features:** s-memrealtime

No documentation.


``__builtin_amdgcn_s_memtime``
------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_s_memtime()

**Target Features:** s-memtime-inst

No documentation.


``__builtin_amdgcn_s_monitor_sleep``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_monitor_sleep(_Constant short)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_s_prefetch_data``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_prefetch_data(void const *, unsigned int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_sendmsg``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_sendmsg(_Constant int, unsigned int)

No documentation.


``__builtin_amdgcn_s_sendmsg_rtn``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_s_sendmsg_rtn(_Constant unsigned int)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_s_sendmsg_rtnl``
-----------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_s_sendmsg_rtnl(_Constant unsigned int)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_s_sendmsghalt``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_sendmsghalt(_Constant int, unsigned int)

No documentation.


``__builtin_amdgcn_s_setprio``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_setprio(_Constant short)

No documentation.


``__builtin_amdgcn_s_setprio_inc_wg``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_setprio_inc_wg(_Constant short)

**Target Features:** setprio-inc-wg-inst

No documentation.


``__builtin_amdgcn_s_setreg``
-----------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_setreg(_Constant int, unsigned int)

No documentation.


``__builtin_amdgcn_s_sleep``
----------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_sleep(_Constant int)

No documentation.


``__builtin_amdgcn_s_sleep_var``
--------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_sleep_var(unsigned int)

**Target Features:** gfx12-insts

No documentation.


``__builtin_amdgcn_s_ttracedata``
---------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_ttracedata(int)

No documentation.


``__builtin_amdgcn_s_ttracedata_imm``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_ttracedata_imm(_Constant short)

**Target Features:** gfx10-insts

No documentation.


``__builtin_amdgcn_s_wait_asynccnt``
------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_wait_asynccnt(_Constant unsigned short)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_s_wait_event``
---------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_wait_event(_Constant short)

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_s_wait_event_export_ready``
----------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_wait_event_export_ready()

**Target Features:** gfx11-insts

No documentation.


``__builtin_amdgcn_s_wait_tensorcnt``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_wait_tensorcnt(_Constant unsigned short)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_s_waitcnt``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_waitcnt(_Constant int)

No documentation.


``__builtin_amdgcn_s_wakeup_barrier``
-------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_s_wakeup_barrier(void *)

**Target Features:** s-wakeup-barrier-inst

No documentation.


``__builtin_amdgcn_sad_hi_u8``
------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_sad_hi_u8(unsigned int, unsigned int, unsigned int)

**Target Features:** sad-insts

No documentation.


``__builtin_amdgcn_sad_u16``
----------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_sad_u16(unsigned int, unsigned int, unsigned int)

**Target Features:** sad-insts

No documentation.


``__builtin_amdgcn_sad_u8``
---------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_sad_u8(unsigned int, unsigned int, unsigned int)

**Target Features:** sad-insts

No documentation.


``__builtin_amdgcn_sat_pk4_i4_i8``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_sat_pk4_i4_i8(unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_sat_pk4_u4_u8``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned short __builtin_amdgcn_sat_pk4_u4_u8(unsigned int)

**Target Features:** gfx1250-insts

No documentation.


``__builtin_amdgcn_sbfe``
-------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_sbfe(unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_sched_barrier``
----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_sched_barrier(_Constant int)

No documentation.


``__builtin_amdgcn_sched_group_barrier``
----------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_sched_group_barrier(_Constant int, _Constant int, _Constant int)

No documentation.


``__builtin_amdgcn_sdot2``
--------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_sdot2(_ExtVector<2, short>, _ExtVector<2, short>, int, _Constant bool)

**Target Features:** dot2-insts

No documentation.


``__builtin_amdgcn_sdot4``
--------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_sdot4(int, int, int, _Constant bool)

**Target Features:** dot1-insts

No documentation.


``__builtin_amdgcn_sdot8``
--------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_sdot8(int, int, int, _Constant bool)

**Target Features:** dot1-insts

No documentation.


``__builtin_amdgcn_set_fpenv``
------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_set_fpenv(uint64_t)

No documentation.


``__builtin_amdgcn_sicmp``
--------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_sicmp(int, int, _Constant int)

No documentation.


``__builtin_amdgcn_sicmpl``
---------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_sicmpl(int64_t, int64_t, _Constant int)

No documentation.


``__builtin_amdgcn_sin_bf16``
-----------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_sin_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_sinf``
-------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_sinf(float)

No documentation.


``__builtin_amdgcn_sinh``
-------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_sinh(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x128_bf8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x128_bf8_bf8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x128_bf8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x128_bf8_fp8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x128_fp8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x128_fp8_bf8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x128_fp8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x128_fp8_fp8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x32_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x32_bf16(_ExtVector<4, short>, _ExtVector<8, short>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x32_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x32_f16(_ExtVector<4, _Float16>, _ExtVector<8, _Float16>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_bf16(_ExtVector<8, __bf16>, _ExtVector<16, __bf16>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_bf8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_bf8_bf8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_bf8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_bf8_fp8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_f16(_ExtVector<8, _Float16>, _ExtVector<16, _Float16>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_fp8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_fp8_bf8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_16x16x64_fp8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_smfmac_f32_16x16x64_fp8_fp8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<4, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x16_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x16_bf16(_ExtVector<4, short>, _ExtVector<8, short>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x16_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x16_f16(_ExtVector<4, _Float16>, _ExtVector<8, _Float16>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_bf16(_ExtVector<8, __bf16>, _ExtVector<16, __bf16>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_bf8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_bf8_bf8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_bf8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_bf8_fp8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_f16(_ExtVector<8, _Float16>, _ExtVector<16, _Float16>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_fp8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_fp8_bf8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x32_fp8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x32_fp8_fp8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** fp8-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x64_bf8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x64_bf8_bf8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x64_bf8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x64_bf8_fp8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x64_fp8_bf8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x64_fp8_bf8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_f32_32x32x64_fp8_fp8``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, float> __builtin_amdgcn_smfmac_f32_32x32x64_fp8_fp8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<16, float>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_i32_16x16x128_i8``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_smfmac_i32_16x16x128_i8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<4, int>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_smfmac_i32_16x16x64_i8``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_smfmac_i32_16x16x64_i8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<4, int>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_i32_32x32x32_i8``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_smfmac_i32_32x32x32_i8(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<16, int>, int, _Constant int, _Constant int)

**Target Features:** gfx940-insts

No documentation.


``__builtin_amdgcn_smfmac_i32_32x32x64_i8``
-------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<16, int> __builtin_amdgcn_smfmac_i32_32x32x64_i8(_ExtVector<4, int>, _ExtVector<8, int>, _ExtVector<16, int>, int, _Constant int, _Constant int)

**Target Features:** gfx950-insts

No documentation.


``__builtin_amdgcn_sqrt``
-------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_sqrt(double)

No documentation.


``__builtin_amdgcn_sqrt_bf16``
------------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_sqrt_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_sqrtf``
--------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_sqrtf(float)

No documentation.


``__builtin_amdgcn_sqrth``
--------------------------

**Prototype:**

.. code-block:: c

  _Float16 __builtin_amdgcn_sqrth(_Float16)

**Target Features:** 16-bit-insts

No documentation.


``__builtin_amdgcn_struct_buffer_load_format_v4f16``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, _Float16> __builtin_amdgcn_struct_buffer_load_format_v4f16(__amdgpu_buffer_rsrc_t, int, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_struct_buffer_load_format_v4f32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_struct_buffer_load_format_v4f32(__amdgpu_buffer_rsrc_t, int, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_struct_buffer_store_format_v4f16``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_struct_buffer_store_format_v4f16(_ExtVector<4, _Float16>, __amdgpu_buffer_rsrc_t, int, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_struct_buffer_store_format_v4f32``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_struct_buffer_store_format_v4f32(_ExtVector<4, float>, __amdgpu_buffer_rsrc_t, int, int, int, _Constant int)

No documentation.


``__builtin_amdgcn_struct_ptr_buffer_load_async_lds``
-----------------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_struct_ptr_buffer_load_async_lds(__amdgpu_buffer_rsrc_t, void address_space<3> *, _Constant unsigned int, int, int, int, _Constant int, _Constant int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_struct_ptr_buffer_load_lds``
-----------------------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_struct_ptr_buffer_load_lds(__amdgpu_buffer_rsrc_t, void address_space<3> *, _Constant unsigned int, int, int, int, _Constant int, _Constant int)

**Target Features:** vmem-to-lds-load-insts

No documentation.


``__builtin_amdgcn_sudot4``
---------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_sudot4(_Constant bool, int, _Constant bool, int, int, _Constant bool)

**Target Features:** dot8-insts

No documentation.


``__builtin_amdgcn_sudot8``
---------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_sudot8(_Constant bool, int, _Constant bool, int, int, _Constant bool)

**Target Features:** dot8-insts

No documentation.


``__builtin_amdgcn_swmmac_bf16_16x16x32_bf16_w32``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, short> __builtin_amdgcn_swmmac_bf16_16x16x32_bf16_w32(_ExtVector<8, short>, _ExtVector<16, short>, _ExtVector<8, short>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_bf16_16x16x32_bf16_w64``
--------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, short> __builtin_amdgcn_swmmac_bf16_16x16x32_bf16_w64(_ExtVector<4, short>, _ExtVector<8, short>, _ExtVector<4, short>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_bf16_16x16x64_bf16``
----------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __bf16> __builtin_amdgcn_swmmac_bf16_16x16x64_bf16(_Constant bool, _ExtVector<16, __bf16>, _Constant bool, _ExtVector<32, __bf16>, _ExtVector<8, __bf16>, int, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_bf16f32_16x16x64_bf16``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_bf16f32_16x16x64_bf16(_Constant bool, _ExtVector<16, __bf16>, _Constant bool, _ExtVector<32, __bf16>, _ExtVector<8, float>, int, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x128_bf8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_swmmac_f16_16x16x128_bf8_bf8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, _Float16>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x128_bf8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_swmmac_f16_16x16x128_bf8_fp8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, _Float16>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x128_fp8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_swmmac_f16_16x16x128_fp8_bf8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, _Float16>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x128_fp8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_swmmac_f16_16x16x128_fp8_fp8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, _Float16>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x32_f16_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, __fp16> __builtin_amdgcn_swmmac_f16_16x16x32_f16_w32(_ExtVector<8, __fp16>, _ExtVector<16, __fp16>, _ExtVector<8, __fp16>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x32_f16_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, __fp16> __builtin_amdgcn_swmmac_f16_16x16x32_f16_w64(_ExtVector<4, __fp16>, _ExtVector<8, __fp16>, _ExtVector<4, __fp16>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f16_16x16x64_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, _Float16> __builtin_amdgcn_swmmac_f16_16x16x64_f16(_Constant bool, _ExtVector<16, _Float16>, _Constant bool, _ExtVector<32, _Float16>, _ExtVector<8, _Float16>, int, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x128_bf8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x128_bf8_bf8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, float>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x128_bf8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x128_bf8_fp8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, float>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x128_fp8_bf8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x128_fp8_bf8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, float>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x128_fp8_fp8``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x128_fp8_fp8(_ExtVector<8, int>, _ExtVector<16, int>, _ExtVector<8, float>, _ExtVector<2, int>, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf16_w32``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf16_w32(_ExtVector<8, short>, _ExtVector<16, short>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf16_w64``
-------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf16_w64(_ExtVector<4, short>, _ExtVector<8, short>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf8_bf8_w32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf8_bf8_w32(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf8_bf8_w64``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf8_bf8_w64(int, _ExtVector<2, int>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf8_fp8_w32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf8_fp8_w32(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_bf8_fp8_w64``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_bf8_fp8_w64(int, _ExtVector<2, int>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_f16_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_f16_w32(_ExtVector<8, __fp16>, _ExtVector<16, __fp16>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_f16_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_f16_w64(_ExtVector<4, __fp16>, _ExtVector<8, __fp16>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_fp8_bf8_w32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_fp8_bf8_w32(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_fp8_bf8_w64``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_fp8_bf8_w64(int, _ExtVector<2, int>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_fp8_fp8_w32``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x32_fp8_fp8_w32(_ExtVector<2, int>, _ExtVector<4, int>, _ExtVector<8, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x32_fp8_fp8_w64``
----------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, float> __builtin_amdgcn_swmmac_f32_16x16x32_fp8_fp8_w64(int, _ExtVector<2, int>, _ExtVector<4, float>, int)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x64_bf16``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x64_bf16(_Constant bool, _ExtVector<16, __bf16>, _Constant bool, _ExtVector<32, __bf16>, _ExtVector<8, float>, int, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_f32_16x16x64_f16``
--------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, float> __builtin_amdgcn_swmmac_f32_16x16x64_f16(_Constant bool, _ExtVector<16, _Float16>, _Constant bool, _ExtVector<32, _Float16>, _ExtVector<8, float>, int, _Constant bool, _Constant bool)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x128_iu8``
---------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_swmmac_i32_16x16x128_iu8(_Constant bool, _ExtVector<8, int>, _Constant bool, _ExtVector<16, int>, _ExtVector<8, int>, _ExtVector<2, int>, _Constant bool, _Constant bool, ...)

**Target Features:** swmmac-gfx1250-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x32_iu4_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_swmmac_i32_16x16x32_iu4_w32(_Constant bool, int, _Constant bool, _ExtVector<2, int>, _ExtVector<8, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x32_iu4_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_swmmac_i32_16x16x32_iu4_w64(_Constant bool, int, _Constant bool, int, _ExtVector<4, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x32_iu8_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_swmmac_i32_16x16x32_iu8_w32(_Constant bool, _ExtVector<2, int>, _Constant bool, _ExtVector<4, int>, _ExtVector<8, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x32_iu8_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_swmmac_i32_16x16x32_iu8_w64(_Constant bool, int, _Constant bool, _ExtVector<2, int>, _ExtVector<4, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x64_iu4_w32``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<8, int> __builtin_amdgcn_swmmac_i32_16x16x64_iu4_w32(_Constant bool, _ExtVector<2, int>, _Constant bool, _ExtVector<4, int>, _ExtVector<8, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize32

No documentation.


``__builtin_amdgcn_swmmac_i32_16x16x64_iu4_w64``
------------------------------------------------

**Prototype:**

.. code-block:: c

  _ExtVector<4, int> __builtin_amdgcn_swmmac_i32_16x16x64_iu4_w64(_Constant bool, int, _Constant bool, _ExtVector<2, int>, _ExtVector<4, int>, int, _Constant bool)

**Target Features:** swmmac-gfx1200-insts,wavefrontsize64

No documentation.


``__builtin_amdgcn_tanh_bf16``
------------------------------

**Prototype:**

.. code-block:: c

  __bf16 __builtin_amdgcn_tanh_bf16(__bf16)

**Target Features:** bf16-trans-insts

No documentation.


``__builtin_amdgcn_tanhf``
--------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_tanhf(float)

**Target Features:** tanh-insts

No documentation.


``__builtin_amdgcn_tanhh``
--------------------------

**Prototype:**

.. code-block:: c

  __fp16 __builtin_amdgcn_tanhh(__fp16)

**Target Features:** tanh-insts

No documentation.


``__builtin_amdgcn_trig_preop``
-------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_trig_preop(double, int)

No documentation.


``__builtin_amdgcn_trig_preopf``
--------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_trig_preopf(float, int)

No documentation.


``__builtin_amdgcn_ubfe``
-------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_ubfe(unsigned int, unsigned int, unsigned int)

No documentation.


``__builtin_amdgcn_udot2``
--------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_udot2(_ExtVector<2, unsigned short>, _ExtVector<2, unsigned short>, unsigned int, _Constant bool)

**Target Features:** dot2-insts

No documentation.


``__builtin_amdgcn_udot4``
--------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_udot4(unsigned int, unsigned int, unsigned int, _Constant bool)

**Target Features:** dot7-insts

No documentation.


``__builtin_amdgcn_udot8``
--------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_udot8(unsigned int, unsigned int, unsigned int, _Constant bool)

**Target Features:** dot7-insts

No documentation.


``__builtin_amdgcn_uicmp``
--------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_uicmp(unsigned int, unsigned int, _Constant int)

No documentation.


``__builtin_amdgcn_uicmpl``
---------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_uicmpl(uint64_t, uint64_t, _Constant int)

No documentation.


``__builtin_amdgcn_update_dpp``
-------------------------------

**Prototype:**

.. code-block:: c

  int __builtin_amdgcn_update_dpp(int, int, _Constant int, _Constant int, _Constant int, _Constant bool)

**Target Features:** dpp

No documentation.


``__builtin_amdgcn_wait_asyncmark``
-----------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_wait_asyncmark(_Constant unsigned short)

**Target Features:** vmem-to-lds-load-insts|asynccnt

No documentation.


``__builtin_amdgcn_wave_barrier``
---------------------------------

**Prototype:**

.. code-block:: c

  void __builtin_amdgcn_wave_barrier()

No documentation.


``__builtin_amdgcn_wave_reduce_add_u32``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_wave_reduce_add_u32(uint32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_add_u64``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_wave_reduce_add_u64(uint64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_and_b32``
----------------------------------------

**Prototype:**

.. code-block:: c

  int32_t __builtin_amdgcn_wave_reduce_and_b32(int32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_and_b64``
----------------------------------------

**Prototype:**

.. code-block:: c

  int64_t __builtin_amdgcn_wave_reduce_and_b64(int64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fadd_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_wave_reduce_fadd_f32(float, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fadd_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_wave_reduce_fadd_f64(double, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fmax_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_wave_reduce_fmax_f32(float, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fmax_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_wave_reduce_fmax_f64(double, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fmin_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_wave_reduce_fmin_f32(float, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fmin_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_wave_reduce_fmin_f64(double, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fsub_f32``
-----------------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_amdgcn_wave_reduce_fsub_f32(float, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_fsub_f64``
-----------------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_amdgcn_wave_reduce_fsub_f64(double, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_max_i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  int32_t __builtin_amdgcn_wave_reduce_max_i32(int32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_max_i64``
----------------------------------------

**Prototype:**

.. code-block:: c

  int64_t __builtin_amdgcn_wave_reduce_max_i64(int64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_max_u32``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_wave_reduce_max_u32(uint32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_max_u64``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_wave_reduce_max_u64(uint64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_min_i32``
----------------------------------------

**Prototype:**

.. code-block:: c

  int32_t __builtin_amdgcn_wave_reduce_min_i32(int32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_min_i64``
----------------------------------------

**Prototype:**

.. code-block:: c

  int64_t __builtin_amdgcn_wave_reduce_min_i64(int64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_min_u32``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_wave_reduce_min_u32(uint32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_min_u64``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_wave_reduce_min_u64(uint64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_or_b32``
---------------------------------------

**Prototype:**

.. code-block:: c

  int32_t __builtin_amdgcn_wave_reduce_or_b32(int32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_or_b64``
---------------------------------------

**Prototype:**

.. code-block:: c

  int64_t __builtin_amdgcn_wave_reduce_or_b64(int64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_sub_u32``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint32_t __builtin_amdgcn_wave_reduce_sub_u32(uint32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_sub_u64``
----------------------------------------

**Prototype:**

.. code-block:: c

  uint64_t __builtin_amdgcn_wave_reduce_sub_u64(uint64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_xor_b32``
----------------------------------------

**Prototype:**

.. code-block:: c

  int32_t __builtin_amdgcn_wave_reduce_xor_b32(int32_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wave_reduce_xor_b64``
----------------------------------------

**Prototype:**

.. code-block:: c

  int64_t __builtin_amdgcn_wave_reduce_xor_b64(int64_t, _Constant int32_t)

No documentation.


``__builtin_amdgcn_wavefrontsize``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_amdgcn_wavefrontsize()

No documentation.


``__builtin_r600_implicitarg_ptr``
----------------------------------

**Prototype:**

.. code-block:: c

  unsigned char address_space<7> * __builtin_r600_implicitarg_ptr()

No documentation.


``__builtin_r600_read_tgid_x``
------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tgid_x()

No documentation.


``__builtin_r600_read_tgid_y``
------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tgid_y()

No documentation.


``__builtin_r600_read_tgid_z``
------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tgid_z()

No documentation.


``__builtin_r600_read_tidig_x``
-------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tidig_x()

No documentation.


``__builtin_r600_read_tidig_y``
-------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tidig_y()

No documentation.


``__builtin_r600_read_tidig_z``
-------------------------------

**Prototype:**

.. code-block:: c

  unsigned int __builtin_r600_read_tidig_z()

No documentation.


``__builtin_r600_recipsqrt_ieee``
---------------------------------

**Prototype:**

.. code-block:: c

  double __builtin_r600_recipsqrt_ieee(double)

No documentation.


``__builtin_r600_recipsqrt_ieeef``
----------------------------------

**Prototype:**

.. code-block:: c

  float __builtin_r600_recipsqrt_ieeef(float)

No documentation.


