Jatin Chaudhary 80f02a1534 SWDEV-379007 - use avx instruction for bf16 cvt ...

AMD CPUs have had avx512_bf16 support for quite some time now (from
consumer Ryzen 7000 series to enterprise grade CPUs). This
patch should allow users to use the hardware bf16 unit when running the
__host__ variants of the function. This can be enabled via `hipcc ...
-mavx512vl -mavx512bf16`.

Change-Id: I67c377afc95ddfe8d45a048dce078a247d4a1878


[ROCm/clr commit: 49349f168c]

2024-04-16 18:35:08 -04:00

..

amd_channel_descriptor.h

SWDEV-430963 - Address MIGraphX failures due to unsafe buffer access usage

2023-11-09 04:39:49 -04:00

amd_device_functions.h

SWDEV-445960 - Fix __umul64hi __mul64hi on windows

2024-03-03 12:42:42 -05:00

amd_hip_atomic.h

SWDEV-448288 - Remove references to deprecated llvm references.

2024-03-26 10:23:18 -04:00

amd_hip_bf16.h

SWDEV-379007 - use avx instruction for bf16 cvt

2024-04-16 18:35:08 -04:00

amd_hip_bfloat16.h

Revert "SWDEV-431175 - Remove datatypes from HIPRTC"

2023-12-05 02:32:38 -05:00

amd_hip_common.h

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

amd_hip_complex.h

SWDEV-430963 - Address MIGraphX failures due to unsafe buffer access usage

2023-11-09 04:39:49 -04:00

amd_hip_cooperative_groups.h

SWDEV-448281 - Fix for CG Tile Partition Dynamic Size

2024-03-04 13:00:24 -05:00

amd_hip_fp8.h

SWDEV-379007 initial implementation of fp8 header

2024-04-16 05:37:59 -04:00

amd_hip_fp16.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00

amd_hip_gl_interop.h

SWDEV-410522 - Fix a doxygen warning in a header in AMD platform

2024-01-11 13:40:25 -05:00

amd_hip_math_constants.h

SWDEV-395996 - Add double precision constants to math constants header

2023-06-05 03:31:27 -04:00

amd_hip_runtime_pt_api.h

SWDEV-433821 - Make hipStreamWaitEvent flags a default argument

2023-12-06 12:37:35 -05:00

amd_hip_runtime.h

SWDEV-435750 - Make threadIdx, blockIdx compatible with CUDA

2024-01-18 02:16:15 -05:00

amd_hip_unsafe_atomics.h

SWDEV-430963 - Address MIGraphX failures due to unsafe buffer access usage

2023-11-09 04:39:49 -04:00

amd_hip_vector_types.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00

amd_math_functions.h

SWDEV-434362 - Replace __HIP_CLANG_ONLY__ as it is not supported on HIPRTC

2023-11-29 04:22:19 -05:00

amd_surface_functions.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00

amd_warp_functions.h

SWDEV-295298: hide warp sync builtins with a macro in ROCm 6.1

2024-02-05 05:08:11 +00:00

amd_warp_sync_functions.h

SWDEV-295298: hide warp sync builtins in ROCm 6.2 (absent in 6.1)

2024-03-11 02:44:20 -04:00

concepts.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

device_library_decls.h

Revert "SWDEV-431175 - Remove datatypes from HIPRTC"

2023-12-05 02:32:38 -05:00

functional_grid_launch.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

grid_launch_GGL.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

grid_launch.h

ROCMOPS-1956 - Push restructured code to hipamd

2021-06-07 21:42:44 +00:00

grid_launch.hpp

ROCMOPS-1956 - Push restructured code to hipamd

2021-06-07 21:42:44 +00:00

helpers.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

hip_api_trace.hpp

SWDEV-444021 - Implement hipGetFuncBySymbol

2024-03-14 02:56:30 -04:00

hip_assert.h

SWDEV-437090: move the __hip_assert macro to a common place

2023-12-14 09:03:14 -05:00

hip_cooperative_groups_helper.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00

hip_fp16_gcc.h

ROCMOPS-1956 - Push restructured code to hipamd

2021-06-07 21:42:44 +00:00

hip_fp16_math_fwd.h

SWDEV-395413 - Fix fp16 header errors with -Werror in hiprtc

2023-04-26 13:05:20 -04:00

hip_ldg.h

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

hip_prof_str.h

SWDEV-444021 - Implement hipGetFuncBySymbol

2024-03-14 02:56:30 -04:00

hip_runtime_prof.h

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

host_defines.h

SWDEV-379007 initial implementation of fp8 header

2024-04-16 05:37:59 -04:00

hsa_helpers.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

macro_based_grid_launch.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

math_fwd.h

SWDEV-430963 - Address MIGraphX failures due to unsafe buffer access usage

2023-11-09 04:39:49 -04:00

ockl_image.h

SWDEV-430963 - Address MIGraphX failures due to unsafe buffer access usage

2023-11-09 04:39:49 -04:00

program_state.hpp

SWDEV-293742 - Update copyrights end year for hipamd

2021-07-09 12:08:39 -04:00

texture_fetch_functions.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00

texture_indirect_functions.h

Revert "SWDEV-419480 - HIPRTC implementation of standard headers shouldn't be in std namespace"

2023-12-06 03:11:59 -05:00