[Feature] Enable AMX FP16 on next generation CPU #456
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
changqi1
reviewed
Jun 19, 2024
src/utils/matmul_helper.h
Outdated
| TimeLine t("onednn_amx_sgemm_f32bf16f32_compute"); | ||
| TimeLine t1("onednn_amx_sgemm_f32bf16f32_compute.create_primitive"); | ||
| template <typename Tin, typename Tweight, typename Tout> | ||
| void onednn_amx_sgemm_f32x16f32_compute(bool transA, int M, int N, int K, float alpha, const Tin *A, int lda, |
There was a problem hiding this comment.
reanme onednn_amx_gemm_compute?
src/utils/matmul_helper.h
Outdated
| void onednn_amx_sgemm_f32bf16f32_compute_residential(bool transA, int M, int N, int K, float alpha, const Tin *A, | ||
| int lda, const bfloat16_t *packedB, float beta, Tout *C, int ldc, const float *bias, const Tin *res, | ||
| template <typename Tin, typename Tweight, typename Tout> | ||
| void onednn_amx_sgemm_f32x16f32_compute_residential(bool transA, int M, int N, int K, float alpha, const Tin *A, |
There was a problem hiding this comment.
reanme onednn_amx_gemm_compute_residential?
|
除了compute和compute_resident，其他的kernel也需要支持这个amx fp16的支持 |
wenhuanh
force-pushed
the
feature/fp16_amx
branch
from
ad11ced
to
ee2c349
Compare
|
revised and updated. |
|
@wenhuanh We will need to check ISA to select AMX_FP16 or AVX512_FP16. But currently this is OK. |
need export env DNNL_MAX_CPU_ISA=AVX512_CORE_AMX_FP16 when running Signed-off-by: Wenhuan Huang <wenhuan.huang@intel.com>
wenhuanh
force-pushed
the
feature/fp16_amx
branch
from
ee2c349
to
2190d1f
Compare
|
Perf status was updated in related JIRA. |
changqi1
approved these changes
Jun 25, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
need export env DNNL_MAX_CPU_ISA=AVX512_CORE_AMX_FP16 when running