haozhe.zhu 7cd6e6acad add bf16 in fp32 out fast path for embedingbag in caffe2 perfkernel (#89198) ...

Add BF16 in FP32 out kernel into Caffe2 emb perfkernels. And also update the python code-gen files to generate the kernel.
The ut will be covered in the next PR(#89199) in this stack ( Tested by nn.EmbeddingBag with BF16 data type)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/89198
Approved by: https://github.com/jgong5, https://github.com/kit1980

2022-11-30 13:06:13 +00:00

..

__init__.py

remediation of S205607

2020-07-17 17:19:47 -07:00

adagrad_avx2.cc

Fix sign-compare in caffe2

2022-04-05 00:08:05 +00:00

adagrad.cc

[NOOP][clangformat][codemod] Enable CLANGFORMAT for caffe2/caffe2/* (#67624)

2021-11-02 22:14:04 -07:00

adagrad.h

use irange for loops (#70248)

2022-01-06 23:14:29 -08:00

batch_box_cox_avx2.cc

[tourch] BatchBoxCox - fix numerical issue in vectorized code (#88875)

2022-11-11 21:58:23 +00:00

batch_box_cox.cc

[torch] Unify batch_box_cox implementations into perfkernels folder (#86569)

2022-10-23 19:29:25 +00:00

batch_box_cox.h

[torch] Unify batch_box_cox implementations into perfkernels folder (#86569)

2022-10-23 19:29:25 +00:00

CMakeLists.txt

Remove caffe2 mobile (#84338)

2022-09-08 01:49:55 +00:00

common_avx2.cc

Remove Apache headers from source.

2018-03-27 13:10:18 -07:00

common_avx512.cc

include avx512vl to avx512 code path (#14733)

2018-12-05 00:50:51 -08:00

common_avx.cc

Remove Apache headers from source.

2018-03-27 13:10:18 -07:00

common.h

[torch] Unify batch_box_cox implementations into perfkernels folder (#86569)

2022-10-23 19:29:25 +00:00

cvtsh_ss_bugfix.h

[NOOP][clangformat][codemod] Enable CLANGFORMAT for caffe2/caffe2/* (#67624)

2021-11-02 22:14:04 -07:00

embedding_lookup_avx2.cc

[clang-tidy] Exclude cppcoreguidelines-avoid-magic-numbers (#57841)

2021-05-07 20:02:33 -07:00

embedding_lookup_fused_8bit_rowwise_avx2.cc

[clang-tidy] Exclude cppcoreguidelines-avoid-magic-numbers (#57841)

2021-05-07 20:02:33 -07:00

embedding_lookup_fused_8bit_rowwise_idx_avx2.cc

[clang-tidy] Exclude cppcoreguidelines-avoid-magic-numbers (#57841)

2021-05-07 20:02:33 -07:00

embedding_lookup_idx_avx2.cc

add bf16 in fp32 out fast path for embedingbag in caffe2 perfkernel (#89198)

2022-11-30 13:06:13 +00:00

embedding_lookup_idx.cc

add bf16 in fp32 out fast path for embedingbag in caffe2 perfkernel (#89198)

2022-11-30 13:06:13 +00:00

embedding_lookup_idx.h

Support int32 indices and offsets in nn.EmbeddingBag (#46758)

2020-11-03 23:33:50 -08:00

embedding_lookup.cc

[NOOP][clangformat][codemod] Enable CLANGFORMAT for caffe2/caffe2/* (#67624)

2021-11-02 22:14:04 -07:00

embedding_lookup.h

minor comment fix (#22140)

2019-08-08 21:08:47 -07:00

fused_8bit_rowwise_embedding_lookup_idx.cc

[NOOP][clangformat][codemod] Enable CLANGFORMAT for caffe2/caffe2/* (#67624)

2021-11-02 22:14:04 -07:00

fused_8bit_rowwise_embedding_lookup_idx.h

Support int32 indices and offsets in nn.EmbeddingBag (#46758)

2020-11-03 23:33:50 -08:00

fused_8bit_rowwise_embedding_lookup.cc

[NOOP][clangformat][codemod] Enable CLANGFORMAT for caffe2/caffe2/* (#67624)

2021-11-02 22:14:04 -07:00

fused_8bit_rowwise_embedding_lookup.h

fix comment index_size->output_size (#29831)

2019-11-16 01:49:02 -08:00

fused_nbit_rowwise_conversion.cc

Allow row sizes to support int64/size_t. (#69303)

2021-12-14 10:09:08 -08:00

fused_nbit_rowwise_conversion.h

Allow row sizes to support int64/size_t. (#69303)

2021-12-14 10:09:08 -08:00

hp_emblookup_codegen.py

add bf16 in fp32 out fast path for embedingbag in caffe2 perfkernel (#89198)

2022-11-30 13:06:13 +00:00

lstm_unit_cpu_avx2.cc

[caffe2] Explicit vectorization of LSTM operator (#35556)

2020-04-01 17:19:56 -07:00

lstm_unit_cpu_common.cc

[caffe2] Explicit vectorization of LSTM operator (#35556)

2020-04-01 17:19:56 -07:00

lstm_unit_cpu_common.h

[caffe2] Explicit vectorization of LSTM operator (#35556)

2020-04-01 17:19:56 -07:00

lstm_unit_cpu-impl.h

[caffe2][tourch] Optimize BatchBoxCox (#87585)

2022-11-10 06:11:05 +00:00

lstm_unit_cpu.h

[caffe2] Explicit vectorization of LSTM operator (#35556)

2020-04-01 17:19:56 -07:00

math_cpu_avx2.cc

[clang-tidy] Exclude cppcoreguidelines-avoid-magic-numbers (#57841)

2021-05-07 20:02:33 -07:00

math_cpu_base.cc

[clang-tidy] Exclude cppcoreguidelines-avoid-magic-numbers (#57841)

2021-05-07 20:02:33 -07:00

math.h

Resolve errors in perfkernel for Windows (#16031)

2019-01-16 21:51:00 -08:00

typed_axpy_avx2.cc

Make PyTorch code-base clang-tidy compliant (#56892)

2021-04-28 14:10:25 -07:00

typed_axpy_avx.cc

Make PyTorch code-base clang-tidy compliant (#56892)

2021-04-28 14:10:25 -07:00

typed_axpy.cc

Make PyTorch code-base clang-tidy compliant (#56892)

2021-04-28 14:10:25 -07:00

typed_axpy.h

Remove Apache headers from source.

2018-03-27 13:10:18 -07:00

vectorizer.h

[caffe2][tourch] Optimize BatchBoxCox (#87585)

2022-11-10 06:11:05 +00:00