Added tesnorizeation for avx2 based gemm. (#3982)
authorKimish Patel <kimishpatel@fb.com>
Wed, 25 Sep 2019 16:52:09 +0000 (09:52 -0700)
committerTianqi Chen <tqchen@users.noreply.github.com>
Wed, 25 Sep 2019 16:52:09 +0000 (09:52 -0700)
commit23727eb49ea71609fc29963b996a68a14fddf79c
treed35f8f267e47a2c0ef47c499790e4152dd089667
parent9baff0099f5b97d8eb82d0dd1ac1aeee0ecb0615
Added tesnorizeation for avx2 based gemm. (#3982)

* Added tesnorizeation for avx2 based gemm.

Summary:
Tensorized the same region as avx512. Names produce 16x1 int32 results.
Does by doing two sets of AVX2 instructions to do reduction on 8x4 int8
kernel with 1x4 data.

Test Plan:
on avx2 machine:
python tests/python/contrib/test_gemm_avx2_acc32.py

Reviewers:

Subscribers:

Tasks:

Tags:

* Fix lint errors. Removed commented out code.

Summary:

Test Plan:

Reviewers:

Subscribers:

Tasks:

Tags:
tests/python/contrib/test_gemm_avx2_acc32.py [new file with mode: 0644]
topi/python/topi/x86/tensor_intrin.py