Faster versions of fvec_op_ny_Dx for AVX2 (#2811) · alexanderguzhva/faiss@d7ccdc4 · GitHub

Commit

Faster versions of fvec_op_ny_Dx for AVX2 (facebookresearch#2811)

Browse files

Summary:
Pull Request resolved: facebookresearch#2811

Use transpose_AxB kernels to speedup computations.

Differential Revision: D44726814

fbshipit-source-id: 566b1179e4deea53cce24f45a5106e7b52b23641

Loading branch information

Alexandr Guzhva authored and facebook-github-bot committed Apr 7, 2023

1 parent c22b980 commit d7ccdc4