Commit daf99296 authored by Sanjoy Das's avatar Sanjoy Das Committed by TensorFlower Gardener
Browse files

[XLA:CPU] Enable LLVM IR GEMM for small matrices. 

We seem to be comfortably better than Eigen for matrices smaller than
32x128x128 on my local machine.

Also rename "block panel" to "tiled small gemm".  "block panel" was never the
right name for the kernel -- "block panel" is a way of tiling a large gemm into
smaller chunks, not a way of implementing the small gemm.

PiperOrigin-RevId: 209671700
parent b1b2cb38
Show whitespace changes
Inline Side-by-side
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please to comment