From a61382e9dec22ea1ff3fea2d59d3c605c0aa2aac Mon Sep 17 00:00:00 2001 From: Sebastiano Tronto Date: Thu, 2 Nov 2023 20:05:29 +0100 Subject: Optimized compose for avx2 --- README.md | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'README.md') diff --git a/README.md b/README.md index a0d2ae2..2990c29 100644 --- a/README.md +++ b/README.md @@ -17,7 +17,8 @@ $ make test * inline moves for avx2 * fix base get_ and set_ macros (constant arguments?) -* optimize things that use get_ and set_ +* optimize inverse for avx2 +* other things to optimize? ### Documentation and interface @@ -38,12 +39,9 @@ $ make test ### Things I need to learn: -* Use AVX2 instructions, in particular - [_mm256_shuffle_epi8](https://www.intel.com/content/www/us/en/docs/cpp-compiler/developer-guide-reference/2021-10/mm256-shuffle-epi8.html)) * Inspect compiled assembly * Use valgrind tool cachegrind and other profiling tools - ## Internal representation of the cube The plan (TODO) is to have multiple implementations: some that -- cgit v1.3