Re: git: 8b4684afcde3 - main - lib/libmd: add optimised SHA1 implementations for amd64

In reply to: Robert Clausecker : "git: 8b4684afcde3 - main - lib/libmd: add optimised SHA1 implementations for amd64"
Go to: [ bottom of page ] [ top of archives ] [ this month ]

From: Alexey Dokuchaev <danfe_at_freebsd.org>
Date: Thu, 15 May 2025 03:03:16 UTC

On Wed, May 14, 2025 at 11:40:38PM +0000, Robert Clausecker wrote:
> commit 8b4684afcde3930eb49490f0b8431c4cb2ad9a46
> 
>   lib/libmd: add optimised SHA1 implementations for amd64
> 
>   Three implementations are provided: one using just scalar
>   instructions, one using AVX2, and one using the SHA instructions
>   (SHANI).  The AVX2 version uses a complicated multi-block carry
>   scheme described in an Intel whitepaper

Sounds pretty cool!  How feasible would be coming up with AVX-only code?
My CPU does not support AVX2 but I often need to compute SHA1 sums.

./danfe