Re: [PATCH] wc: Add AVX2 optimization when counting only lines

[Pádraig Brady]

On 21/04/2021 19:13, Assaf Gordon wrote:
> Hello,
>
> On 2021-03-29 7:21 a.m., Pádraig Brady wrote:
>
> > On 28/03/2021 18:29, Kristoffer Brånemyr via GNU coreutils General
> > > I wanted to practice some more using vector intrinsics, so I made a
> > > small AVX2 optimization for wc -l. Depending on line length it is
> > > about 2-5x faster than previous version. (Well, only looking at user
> > > time it is much faster than that even.)
> > Excellent results.
> > I'll review this very soon.
>
> I'm attaching the patch (copied from the Github's pull-request),
> hopefully we can continue the discussion here on the mailing list.

I plan to push the attached 2 commits tomorrow.
The first adjusts the original patch to pass `make syntax-check`.
Also I noticed an inconsistency in the new wc_lines() function,
between BUFFER_SIZE and BUFSIZ, and changed uses to the former.
The second commit, adds a --debug option to indicate
the now runtime variable behavior or which implementation is used.

cheers,
Pádraig

wc-avx.patch
Description: Text Data