From: "H.J. Lu" <hjl.tools@gmail.com>
To: Noah Goldstein <goldstein.w.n@gmail.com>
Cc: GNU C Library <libc-alpha@sourceware.org>,
"Carlos O'Donell" <carlos@systemhalted.org>
Subject: Re: [PATCH v1 2/3] x86: Remove memcmp-sse4.S
Date: Fri, 15 Apr 2022 10:20:34 -0700 [thread overview]
Message-ID: <CAMe9rOrgnfAfCOggWy78epsAzE+uN8PJ+rec9e_fkdJhXBJmRQ@mail.gmail.com> (raw)
In-Reply-To: <20220415055132.1257272-2-goldstein.w.n@gmail.com>
On Thu, Apr 14, 2022 at 10:51 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
>
> Code didn't actually use any sse4 instructions. The new memcmp-sse2
> implementation is also faster.
Please mention that SSE4.1 ptest instruction was removed by
commit 2f9062d7171850451e6044ef78d91ff8c017b9c0
Author: Noah Goldstein <goldstein.w.n@gmail.com>
Date: Wed Nov 10 16:18:56 2021 -0600
x86: Shrink memcmp-sse4.S code size
> geometric_mean(N=20) of page cross cases SSE2 / SSE4: 0.905
>
> Note there are two regressions prefering SSE2 for Size = 1 and Size =
> 65.
>
> Size = 1:
> size, align0, align1, ret, New Time/Old Time
> 1, 1, 1, 0, 1.2
> 1, 1, 1, 1, 1.197
> 1, 1, 1, -1, 1.2
>
> This is intentional. Size == 1 is significantly less hot based on
> profiles of GCC11 and Python3 than sizes [4, 8] (which is made
> hotter).
>
> Python3 Size = 1 -> 13.64%
> Python3 Size = [4, 8] -> 60.92%
>
> GCC11 Size = 1 -> 1.29%
> GCC11 Size = [4, 8] -> 33.86%
>
> size, align0, align1, ret, New Time/Old Time
> 4, 4, 4, 0, 0.622
> 4, 4, 4, 1, 0.797
> 4, 4, 4, -1, 0.805
> 5, 5, 5, 0, 0.623
> 5, 5, 5, 1, 0.777
> 5, 5, 5, -1, 0.802
> 6, 6, 6, 0, 0.625
> 6, 6, 6, 1, 0.813
> 6, 6, 6, -1, 0.788
> 7, 7, 7, 0, 0.625
> 7, 7, 7, 1, 0.799
> 7, 7, 7, -1, 0.795
> 8, 8, 8, 0, 0.625
> 8, 8, 8, 1, 0.848
> 8, 8, 8, -1, 0.914
> 9, 9, 9, 0, 0.625
>
> Size = 65:
> size, align0, align1, ret, New Time/Old Time
> 65, 0, 0, 0, 1.103
> 65, 0, 0, 1, 1.216
> 65, 0, 0, -1, 1.227
> 65, 65, 0, 0, 1.091
> 65, 0, 65, 1, 1.19
> 65, 65, 65, -1, 1.215
>
> This is because A) the checks in range [65, 96] are now unrolled 2x
> and B) because smaller values <= 16 are now given a hotter path. By
> contrast the SSE4 version has a branch for Size = 80. The unrolled
> version has get better performance for returns which need both
> comparisons.
>
> size, align0, align1, ret, New Time/Old Time
> 128, 4, 8, 0, 0.858
> 128, 4, 8, 1, 0.879
> 128, 4, 8, -1, 0.888
>
> As well, out of microbenchmark environments that are not full
> predictable the branch will have a real-cost.
> ---
> sysdeps/x86_64/multiarch/Makefile | 2 --
> sysdeps/x86_64/multiarch/ifunc-impl-list.c | 4 ----
> sysdeps/x86_64/multiarch/ifunc-memcmp.h | 4 ----
> 3 files changed, 10 deletions(-)
>
Please also remove memcmp-sse4.S.
> diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile
> index b573966966..0400ea332b 100644
> --- a/sysdeps/x86_64/multiarch/Makefile
> +++ b/sysdeps/x86_64/multiarch/Makefile
> @@ -11,7 +11,6 @@ sysdep_routines += \
> memcmp-avx2-movbe-rtm \
> memcmp-evex-movbe \
> memcmp-sse2 \
> - memcmp-sse4 \
> memcmpeq-avx2 \
> memcmpeq-avx2-rtm \
> memcmpeq-evex \
> @@ -164,7 +163,6 @@ sysdep_routines += \
> wmemcmp-avx2-movbe-rtm \
> wmemcmp-evex-movbe \
> wmemcmp-sse2 \
> - wmemcmp-sse4 \
> # sysdep_routines
> endif
>
> diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> index c6008a73ed..a8afcf81bb 100644
> --- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> +++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c
> @@ -96,8 +96,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
> && CPU_FEATURE_USABLE (BMI2)
> && CPU_FEATURE_USABLE (MOVBE)),
> __memcmp_evex_movbe)
> - IFUNC_IMPL_ADD (array, i, memcmp, CPU_FEATURE_USABLE (SSE4_1),
> - __memcmp_sse4_1)
> IFUNC_IMPL_ADD (array, i, memcmp, 1, __memcmp_sse2))
>
> #ifdef SHARED
> @@ -809,8 +807,6 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array,
> && CPU_FEATURE_USABLE (BMI2)
> && CPU_FEATURE_USABLE (MOVBE)),
> __wmemcmp_evex_movbe)
> - IFUNC_IMPL_ADD (array, i, wmemcmp, CPU_FEATURE_USABLE (SSE4_1),
> - __wmemcmp_sse4_1)
> IFUNC_IMPL_ADD (array, i, wmemcmp, 1, __wmemcmp_sse2))
>
> /* Support sysdeps/x86_64/multiarch/wmemset.c. */
> diff --git a/sysdeps/x86_64/multiarch/ifunc-memcmp.h b/sysdeps/x86_64/multiarch/ifunc-memcmp.h
> index 44759a3ad5..c743970fe3 100644
> --- a/sysdeps/x86_64/multiarch/ifunc-memcmp.h
> +++ b/sysdeps/x86_64/multiarch/ifunc-memcmp.h
> @@ -20,7 +20,6 @@
> # include <init-arch.h>
>
> extern __typeof (REDIRECT_NAME) OPTIMIZE (sse2) attribute_hidden;
> -extern __typeof (REDIRECT_NAME) OPTIMIZE (sse4_1) attribute_hidden;
> extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2_movbe) attribute_hidden;
> extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2_movbe_rtm) attribute_hidden;
> extern __typeof (REDIRECT_NAME) OPTIMIZE (evex_movbe) attribute_hidden;
> @@ -46,8 +45,5 @@ IFUNC_SELECTOR (void)
> return OPTIMIZE (avx2_movbe);
> }
>
> - if (CPU_FEATURE_USABLE_P (cpu_features, SSE4_1))
> - return OPTIMIZE (sse4_1);
> -
> return OPTIMIZE (sse2);
> }
> --
> 2.25.1
>
Thanks.
--
H.J.
next prev parent reply other threads:[~2022-04-15 17:21 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-15 5:51 [PATCH v1 1/3] x86: Optimize memcmp SSE2 in memcmp.S Noah Goldstein
2022-04-15 5:51 ` [PATCH v1 2/3] x86: Remove memcmp-sse4.S Noah Goldstein
2022-04-15 17:20 ` H.J. Lu [this message]
2022-04-15 17:29 ` Noah Goldstein
2022-04-15 5:51 ` [PATCH v1 3/3] x86: Cleanup page cross code in memcmp-avx2-movbe.S Noah Goldstein
2022-04-15 17:21 ` H.J. Lu
2022-04-15 17:13 ` [PATCH v1 1/3] x86: Optimize memcmp SSE2 in memcmp.S H.J. Lu
2022-04-15 17:25 ` [PATCH v2 " Noah Goldstein
2022-04-15 17:25 ` [PATCH v2 2/3] x86: Remove memcmp-sse4.S Noah Goldstein
2022-04-15 17:25 ` [PATCH v2 3/3] x86: Cleanup page cross code in memcmp-avx2-movbe.S Noah Goldstein
2022-04-15 17:27 ` [PATCH v3 1/3] x86: Optimize memcmp SSE2 in memcmp.S Noah Goldstein
2022-04-15 17:28 ` [PATCH v3 2/3] x86: Remove memcmp-sse4.S Noah Goldstein
2022-04-15 17:32 ` H.J. Lu
2022-05-12 20:01 ` Sunil Pandey
2022-04-15 17:28 ` [PATCH v3 3/3] x86: Cleanup page cross code in memcmp-avx2-movbe.S Noah Goldstein
2022-04-15 17:33 ` H.J. Lu
2022-05-12 20:03 ` Sunil Pandey
2022-04-15 17:33 ` [PATCH v3 1/3] x86: Optimize memcmp SSE2 in memcmp.S H.J. Lu
2022-04-15 18:44 ` Noah Goldstein
2022-04-19 20:30 ` Joseph Myers
2022-04-19 20:53 ` Noah Goldstein
2022-04-19 22:56 ` Noah Goldstein
2022-04-20 1:19 ` Noah Goldstein
2022-05-12 19:59 ` Sunil Pandey
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAMe9rOrgnfAfCOggWy78epsAzE+uN8PJ+rec9e_fkdJhXBJmRQ@mail.gmail.com \
--to=hjl.tools@gmail.com \
--cc=carlos@systemhalted.org \
--cc=goldstein.w.n@gmail.com \
--cc=libc-alpha@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).