From: Noah Goldstein <goldstein.w.n@gmail.com>
To: "H.J. Lu" <hjl.tools@gmail.com>
Cc: libc-alpha@sourceware.org, carlos@systemhalted.org
Subject: Re: [PATCH v1 7/7] Bench: Improve benchtests for memchr, strchr, strnlen, strrchr
Date: Tue, 18 Oct 2022 15:58:33 -0700 [thread overview]
Message-ID: <CAFUsyfJDMGrXXRUqVWQUBJANPL=WvVKPKZe3+E_U_2NuyujcBQ@mail.gmail.com> (raw)
In-Reply-To: <CAMe9rOpKvtHhqM776O0rFgzFTB2RkOmPXHehKwboi=0NV2CZMQ@mail.gmail.com>
On Tue, Oct 18, 2022 at 2:54 PM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Tue, Oct 18, 2022 at 2:05 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> >
> > On Tue, Oct 18, 2022 at 4:01 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> > >
> > > On Mon, Oct 17, 2022 at 7:49 PM Noah Goldstein <goldstein.w.n@gmail.com> wrote:
> > > >
> > > > 1. Add more complete coverage in the medium size range.
> > > > 2. In strnlen remove the `1 << i` which was UB (`i` could go beyond
> > > > 32/64)
> > > > 3. Add timer for total benchmark runtime (useful for deciding about
> > > > tradeoff between coverage and runtime).
> > >
> > > So this is only used for total runtime and won't be used for performance
> > > comparison. Will "time ./bench" be sufficient?
> >
> > Yes but when running something like:
> > make bench BENCHSET="string-benchset"
> > its hard to get the time of an individual test. We also do timeout on
> > an individual test basis so seems reasonable to get total runtime
> > on individual test basis.
>
> Can you make it a separate patch to add total time? One can tell
> which bench takes too much time when running
>
> make bench BENCHSET="string-benchset"
>
> and use "time ./bench" on that bench.
Its not that its not doable, this is just something I found
convenient when adding benchtests and figured others
might as well.
I'll remove it from this patchset for now.
Once this stuff is done i can resubmit and we can weigh
the pros/cons.
>
> > >
> > > > ---
> > > > benchtests/bench-memchr.c | 83 +++++++++++++++++++++++++-----------
> > > > benchtests/bench-rawmemchr.c | 36 ++++++++++++++--
> > > > benchtests/bench-strchr.c | 42 +++++++++++++-----
> > > > benchtests/bench-strnlen.c | 19 ++++++---
> > > > benchtests/bench-strrchr.c | 33 +++++++++++++-
> > > > 5 files changed, 166 insertions(+), 47 deletions(-)
> > > >
> > > > diff --git a/benchtests/bench-memchr.c b/benchtests/bench-memchr.c
> > > > index 0facda2fa0..c4d758ae61 100644
> > > > --- a/benchtests/bench-memchr.c
> > > > +++ b/benchtests/bench-memchr.c
> > > > @@ -126,9 +126,10 @@ do_test (json_ctx_t *json_ctx, size_t align, size_t pos, size_t len,
> > > > int
> > > > test_main (void)
> > > > {
> > > > - size_t i;
> > > > + size_t i, j, al, al_max;
> > > > int repeats;
> > > > json_ctx_t json_ctx;
> > > > + timing_t bench_start, bench_stop, bench_total_time;
> > > > test_init ();
> > > >
> > > > json_init (&json_ctx, 0, stdout);
> > > > @@ -147,35 +148,47 @@ test_main (void)
> > > >
> > > > json_array_begin (&json_ctx, "results");
> > > >
> > > > + TIMING_NOW (bench_start);
> > > > + al_max = 0;
> > > > +#ifdef USE_AS_MEMRCHR
> > > > + al_max = getpagesize () / 2;
> > > > +#endif
> > > > +
> > > > for (repeats = 0; repeats < 2; ++repeats)
> > > > {
> > > > - for (i = 1; i < 8; ++i)
> > > > + for (al = 0; al <= al_max; al += getpagesize () / 2)
> > > > {
> > > > - do_test (&json_ctx, 0, 16 << i, 2048, 23, repeats);
> > > > - do_test (&json_ctx, i, 64, 256, 23, repeats);
> > > > - do_test (&json_ctx, 0, 16 << i, 2048, 0, repeats);
> > > > - do_test (&json_ctx, i, 64, 256, 0, repeats);
> > > > -
> > > > - do_test (&json_ctx, getpagesize () - 15, 64, 256, 0, repeats);
> > > > + for (i = 1; i < 8; ++i)
> > > > + {
> > > > + do_test (&json_ctx, al, 16 << i, 2048, 23, repeats);
> > > > + do_test (&json_ctx, al + i, 64, 256, 23, repeats);
> > > > + do_test (&json_ctx, al, 16 << i, 2048, 0, repeats);
> > > > + do_test (&json_ctx, al + i, 64, 256, 0, repeats);
> > > > +
> > > > + do_test (&json_ctx, al + getpagesize () - 15, 64, 256, 0,
> > > > + repeats);
> > > > #ifdef USE_AS_MEMRCHR
> > > > - /* Also test the position close to the beginning for memrchr. */
> > > > - do_test (&json_ctx, 0, i, 256, 23, repeats);
> > > > - do_test (&json_ctx, 0, i, 256, 0, repeats);
> > > > - do_test (&json_ctx, i, i, 256, 23, repeats);
> > > > - do_test (&json_ctx, i, i, 256, 0, repeats);
> > > > + /* Also test the position close to the beginning for memrchr. */
> > > > + do_test (&json_ctx, al, i, 256, 23, repeats);
> > > > + do_test (&json_ctx, al, i, 256, 0, repeats);
> > > > + do_test (&json_ctx, al + i, i, 256, 23, repeats);
> > > > + do_test (&json_ctx, al + i, i, 256, 0, repeats);
> > > > #endif
> > > > + }
> > > > + for (i = 1; i < 8; ++i)
> > > > + {
> > > > + do_test (&json_ctx, al + i, i << 5, 192, 23, repeats);
> > > > + do_test (&json_ctx, al + i, i << 5, 192, 0, repeats);
> > > > + do_test (&json_ctx, al + i, i << 5, 256, 23, repeats);
> > > > + do_test (&json_ctx, al + i, i << 5, 256, 0, repeats);
> > > > + do_test (&json_ctx, al + i, i << 5, 512, 23, repeats);
> > > > + do_test (&json_ctx, al + i, i << 5, 512, 0, repeats);
> > > > +
> > > > + do_test (&json_ctx, al + getpagesize () - 15, i << 5, 256, 23,
> > > > + repeats);
> > > > + }
> > > > }
> > > > - for (i = 1; i < 8; ++i)
> > > > - {
> > > > - do_test (&json_ctx, i, i << 5, 192, 23, repeats);
> > > > - do_test (&json_ctx, i, i << 5, 192, 0, repeats);
> > > > - do_test (&json_ctx, i, i << 5, 256, 23, repeats);
> > > > - do_test (&json_ctx, i, i << 5, 256, 0, repeats);
> > > > - do_test (&json_ctx, i, i << 5, 512, 23, repeats);
> > > > - do_test (&json_ctx, i, i << 5, 512, 0, repeats);
> > > > -
> > > > - do_test (&json_ctx, getpagesize () - 15, i << 5, 256, 23, repeats);
> > > > - }
> > > > +
> > > > for (i = 1; i < 32; ++i)
> > > > {
> > > > do_test (&json_ctx, 0, i, i + 1, 23, repeats);
> > > > @@ -207,11 +220,33 @@ test_main (void)
> > > > do_test (&json_ctx, 0, 2, i + 1, 0, repeats);
> > > > #endif
> > > > }
> > > > + for (al = 0; al <= al_max; al += getpagesize () / 2)
> > > > + {
> > > > + for (i = (16 / sizeof (CHAR)); i <= (8192 / sizeof (CHAR)); i += i)
> > > > + {
> > > > + for (j = 0; j <= (384 / sizeof (CHAR));
> > > > + j += (32 / sizeof (CHAR)))
> > > > + {
> > > > + do_test (&json_ctx, al, i + j, i, 23, repeats);
> > > > + do_test (&json_ctx, al, i, i + j, 23, repeats);
> > > > + if (j < i)
> > > > + {
> > > > + do_test (&json_ctx, al, i - j, i, 23, repeats);
> > > > + do_test (&json_ctx, al, i, i - j, 23, repeats);
> > > > + }
> > > > + }
> > > > + }
> > > > + }
> > > > +
> > > > #ifndef USE_AS_MEMRCHR
> > > > break;
> > > > #endif
> > > > }
> > > >
> > > > + TIMING_NOW (bench_stop);
> > > > + TIMING_DIFF (bench_total_time, bench_start, bench_stop);
> > > > + json_attr_double (&json_ctx, "benchtime", bench_total_time);
> > > > +
> > > > json_array_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > diff --git a/benchtests/bench-rawmemchr.c b/benchtests/bench-rawmemchr.c
> > > > index b1803afc14..667ecd48f9 100644
> > > > --- a/benchtests/bench-rawmemchr.c
> > > > +++ b/benchtests/bench-rawmemchr.c
> > > > @@ -70,7 +70,7 @@ do_test (json_ctx_t *json_ctx, size_t align, size_t pos, size_t len, int seek_ch
> > > > size_t i;
> > > > char *result;
> > > >
> > > > - align &= 7;
> > > > + align &= getpagesize () - 1;
> > > > if (align + len >= page_size)
> > > > return;
> > > >
> > > > @@ -106,7 +106,7 @@ test_main (void)
> > > > {
> > > > json_ctx_t json_ctx;
> > > > size_t i;
> > > > -
> > > > + timing_t bench_start, bench_stop, bench_total_time;
> > > > test_init ();
> > > >
> > > > json_init (&json_ctx, 0, stdout);
> > > > @@ -120,11 +120,12 @@ test_main (void)
> > > >
> > > > json_array_begin (&json_ctx, "ifuncs");
> > > > FOR_EACH_IMPL (impl, 0)
> > > > - json_element_string (&json_ctx, impl->name);
> > > > + json_element_string (&json_ctx, impl->name);
> > > > json_array_end (&json_ctx);
> > > >
> > > > json_array_begin (&json_ctx, "results");
> > > >
> > > > + TIMING_NOW (bench_start);
> > > > for (i = 1; i < 7; ++i)
> > > > {
> > > > do_test (&json_ctx, 0, 16 << i, 2048, 23);
> > > > @@ -137,6 +138,35 @@ test_main (void)
> > > > do_test (&json_ctx, 0, i, i + 1, 23);
> > > > do_test (&json_ctx, 0, i, i + 1, 0);
> > > > }
> > > > + for (; i < 256; i += 32)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + 1, 23);
> > > > + do_test (&json_ctx, 0, i - 1, i, 23);
> > > > + }
> > > > + for (; i < 512; i += 64)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + 1, 23);
> > > > + do_test (&json_ctx, 0, i - 1, i, 23);
> > > > + }
> > > > + for (; i < 1024; i += 128)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + 1, 23);
> > > > + do_test (&json_ctx, 0, i - 1, i, 23);
> > > > + }
> > > > + for (; i < 2048; i += 256)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + 1, 23);
> > > > + do_test (&json_ctx, 0, i - 1, i, 23);
> > > > + }
> > > > + for (; i < 4096; i += 512)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + 1, 23);
> > > > + do_test (&json_ctx, 0, i - 1, i, 23);
> > > > + }
> > > > +
> > > > + TIMING_NOW (bench_stop);
> > > > + TIMING_DIFF (bench_total_time, bench_start, bench_stop);
> > > > + json_attr_double (&json_ctx, "benchtime", bench_total_time);
> > > >
> > > > json_array_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > diff --git a/benchtests/bench-strchr.c b/benchtests/bench-strchr.c
> > > > index 54640bde7e..af325806ce 100644
> > > > --- a/benchtests/bench-strchr.c
> > > > +++ b/benchtests/bench-strchr.c
> > > > @@ -287,8 +287,8 @@ int
> > > > test_main (void)
> > > > {
> > > > json_ctx_t json_ctx;
> > > > - size_t i;
> > > > -
> > > > + size_t i, j;
> > > > + timing_t bench_start, bench_stop, bench_total_time;
> > > > test_init ();
> > > >
> > > > json_init (&json_ctx, 0, stdout);
> > > > @@ -307,6 +307,7 @@ test_main (void)
> > > >
> > > > json_array_begin (&json_ctx, "results");
> > > >
> > > > + TIMING_NOW (bench_start);
> > > > for (i = 1; i < 8; ++i)
> > > > {
> > > > do_test (&json_ctx, 0, 16 << i, 2048, SMALL_CHAR, MIDDLE_CHAR);
> > > > @@ -367,15 +368,34 @@ test_main (void)
> > > > do_test (&json_ctx, 0, i, i + 1, 0, BIG_CHAR);
> > > > }
> > > >
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.0);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.1);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.25);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.33);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.5);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.66);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.75);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 0.9);
> > > > - DO_RAND_TEST(&json_ctx, 0, 15, 16, 1.0);
> > > > + for (i = 16 / sizeof (CHAR); i <= 8192 / sizeof (CHAR); i += i)
> > > > + {
> > > > + for (j = 32 / sizeof (CHAR); j <= 320 / sizeof (CHAR);
> > > > + j += 32 / sizeof (CHAR))
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i + j, 0, MIDDLE_CHAR);
> > > > + do_test (&json_ctx, 0, i + j, i, 0, MIDDLE_CHAR);
> > > > + if (i > j)
> > > > + {
> > > > + do_test (&json_ctx, 0, i, i - j, 0, MIDDLE_CHAR);
> > > > + do_test (&json_ctx, 0, i - j, i, 0, MIDDLE_CHAR);
> > > > + }
> > > > + }
> > > > + }
> > > > +
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.0);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.1);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.25);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.33);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.5);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.66);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.75);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 0.9);
> > > > + DO_RAND_TEST (&json_ctx, 0, 15, 16, 1.0);
> > > > +
> > > > + TIMING_NOW (bench_stop);
> > > > + TIMING_DIFF (bench_total_time, bench_start, bench_stop);
> > > > + json_attr_double (&json_ctx, "benchtime", bench_total_time);
> > > >
> > > > json_array_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > diff --git a/benchtests/bench-strnlen.c b/benchtests/bench-strnlen.c
> > > > index 13b46b3f57..c6281b6373 100644
> > > > --- a/benchtests/bench-strnlen.c
> > > > +++ b/benchtests/bench-strnlen.c
> > > > @@ -117,7 +117,7 @@ test_main (void)
> > > > {
> > > > size_t i, j;
> > > > json_ctx_t json_ctx;
> > > > -
> > > > + timing_t bench_start, bench_stop, bench_total_time;
> > > > test_init ();
> > > >
> > > > json_init (&json_ctx, 0, stdout);
> > > > @@ -136,6 +136,7 @@ test_main (void)
> > > >
> > > > json_array_begin (&json_ctx, "results");
> > > >
> > > > + TIMING_NOW (bench_start);
> > > > for (i = 0; i <= 1; ++i)
> > > > {
> > > > do_test (&json_ctx, i, 1, 128, MIDDLE_CHAR);
> > > > @@ -195,23 +196,27 @@ test_main (void)
> > > > {
> > > > for (j = 0; j <= (704 / sizeof (CHAR)); j += (32 / sizeof (CHAR)))
> > > > {
> > > > - do_test (&json_ctx, 0, 1 << i, (i + j), BIG_CHAR);
> > > > do_test (&json_ctx, 0, i + j, i, BIG_CHAR);
> > > > -
> > > > - do_test (&json_ctx, 64, 1 << i, (i + j), BIG_CHAR);
> > > > do_test (&json_ctx, 64, i + j, i, BIG_CHAR);
> > > >
> > > > + do_test (&json_ctx, 0, i, i + j, BIG_CHAR);
> > > > + do_test (&json_ctx, 64, i, i + j, BIG_CHAR);
> > > > +
> > > > if (j < i)
> > > > {
> > > > - do_test (&json_ctx, 0, 1 << i, i - j, BIG_CHAR);
> > > > do_test (&json_ctx, 0, i - j, i, BIG_CHAR);
> > > > -
> > > > - do_test (&json_ctx, 64, 1 << i, i - j, BIG_CHAR);
> > > > do_test (&json_ctx, 64, i - j, i, BIG_CHAR);
> > > > +
> > > > + do_test (&json_ctx, 0, i, i - j, BIG_CHAR);
> > > > + do_test (&json_ctx, 64, i, i - j, BIG_CHAR);
> > > > }
> > > > }
> > > > }
> > > >
> > > > + TIMING_NOW (bench_stop);
> > > > + TIMING_DIFF (bench_total_time, bench_start, bench_stop);
> > > > + json_attr_double (&json_ctx, "benchtime", bench_total_time);
> > > > +
> > > > json_array_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > diff --git a/benchtests/bench-strrchr.c b/benchtests/bench-strrchr.c
> > > > index 7cd2a15484..e6d8163047 100644
> > > > --- a/benchtests/bench-strrchr.c
> > > > +++ b/benchtests/bench-strrchr.c
> > > > @@ -151,8 +151,9 @@ int
> > > > test_main (void)
> > > > {
> > > > json_ctx_t json_ctx;
> > > > - size_t i, j;
> > > > + size_t i, j, k;
> > > > int seek;
> > > > + timing_t bench_start, bench_stop, bench_total_time;
> > > >
> > > > test_init ();
> > > > json_init (&json_ctx, 0, stdout);
> > > > @@ -171,9 +172,10 @@ test_main (void)
> > > >
> > > > json_array_begin (&json_ctx, "results");
> > > >
> > > > + TIMING_NOW (bench_start);
> > > > for (seek = 0; seek <= 23; seek += 23)
> > > > {
> > > > - for (j = 1; j < 32; j += j)
> > > > + for (j = 1; j <= 256; j = (j * 4))
> > > > {
> > > > for (i = 1; i < 9; ++i)
> > > > {
> > > > @@ -197,12 +199,39 @@ test_main (void)
> > > > do_test (&json_ctx, getpagesize () - i / 2 - 1, i, i + 1, seek,
> > > > SMALL_CHAR, j);
> > > > }
> > > > +
> > > > + for (i = (16 / sizeof (CHAR)); i <= (288 / sizeof (CHAR)); i += 32)
> > > > + {
> > > > + do_test (&json_ctx, 0, i - 16, i, seek, SMALL_CHAR, j);
> > > > + do_test (&json_ctx, 0, i, i + 16, seek, SMALL_CHAR, j);
> > > > + }
> > > > +
> > > > + for (i = (16 / sizeof (CHAR)); i <= (2048 / sizeof (CHAR)); i += i)
> > > > + {
> > > > + for (k = 0; k <= (288 / sizeof (CHAR));
> > > > + k += (48 / sizeof (CHAR)))
> > > > + {
> > > > + do_test (&json_ctx, 0, k, i, seek, SMALL_CHAR, j);
> > > > + do_test (&json_ctx, 0, i, i + k, seek, SMALL_CHAR, j);
> > > > +
> > > > + if (k < i)
> > > > + {
> > > > + do_test (&json_ctx, 0, i - k, i, seek, SMALL_CHAR, j);
> > > > + do_test (&json_ctx, 0, k, i - k, seek, SMALL_CHAR, j);
> > > > + do_test (&json_ctx, 0, i, i - k, seek, SMALL_CHAR, j);
> > > > + }
> > > > + }
> > > > + }
> > > > +
> > > > if (seek == 0)
> > > > {
> > > > break;
> > > > }
> > > > }
> > > > }
> > > > + TIMING_NOW (bench_stop);
> > > > + TIMING_DIFF (bench_total_time, bench_start, bench_stop);
> > > > + json_attr_double (&json_ctx, "benchtime", bench_total_time);
> > > >
> > > > json_array_end (&json_ctx);
> > > > json_attr_object_end (&json_ctx);
> > > > --
> > > > 2.34.1
> > > >
> > >
> > >
> > > --
> > > H.J.
>
>
>
> --
> H.J.
next prev parent reply other threads:[~2022-10-18 22:58 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-18 2:48 [PATCH v1 1/7] x86: Optimize memchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-18 2:48 ` [PATCH v1 2/7] x86: Shrink / minorly optimize strchr-evex " Noah Goldstein
2022-10-18 2:51 ` Noah Goldstein
2022-10-18 2:48 ` [PATCH v1 3/7] x86: Optimize strnlen-evex.S " Noah Goldstein
2022-10-18 2:51 ` Noah Goldstein
2022-10-18 2:48 ` [PATCH v1 4/7] x86: Optimize memrchr-evex.S Noah Goldstein
2022-10-18 2:51 ` Noah Goldstein
2022-10-18 2:48 ` [PATCH v1 5/7] x86: Optimize strrchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-18 2:52 ` Noah Goldstein
2022-10-18 2:49 ` [PATCH v1 6/7] x86: Add support for VEC_SIZE == 64 in strcmp-evex.S impl Noah Goldstein
2022-10-20 2:15 ` [PATCH v4] " Noah Goldstein
2022-10-20 3:46 ` H.J. Lu
2022-10-18 2:49 ` [PATCH v1 7/7] Bench: Improve benchtests for memchr, strchr, strnlen, strrchr Noah Goldstein
2022-10-18 21:00 ` H.J. Lu
2022-10-18 21:05 ` Noah Goldstein
2022-10-18 21:53 ` H.J. Lu
2022-10-18 22:58 ` Noah Goldstein [this message]
2022-10-18 2:50 ` [PATCH v1 1/7] x86: Optimize memchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 " Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 2/7] x86: Shrink / minorly optimize strchr-evex " Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 3/7] x86: Optimize strnlen-evex.S " Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 4/7] x86: Optimize memrchr-evex.S Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 5/7] x86: Optimize strrchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 6/7] x86: Add support for VEC_SIZE == 64 in strcmp-evex.S impl Noah Goldstein
2022-10-18 23:19 ` [PATCH v2 7/7] Bench: Improve benchtests for memchr, strchr, strnlen, strrchr Noah Goldstein
2022-10-19 0:01 ` H.J. Lu
2022-10-19 0:44 ` Noah Goldstein
2022-10-19 0:44 ` [PATCH v3 1/7] x86: Optimize memchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-19 0:44 ` [PATCH v3 2/7] x86: Shrink / minorly optimize strchr-evex " Noah Goldstein
2022-10-19 16:53 ` H.J. Lu
2022-10-19 0:44 ` [PATCH v3 3/7] x86: Optimize strnlen-evex.S " Noah Goldstein
2022-10-19 16:57 ` H.J. Lu
2022-10-19 0:44 ` [PATCH v3 4/7] x86: Optimize memrchr-evex.S Noah Goldstein
2022-10-19 16:58 ` H.J. Lu
2022-10-19 0:44 ` [PATCH v3 5/7] x86: Optimize strrchr-evex.S and implement with VMM headers Noah Goldstein
2022-10-19 16:58 ` H.J. Lu
2022-10-19 0:44 ` [PATCH v3 6/7] x86: Add support for VEC_SIZE == 64 in strcmp-evex.S impl Noah Goldstein
2022-10-19 16:59 ` H.J. Lu
2022-10-19 0:44 ` [PATCH v3 7/7] Bench: Improve benchtests for memchr, strchr, strnlen, strrchr Noah Goldstein
2022-10-19 17:00 ` H.J. Lu
2022-10-19 16:52 ` [PATCH v3 1/7] x86: Optimize memchr-evex.S and implement with VMM headers H.J. Lu
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='CAFUsyfJDMGrXXRUqVWQUBJANPL=WvVKPKZe3+E_U_2NuyujcBQ@mail.gmail.com' \
--to=goldstein.w.n@gmail.com \
--cc=carlos@systemhalted.org \
--cc=hjl.tools@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).