diff options
| author | Adhemerval Zanella <adhemerval.zanella@linaro.org> | 2024-02-08 10:08:39 -0300 |
|---|---|---|
| committer | Arjun Shankar <arjun@redhat.com> | 2024-04-04 12:12:11 +0200 |
| commit | 6484a92698039c4a7a510f0214e22d067b0d78b3 (patch) | |
| tree | 98ec54b00f89ecb05fabafba078d296f238c3894 /sysdeps | |
| parent | aa4249266e9906c4bc833e4847f4d8feef59504f (diff) | |
| download | glibc-6484a92698039c4a7a510f0214e22d067b0d78b3.zip glibc-6484a92698039c4a7a510f0214e22d067b0d78b3.tar.gz glibc-6484a92698039c4a7a510f0214e22d067b0d78b3.tar.bz2 | |
x86: Do not prefer ERMS for memset on Zen3+
For AMD Zen3+ architecture, the performance of the vectorized loop is
slightly better than ERMS.
Checked on x86_64-linux-gnu on Zen3.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
(cherry picked from commit 272708884cb750f12f5c74a00e6620c19dc6d567)
Diffstat (limited to 'sysdeps')
| -rw-r--r-- | sysdeps/x86/dl-cacheinfo.h | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/sysdeps/x86/dl-cacheinfo.h b/sysdeps/x86/dl-cacheinfo.h index f34d128..5a98f70 100644 --- a/sysdeps/x86/dl-cacheinfo.h +++ b/sysdeps/x86/dl-cacheinfo.h @@ -1021,6 +1021,11 @@ dl_init_cacheinfo (struct cpu_features *cpu_features) minimum value is fixed. */ rep_stosb_threshold = TUNABLE_GET (x86_rep_stosb_threshold, long int, NULL); + if (cpu_features->basic.kind == arch_kind_amd + && !TUNABLE_IS_INITIALIZED (x86_rep_stosb_threshold)) + /* For AMD Zen3+ architecture, the performance of the vectorized loop is + slightly better than ERMS. */ + rep_stosb_threshold = SIZE_MAX; TUNABLE_SET_WITH_BOUNDS (x86_data_cache_size, data, 0, SIZE_MAX); TUNABLE_SET_WITH_BOUNDS (x86_shared_cache_size, shared, 0, SIZE_MAX); |