public inbox for glibc-cvs@sourceware.org help / color / mirror / Atom feed
From: Wilco Dijkstra <wilco@sourceware.org> To: glibc-cvs@sourceware.org Subject: [glibc] benchtests: Add random() benchmark Date: Thu, 18 Apr 2024 13:31:02 +0000 (GMT) [thread overview] Message-ID: <20240418133102.380F13858D38@sourceware.org> (raw) https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=0997c3d0c87433ac8c78043aaa9b6b7e91df2882 commit 0997c3d0c87433ac8c78043aaa9b6b7e91df2882 Author: Wilco Dijkstra <wilco.dijkstra@arm.com> Date: Tue Nov 28 13:33:56 2023 +0000 benchtests: Add random() benchmark Add a simple benchmark to measure the overhead of internal libc locks in the random() implementation on both single- and multi-threaded cases. This relies on the implementation of random using internal locks to access shared global data, and that the runtime uses multi-threaded locking once a thread has been created (even after it finishes). Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org> Diff: --- benchtests/Makefile | 1 + benchtests/bench-random-lock.c | 106 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 107 insertions(+) diff --git a/benchtests/Makefile b/benchtests/Makefile index 05b8751c55..7e73b8504e 100644 --- a/benchtests/Makefile +++ b/benchtests/Makefile @@ -257,6 +257,7 @@ hash-benchset := \ stdlib-benchset := \ arc4random \ + random-lock \ strtod \ # stdlib-benchset diff --git a/benchtests/bench-random-lock.c b/benchtests/bench-random-lock.c new file mode 100644 index 0000000000..0fbe918137 --- /dev/null +++ b/benchtests/bench-random-lock.c @@ -0,0 +1,106 @@ +/* Benchmark internal libc locking functions used in random. + Copyright (C) 2022-2024 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + <https://www.gnu.org/licenses/>. */ + +#define TEST_MAIN +#define TEST_NAME "random-lock" +#define TEST_FUNCTION test_main +#include <pthread.h> +#include <stdio.h> +#include <stdlib.h> +#include "bench-timing.h" +#include "json-lib.h" + +/* Modern cores run 20M iterations in about 1 second. */ +#define NUM_ITERS 50000000 + + +/* Measure the overhead of __libc_lock_lock and __libc_lock_unlock by + calling random (). */ +static void +bench_random_lock (json_ctx_t *json_ctx, size_t iters) +{ + timing_t start, stop, total; + + srandom (0); + + /* Warmup to reduce variations due to frequency scaling. */ + for (int i = 0; i < iters / 4; i++) + (void) random (); + + TIMING_NOW (start); + + for (int i = 0; i < iters; i++) + (void) random (); + + TIMING_NOW (stop); + + TIMING_DIFF (total, start, stop); + + json_element_double (json_ctx, (double) total / (double) iters); +} + +static void * +thread_start (void *p) +{ + return p; +} + +int +test_main (void) +{ + json_ctx_t json_ctx; + + json_init (&json_ctx, 0, stdout); + + json_document_begin (&json_ctx); + + json_attr_string (&json_ctx, "timing_type", TIMING_TYPE); + json_attr_object_begin (&json_ctx, "functions"); + json_attr_object_begin (&json_ctx, "random"); + json_attr_string (&json_ctx, "bench-variant", "single-threaded"); + json_array_begin (&json_ctx, "results"); + + /* Run benchmark single threaded. */ + bench_random_lock (&json_ctx, NUM_ITERS); + + json_array_end (&json_ctx); + json_attr_object_end (&json_ctx); + + json_attr_object_begin (&json_ctx, "random"); + json_attr_string (&json_ctx, "bench-variant", "multi-threaded"); + json_array_begin (&json_ctx, "results"); + + /* Start a short thread to force SINGLE_THREAD_P == false. This relies on + the runtime disabling single-threaded optimizations when multiple + threads are used, even after they finish. */ + + pthread_t t; + pthread_create (&t, NULL, thread_start, NULL); + pthread_join (t, NULL); + + /* Repeat benchmark with single-threaded optimizations disabled. */ + bench_random_lock (&json_ctx, NUM_ITERS); + + json_array_end (&json_ctx); + json_attr_object_end (&json_ctx); + json_attr_object_end (&json_ctx); + json_document_end (&json_ctx); + return 0; +} + +#include "support/test-driver.c"
reply other threads:[~2024-04-18 13:31 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20240418133102.380F13858D38@sourceware.org \ --to=wilco@sourceware.org \ --cc=glibc-cvs@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: linkBe 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).