Re: [PATCH v6 03/10] benchtests: Add arc4random benchtest

[Noah Goldstein <goldstein.w.n@gmail.com>]

On Wed, May 18, 2022 at 12:15 PM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> It shows both throughput (total bytes obtained in the test duration)
> and latecy for both arc4random and arc4random_buf with different
> sizes.
>
> Checked on x86_64-linux-gnu, aarch64-linux, and powerpc64le-linux-gnu.
> ---
>  benchtests/Makefile           |   5 +-
>  benchtests/bench-arc4random.c | 224 ++++++++++++++++++++++++++++++++++
>  stdlib/arc4random.c           |  14 +--
>  3 files changed, 235 insertions(+), 8 deletions(-)
>  create mode 100644 benchtests/bench-arc4random.c
>
> diff --git a/benchtests/Makefile b/benchtests/Makefile
> index de9de5cf58..76583d45a3 100644
> --- a/benchtests/Makefile
> +++ b/benchtests/Makefile
> @@ -227,7 +227,10 @@ LOCALES := \
>  include ../gen-locales.mk
>  endif
>
> -stdlib-benchset := strtod
> +stdlib-benchset := \
> +  arc4random \
> +  strtod \
> +  # stdlib-benchset
>
>  stdio-common-benchset := sprintf
>
> diff --git a/benchtests/bench-arc4random.c b/benchtests/bench-arc4random.c
> new file mode 100644
> index 0000000000..626f2ba48c
> --- /dev/null
> +++ b/benchtests/bench-arc4random.c
> @@ -0,0 +1,224 @@
> +/* arc4random benchmarks.
> +   Copyright (C) 2022 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/>.  */
> +
> +#include "bench-timing.h"
> +#include "json-lib.h"
> +#include <array_length.h>
> +#include <intprops.h>
> +#include <signal.h>
> +#include <stdbool.h>
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <support/support.h>
> +#include <support/timespec.h>
> +#include <support/xthread.h>
> +
> +/* Prevent compiler to optimize away call.  */
> +#define DO_NOT_OPTIMIZE(__value)               \
> +  ({                                           \
> +    __typeof (__value) __v = (__value);                \
> +    asm volatile("" : : "r,m" (__v) : "memory");\
> +  })

This might be useful to have in the support header.
bench-hash-func.c has essentially the same code.

> +
> +static volatile sig_atomic_t timer_finished;
> +
> +static void timer_callback (int unused)
> +{
> +  timer_finished = 1;
> +}
> +
> +static timer_t timer;
> +
> +/* Run for approximately DURATION seconds, and it does not matter who
> +   receive the signal (so not need to mask it on main thread).  */
> +static void
> +timer_start (void)
> +{
> +  timer_finished = 0;
> +  timer = support_create_timer (DURATION, 0, false, timer_callback);
> +}
> +static void
> +timer_stop (void)
> +{
> +  support_delete_timer (timer);
> +}
> +
> +static const uint32_t sizes[] = { 0, 16, 32, 48, 64, 80, 96, 112, 128 };
> +
> +static double
> +bench_throughput (void)
> +{
> +  uint64_t n = 0;
> +
> +  struct timespec start, end;
> +  clock_gettime (CLOCK_MONOTONIC, &start);
> +  while (1)
> +    {
> +      DO_NOT_OPTIMIZE (arc4random ());
> +      n++;
> +
> +      if (timer_finished == 1)
> +       break;
> +    }
> +  clock_gettime (CLOCK_MONOTONIC, &end);
> +  struct timespec diff = timespec_sub (end, start);
> +
> +  double total = (double) n * sizeof (uint32_t);
> +  double duration = (double) diff.tv_sec
> +    + (double) diff.tv_nsec / TIMESPEC_HZ;
> +
> +  return total / duration;
> +}
> +
> +static double
> +bench_latency (void)
> +{
> +  timing_t start, stop, cur;
> +  const size_t iters = 1024;
> +
> +  TIMING_NOW (start);
> +  for (size_t i = 0; i < iters; i++)
> +    DO_NOT_OPTIMIZE (arc4random ());
> +  TIMING_NOW (stop);
> +
> +  TIMING_DIFF (cur, start, stop);
> +
> +  return (double) (cur) / (double) iters;
> +}
> +
> +static double
> +bench_buf_throughput (size_t len)
> +{
> +  uint8_t buf[len];
> +  uint64_t n = 0;
> +
> +  struct timespec start, end;
> +  clock_gettime (CLOCK_MONOTONIC, &start);
> +  while (1)
> +    {
> +      arc4random_buf (buf, len);
> +      n++;
> +
> +      if (timer_finished == 1)
> +       break;
> +    }
> +  clock_gettime (CLOCK_MONOTONIC, &end);
> +  struct timespec diff = timespec_sub (end, start);
> +
> +  double total = (double) n * len;
> +  double duration = (double) diff.tv_sec
> +    + (double) diff.tv_nsec / TIMESPEC_HZ;
> +
> +  return total / duration;
> +}
> +
> +static double
> +bench_buf_latency (size_t len)
> +{
> +  timing_t start, stop, cur;
> +  const size_t iters = 1024;
> +
> +  uint8_t buf[len];
> +
> +  TIMING_NOW (start);
> +  for (size_t i = 0; i < iters; i++)
> +    arc4random_buf (buf, len);
> +  TIMING_NOW (stop);
> +
> +  TIMING_DIFF (cur, start, stop);
> +
> +  return (double) (cur) / (double) iters;
> +}
> +
> +static void
> +bench_singlethread (json_ctx_t *json_ctx)
> +{
> +  json_element_object_begin (json_ctx);
> +
> +  json_array_begin (json_ctx, "throughput");
> +  for (int i = 0; i < array_length (sizes); i++)
> +    {
> +      timer_start ();
> +      double r = sizes[i] == 0
> +       ? bench_throughput () : bench_buf_throughput (sizes[i]);
> +      timer_stop ();
> +
> +      json_element_double (json_ctx, r);
> +    }
> +  json_array_end (json_ctx);
> +
> +  json_array_begin (json_ctx, "latency");
> +  for (int i = 0; i < array_length (sizes); i++)
> +    {
> +      timer_start ();
> +      double r = sizes[i] == 0
> +       ? bench_latency () : bench_buf_latency (sizes[i]);
> +      timer_stop ();
> +
> +      json_element_double (json_ctx, r);
> +    }
> +  json_array_end (json_ctx);
> +
> +  json_element_object_end (json_ctx);
> +}
> +
> +static void
> +run_bench (json_ctx_t *json_ctx, const char *name,
> +          char *const*fnames, size_t fnameslen,
> +          void (*bench)(json_ctx_t *ctx))
> +{
> +  json_attr_object_begin (json_ctx, name);
> +  json_array_begin (json_ctx, "functions");
> +  for (int i = 0; i < fnameslen; i++)
> +    json_element_string (json_ctx, fnames[i]);
> +  json_array_end (json_ctx);
> +
> +  json_array_begin (json_ctx, "results");
> +  bench (json_ctx);
> +  json_array_end (json_ctx);
> +  json_attr_object_end (json_ctx);
> +}
> +
> +static int
> +do_test (void)
> +{
> +  char *fnames[array_length (sizes)];
> +  for (int i = 0; i < array_length (sizes); i++)
> +    if (sizes[i] == 0)
> +      fnames[i] = xasprintf ("arc4random");
> +    else
> +      fnames[i] = xasprintf ("arc4random_buf(%u)", sizes[i]);
> +
> +  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);
> +
> +  run_bench (&json_ctx, "single-thread", fnames, array_length (fnames),
> +            bench_singlethread);
> +
> +  json_document_end (&json_ctx);
> +
> +  for (int i = 0; i < array_length (sizes); i++)
> +    free (fnames[i]);
> +
> +  return 0;
> +}
> +
> +#include <support/test-driver.c>
> diff --git a/stdlib/arc4random.c b/stdlib/arc4random.c
> index f6553dfd7d..74c917b9b9 100644
> --- a/stdlib/arc4random.c
> +++ b/stdlib/arc4random.c
> @@ -25,9 +25,9 @@
>  #include <sys/param.h>
>  #include <sys/random.h>
>
> -/* Besides the cipher state 'ctx', it keeps two counters: 'have' is the
> -   current valid bytes not yet consumed in 'buf', while 'count' is the maximum
> -   number of bytes until a reseed.
> +/* arc4random keeps two counters: 'have' is the current valid bytes not yet
> +   consumed in 'buf' while 'count' is the maximum number of bytes until a
> +   reseed.
>
>     Both the initial seed and reseed try to obtain entropy from the kernel
>     and abort the process if none could be obtained.
> @@ -81,10 +81,10 @@ arc4random_getrandom_failure (void)
>    __libc_fatal ("Fatal glibc error: Cannot get entropy for arc4random\n");
>  }
>
> -/* Fork detection is done by checking if MADV_WIPEONFORK supported.  If not
> -   the fork callback will reset the state on the fork call.  It does not
> -   handle direct clone calls, nor vfork or _Fork (arc4random is not
> -   async-signal-safe due the internal lock usage).  */
> +/* If the kernel supports MADV_WIPEONFORK it is used to provide fork
> +   detection. Otherwise, the state is resetied with an atfork handler. The
> +   latter does not handle direct clone calls, nor vfork or _Fork (arc4random
> +   is not async-signal-safe due to the internal lock usage).  */
>  static void
>  arc4random_init (uint8_t *buf)
>  {
> --
> 2.34.1
>