From: Con Kolivas <kernel@kolivas.org>
To: Ingo Molnar <mingo@elte.hu>, ck list <ck@vds.kolivas.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>,
Andrew Morton <akpm@linux-foundation.org>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: swap-prefetch: 2.6.22 -mm merge plans
Date: Sat, 5 May 2007 18:42:31 +1000 [thread overview]
Message-ID: <200705051842.32328.kernel@kolivas.org> (raw)
In-Reply-To: <200705042210.15953.kernel@kolivas.org>
[-- Attachment #1: Type: text/plain, Size: 1785 bytes --]
On Friday 04 May 2007 22:10, Con Kolivas wrote:
> On Friday 04 May 2007 18:52, Ingo Molnar wrote:
> > agreed. Con, IIRC you wrote a testcase for this, right? Could you please
> > send us the results of that testing?
>
> Yes, sorry it's a crappy test app but works on 32bit. Timed with prefetch
> disabled and then enabled swap prefetch saves ~5 seconds on average
> hardware on this one test case. I had many users try this and the results
> were between 2 and 10 seconds, but always showed a saving on this testcase.
> This effect easily occurs on printing a big picture, editing a large file,
> compressing an iso image or whatever in real world workloads. Smaller, but
> much more frequent effects of this over the course of a day obviously also
> occur and do add up.
Here's a better swap prefetch tester. Instructions in file.
Machine with 2GB ram and 2GB swapfile
Prefetch disabled:
./sp_tester
Ram 2060352000 Swap 1973420000
Total ram to be malloced: 3047062000 bytes
Starting first malloc of 1523531000 bytes
Starting 1st read of first malloc
Touching this much ram takes 809 milliseconds
Starting second malloc of 1523531000 bytes
Completed second malloc and free
Sleeping for 600 seconds
Important part - starting reread of first malloc
Completed read of first malloc
Timed portion 53397 milliseconds
Enabled:
./sp_tester
Ram 2060352000 Swap 1973420000
Total ram to be malloced: 3047062000 bytes
Starting first malloc of 1523531000 bytes
Starting 1st read of first malloc
Touching this much ram takes 676 milliseconds
Starting second malloc of 1523531000 bytes
Completed second malloc and free
Sleeping for 600 seconds
Important part - starting reread of first malloc
Completed read of first malloc
Timed portion 26351 milliseconds
Note huge time difference.
--
-ck
[-- Attachment #2: sp_tester.c --]
[-- Type: text/x-csrc, Size: 2890 bytes --]
/*
sp_tester.c
Build with:
gcc -o sp_tester sp_tester.c -lrt -W -Wall -O2
How to use:
echo 1 > /proc/sys/vm/overcommit_memory
swapoff -a
swapon -a
./sp_tester
then repeat with changed conditions eg
echo 0 > /proc/sys/vm/swap_prefetch
Each Test takes 10 minutes
*/
#include <stdio.h>
#include <stdarg.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/mman.h>
#include <time.h>
void fatal(const char *format, ...)
{
va_list ap;
if (format) {
va_start(ap, format);
vfprintf(stderr, format, ap);
va_end(ap);
}
fprintf(stderr, "Fatal error - exiting\n");
exit(1);
}
unsigned long ramsize, swapsize;
size_t get_ram(void)
{
FILE *meminfo;
char aux[256];
if(!(meminfo = fopen("/proc/meminfo", "r")))
fatal("fopen\n");
while( !feof(meminfo) && !fscanf(meminfo, "MemTotal: %lu kB", &ramsize) )
fgets(aux,sizeof(aux),meminfo);
while( !feof(meminfo) && !fscanf(meminfo, "SwapTotal: %lu kB", &swapsize) )
fgets(aux,sizeof(aux),meminfo);
if (fclose(meminfo) == -1)
fatal("fclose");
ramsize *= 1000;
swapsize *= 1000;
printf("Ram %lu Swap %lu\n", ramsize, swapsize);
return ramsize + (swapsize / 2);
}
unsigned long get_usecs(struct timespec *myts)
{
if (clock_gettime(CLOCK_REALTIME, myts))
fatal("clock_gettime");
return (myts->tv_sec * 1000000 + myts->tv_nsec / 1000 );
}
int main(void)
{
unsigned long current_time, time_diff;
struct timespec myts;
char *buf1, *buf2, *buf3, *buf4;
size_t size = get_ram();
int sleep_seconds = 600;
if (size > ramsize / 2 * 3)
size = ramsize / 2 * 3;
printf("Total ram to be malloced: %u bytes\n", size);
size /= 2;
printf("Starting first malloc of %u bytes\n", size);
buf1 = malloc(size);
buf4 = malloc(1);
if (buf1 == (char *)-1)
fatal("Failed to malloc 1st buffer\n");
memset(buf1, 0, size);
time_diff = current_time = get_usecs(&myts);
for (buf3 = buf1; buf3 < buf1 + size; buf3++)
*buf4 = *buf3;
printf("Starting 1st read of first malloc\n");
current_time = get_usecs(&myts);
time_diff = current_time - time_diff;
printf("Touching this much ram takes %lu milliseconds\n",time_diff / 1000);
printf("Starting second malloc of %u bytes\n", size);
buf2 = malloc(size);
if (buf2 == (char *)-1)
fatal("Failed to malloc 2nd buffer\n");
memset(buf2, 0, size);
for (buf3 = buf2 + size; buf3 > buf2; buf3--)
*buf4 = *buf3;
free(buf2);
printf("Completed second malloc and free\n");
printf("Sleeping for %u seconds\n", sleep_seconds);
sleep(sleep_seconds);
printf("Important part - starting reread of first malloc\n");
time_diff = current_time = get_usecs(&myts);
for (buf3 = buf1; buf3 < buf1 + size; buf3++)
*buf4 = *buf3;
current_time = get_usecs(&myts);
time_diff = current_time - time_diff;
printf("Completed read of first malloc\n");
printf("Timed portion %lu milliseconds\n",time_diff / 1000);
free(buf1);
free(buf4);
return 0;
}
next prev parent reply other threads:[~2007-05-05 8:42 UTC|newest]
Thread overview: 180+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-04-30 23:20 Andrew Morton
2007-04-30 23:59 ` Bill Irwin
2007-05-01 0:09 ` nfsd/md patches " Neil Brown
2007-05-01 9:08 ` Christoph Hellwig
2007-05-01 9:15 ` Andrew Morton
2007-05-01 9:21 ` Christoph Hellwig
2007-05-01 9:52 ` Neil Brown
2007-05-01 10:15 ` Christoph Hellwig
2007-05-01 14:34 ` Trond Myklebust
2007-05-01 0:54 ` MADV_FREE functionality Rik van Riel
2007-05-01 1:18 ` Andrew Morton
2007-05-01 1:23 ` Rik van Riel
2007-05-01 7:13 ` Jakub Jelinek
2007-05-01 1:23 ` Ulrich Drepper
2007-05-01 1:39 ` 2.6.22 -mm merge plans Stefan Richter
2007-05-01 2:30 ` 2.6.22 -mm merge plans (RE: input) Dmitry Torokhov
2007-05-01 8:11 ` 2.6.22 -mm merge plans -- pfn_valid_within Andy Whitcroft
2007-05-01 8:19 ` Andrew Morton
2007-05-01 8:44 ` 2.6.22 -mm merge plans -- vm bugfixes Nick Piggin
2007-05-01 8:54 ` Andrew Morton
2007-05-01 19:31 ` Hugh Dickins
2007-05-02 3:08 ` Nick Piggin
2007-05-02 9:15 ` Nick Piggin
2007-05-02 14:00 ` Hugh Dickins
2007-05-03 1:32 ` Nick Piggin
2007-05-03 10:37 ` Christoph Hellwig
2007-05-03 12:56 ` Nick Piggin
2007-05-04 9:23 ` Nick Piggin
2007-05-04 9:43 ` Nick Piggin
2007-05-08 3:03 ` Benjamin Herrenschmidt
2007-05-03 12:24 ` Hugh Dickins
2007-05-03 12:43 ` Nick Piggin
2007-05-03 12:58 ` Hugh Dickins
2007-05-03 13:08 ` Nick Piggin
2007-05-03 16:52 ` Andrew Morton
2007-05-04 4:16 ` Nick Piggin
2007-05-09 12:34 ` Nick Piggin
2007-05-09 14:28 ` Hugh Dickins
2007-05-09 14:45 ` Nick Piggin
2007-05-09 15:38 ` Hugh Dickins
2007-05-09 22:24 ` Nick Piggin
2007-05-01 8:46 ` pcmcia ioctl removal Christoph Hellwig
2007-05-01 8:56 ` Russell King
2007-05-01 8:57 ` Willy Tarreau
2007-05-01 9:08 ` Andrew Morton
2007-05-01 14:46 ` Adrian Bunk
2007-05-01 9:16 ` Robert P. J. Day
2007-05-01 9:44 ` Willy Tarreau
2007-05-01 10:16 ` Robert P. J. Day
2007-05-01 10:26 ` Gabriel C
2007-05-01 10:52 ` Willy Tarreau
2007-05-01 10:12 ` Jan Engelhardt
2007-05-01 11:00 ` Willy Tarreau
2007-05-01 12:06 ` Konstantin Münning
2007-05-01 13:56 ` Rogan Dawes
2007-05-01 19:10 ` Russell King
2007-05-01 20:41 ` Jan Engelhardt
2007-05-09 12:54 ` Pavel Machek
2007-05-09 13:00 ` Robert P. J. Day
2007-05-09 13:03 ` Adrian Bunk
2007-05-09 19:11 ` Romano Giannetti
2007-05-10 12:40 ` Adrian Bunk
2007-05-01 8:48 ` pci hotplug patches Christoph Hellwig
2007-05-02 3:57 ` Greg KH
2007-05-13 20:59 ` Christoph Hellwig
2007-05-14 11:48 ` Greg KH
2007-05-01 8:54 ` cache-pipe-buf-page-address-for-non-highmem-arch.patch Christoph Hellwig
2007-05-01 9:04 ` cache-pipe-buf-page-address-for-non-highmem-arch.patch Andrew Morton
2007-05-01 11:31 ` cache-pipe-buf-page-address-for-non-highmem-arch.patch Andi Kleen
2007-05-03 3:48 ` cache-pipe-buf-page-address-for-non-highmem-arch.patch Ken Chen
2007-05-01 8:55 ` consolidate-generic_writepages-and-mpage_writepages.patch Christoph Hellwig
2007-05-01 9:17 ` 2.6.22 -mm merge plans Pekka Enberg
2007-05-01 9:24 ` Christoph Hellwig
2007-05-01 9:37 ` Peter Zijlstra
2007-05-01 12:19 ` Andi Kleen
2007-05-01 17:12 ` Pekka Enberg
2007-05-01 10:16 ` fragmentation avoidance " Mel Gorman
2007-05-01 13:02 ` 2.6.22 -mm merge plans -- lumpy reclaim Andy Whitcroft
2007-05-01 18:03 ` Peter Zijlstra
2007-05-01 19:00 ` Andrew Morton
2007-05-01 14:54 ` fragmentation avoidance Re: 2.6.22 -mm merge plans Christoph Lameter
2007-05-01 19:00 ` Mel Gorman
2007-05-01 18:57 ` Andrew Morton
2007-05-07 13:07 ` Yasunori Goto
2007-05-01 14:31 ` 2.6.22 -mm merge plans: mm-more-rmap-checking Hugh Dickins
2007-05-02 1:42 ` Nick Piggin
2007-05-02 13:17 ` Hugh Dickins
2007-05-03 0:18 ` Nick Piggin
2007-05-01 16:56 ` 2.6.22 -mm merge plans Zan Lynx
2007-05-01 17:06 ` 2.6.22 -mm merge plans: mm-detach_vmas_to_be_unmapped-fix Hugh Dickins
2007-05-01 18:10 ` 2.6.22 -mm merge plans: slub Hugh Dickins
2007-05-01 19:25 ` Christoph Lameter
2007-05-01 19:55 ` Andrew Morton
2007-05-01 20:19 ` Hugh Dickins
2007-05-01 20:36 ` Andrew Morton
2007-05-01 20:46 ` Christoph Lameter
2007-05-01 21:09 ` Andrew Morton
2007-05-02 12:54 ` Hugh Dickins
2007-05-02 17:03 ` Christoph Lameter
2007-05-02 19:11 ` Andrew Morton
2007-05-02 19:42 ` Christoph Lameter
2007-05-02 19:54 ` Sam Ravnborg
2007-05-02 20:14 ` Christoph Lameter
2007-05-02 18:52 ` Siddha, Suresh B
2007-05-02 18:58 ` Christoph Lameter
2007-05-01 21:08 ` Christoph Lameter
2007-05-02 12:45 ` Hugh Dickins
2007-05-02 17:01 ` Christoph Lameter
2007-05-02 18:08 ` Hugh Dickins
2007-05-02 18:28 ` Christoph Lameter
2007-05-02 18:42 ` Andrew Morton
2007-05-02 18:53 ` Christoph Lameter
2007-05-02 17:25 ` Christoph Lameter
2007-05-02 18:36 ` Hugh Dickins
2007-05-02 18:39 ` Christoph Lameter
2007-05-02 18:57 ` Andrew Morton
2007-05-02 19:01 ` Christoph Lameter
2007-05-02 19:18 ` Pekka Enberg
2007-05-02 19:34 ` Christoph Lameter
2007-05-02 19:43 ` Christoph Lameter
2007-05-03 8:15 ` Andrew Morton
2007-05-03 8:27 ` William Lee Irwin III
2007-05-03 16:30 ` Christoph Lameter
2007-05-03 8:46 ` Hugh Dickins
2007-05-03 8:57 ` Andrew Morton
2007-05-03 9:15 ` Hugh Dickins
2007-05-03 21:04 ` 2.6.22 -mm merge plans: slub on PowerPC Hugh Dickins
2007-05-03 21:15 ` Christoph Lameter
2007-05-03 22:41 ` Hugh Dickins
2007-05-04 0:25 ` Benjamin Herrenschmidt
2007-05-04 0:54 ` Christoph Lameter
2007-05-03 16:45 ` 2.6.22 -mm merge plans: slub Christoph Lameter
2007-05-03 15:54 ` swap-prefetch: 2.6.22 -mm merge plans Ingo Molnar
2007-05-03 16:15 ` Michal Piotrowski
2007-05-03 16:23 ` Michal Piotrowski
2007-05-03 22:14 ` Con Kolivas
2007-05-04 7:34 ` Nick Piggin
2007-05-04 8:52 ` Ingo Molnar
2007-05-04 9:09 ` Nick Piggin
2007-05-04 12:10 ` Con Kolivas
2007-05-05 8:42 ` Con Kolivas [this message]
2007-05-06 10:13 ` [ck] " Antonino Ingargiola
2007-05-06 18:22 ` Jory A. Pratt
2007-05-09 23:28 ` Con Kolivas
2007-05-10 0:05 ` Nick Piggin
2007-05-10 1:34 ` Con Kolivas
2007-05-10 1:56 ` Nick Piggin
2007-05-10 3:48 ` Ray Lee
2007-05-10 3:56 ` Nick Piggin
2007-05-10 5:52 ` Ray Lee
2007-05-10 7:04 ` Nick Piggin
2007-05-10 7:20 ` William Lee Irwin III
2007-05-10 12:34 ` Ray Lee
2007-05-12 4:46 ` [PATCH] mm: swap prefetch improvements Con Kolivas
2007-05-12 5:03 ` Paul Jackson
2007-05-12 5:15 ` Con Kolivas
2007-05-12 5:51 ` Paul Jackson
2007-05-12 7:28 ` Con Kolivas
2007-05-12 8:14 ` Paul Jackson
2007-05-12 8:21 ` Con Kolivas
2007-05-12 8:37 ` Paul Jackson
2007-05-12 8:57 ` [PATCH respin] " Con Kolivas
2007-05-21 10:03 ` [PATCH] " Ingo Molnar
2007-05-21 13:44 ` Con Kolivas
2007-05-21 16:00 ` Ingo Molnar
2007-05-22 10:15 ` Antonino Ingargiola
2007-05-22 10:20 ` Con Kolivas
2007-05-22 10:25 ` Ingo Molnar
2007-05-22 10:37 ` Con Kolivas
2007-05-22 10:46 ` Ingo Molnar
2007-05-22 10:54 ` Con Kolivas
2007-05-22 10:57 ` Ingo Molnar
2007-05-22 11:04 ` Con Kolivas
[not found] ` <20070522111104.GA14950@elte.hu>
2007-05-22 11:12 ` Ingo Molnar
2007-05-22 20:18 ` [ck] " Michael Chang
2007-05-22 20:31 ` Ingo Molnar
2007-05-10 3:58 ` swap-prefetch: 2.6.22 -mm merge plans Con Kolivas
2007-05-07 14:28 ` Bill Davidsen
2007-05-07 14:18 ` Bill Davidsen
2007-05-07 17:47 ` Josef Sipek
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=200705051842.32328.kernel@kolivas.org \
--to=kernel@kolivas.org \
--cc=akpm@linux-foundation.org \
--cc=ck@vds.kolivas.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mingo@elte.hu \
--cc=nickpiggin@yahoo.com.au \
/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