* reiserfs v3 vs v4 bonnie++ benchmarks
@ 2004-03-09 16:27 Mihai RUSU
2004-03-09 17:11 ` Hans Reiser
0 siblings, 1 reply; 10+ messages in thread
From: Mihai RUSU @ 2004-03-09 16:27 UTC (permalink / raw)
To: reiserfs-list
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi
Im playing with reiser4 lately and while doing this I have run some
bonnie++ tests on both v3 and v4 reiserfs, on the same machine.
I have this interesting results (I have run the tests 3 times for each FS
but the results were almost the same):
V3:
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
dizzywork 1G 13459 71 62922 20 26061 6 15687 83 54593 8 191.0 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
256:64000:16000/256 1287 27 165 2 11382 77 1230 25 63 1 253 2
V4:
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
dizzywork 1G 16926 87 52469 15 26474 11 16759 85 55349 9 166.6 1
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
256:64000:16000/256 863 40 1243 16 4219 47 873 46 81 1 3980 48
Interesting issues:
- - see that the "sequential writes" test using write with a 8192 bytes
buffer is lower with v4 than with v3 (and it was like that in all tests)
- - both reiserfs v3 and v4 have almost same number of "little" files
created in both sequential or random order but... reiser4 is much slower
than v3 here
- - different than v3, reiser4 has almost the same speed at deleting small
files in either random or sequential order but in the sequential case its
much slower than v3
- - Im very glad to see the BIG speed increase in reading small files in
sequential order although I would love to see more increase in the random
order case (closer to real-life)
System information:
- - P4 2ghz 512 kb cache 2 x 256 MB DDR
- - kernel 2.6.3 with reiser4 patch all.diff.gz downloaded today (from
LATEST dir)
- - mkfs default options for both v3 and v4
- - 40 gb single partition on a 40 gb IDE drive
hdparm -iv /dev/hdb
/dev/hdb:
multcount = 16 (on)
IO_support = 3 (32-bit w/sync)
unmaskirq = 0 (off)
using_dma = 1 (on)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 256 (on)
geometry = 65535/16/63, sectors = 78165360, start = 0
udma5, writecache enabled (8mb cache)
- - kernel options :
$ grep -i reiser .config
CONFIG_REISER4_FS=y
# CONFIG_REISER4_FS_SYSCALL is not set
CONFIG_REISER4_LARGE_KEY=y
# CONFIG_REISER4_CHECK is not set
CONFIG_REISER4_USE_EFLUSH=y
# CONFIG_REISER4_COPY_ON_CAPTURE is not set
# CONFIG_REISER4_BADBLOCKS is not set
CONFIG_REISERFS_FS=y
# CONFIG_REISERFS_CHECK is not set
# CONFIG_REISERFS_PROC_INFO is not set
- --
Mihai RUSU Email: dizzy@roedu.net
GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
"Linux is obsolete" -- AST
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFATfB5PZzOzrZY/1QRAnu5AJ9ZdGt9BE8iPz9iB/QrnzYOr2XIiwCcCgfb
lY0VWu16eyCMiQceZP1d708=
=l+SO
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 16:27 reiserfs v3 vs v4 bonnie++ benchmarks Mihai RUSU
@ 2004-03-09 17:11 ` Hans Reiser
2004-03-09 17:18 ` Mihai RUSU
0 siblings, 1 reply; 10+ messages in thread
From: Hans Reiser @ 2004-03-09 17:11 UTC (permalink / raw)
To: Mihai RUSU; +Cc: reiserfs-list
Mihai RUSU wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>Hi
>
>Im playing with reiser4 lately and while doing this I have run some
>bonnie++ tests on both v3 and v4 reiserfs, on the same machine.
>
>I have this interesting results (I have run the tests 3 times for each FS
>but the results were almost the same):
>
>V3:
>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>dizzywork 1G 13459 71 62922 20 26061 6 15687 83 54593 8 191.0 0
> ------Sequential Create------ --------Random Create--------
> -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>256:64000:16000/256 1287 27 165 2 11382 77 1230 25 63 1 253 2
>
>V4:
>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>dizzywork 1G 16926 87 52469 15 26474 11 16759 85 55349 9 166.6 1
> ------Sequential Create------ --------Random Create--------
> -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>256:64000:16000/256 863 40 1243 16 4219 47 873 46 81 1 3980 48
>
>
>Interesting issues:
>- - see that the "sequential writes" test using write with a 8192 bytes
>
>
which of the above is this, the sequential output test?
>buffer is lower with v4 than with v3 (and it was like that in all tests)
>- - both reiserfs v3 and v4 have almost same number of "little" files
>created in both sequential or random order but... reiser4 is much slower
>than v3 here
>
>
are little files = 64k?
>- - different than v3, reiser4 has almost the same speed at deleting small
>files in either random or sequential order but in the sequential case its
>much slower than v3
>- - Im very glad to see the BIG speed increase in reading small files in
>sequential order although I would love to see more increase in the random
>order case (closer to real-life)
>
>
I need to take the time to read through bonnie++ so that I can really
understand it, sigh, not today...
It does seem clear though that V4 consumes too much CPU for creates and
writes, and that we are CPU bound for this task. After we release it
Zam will tweak it more, and I will review the write code again. He is
tweaking the linux profiler to collect better statistics on parent callers.
Thanks for gathering this data.
>System information:
>- - P4 2ghz 512 kb cache 2 x 256 MB DDR
>- - kernel 2.6.3 with reiser4 patch all.diff.gz downloaded today (from
>LATEST dir)
>- - mkfs default options for both v3 and v4
>- - 40 gb single partition on a 40 gb IDE drive
>hdparm -iv /dev/hdb
>
>/dev/hdb:
> multcount = 16 (on)
> IO_support = 3 (32-bit w/sync)
> unmaskirq = 0 (off)
> using_dma = 1 (on)
> keepsettings = 0 (off)
> readonly = 0 (off)
> readahead = 256 (on)
> geometry = 65535/16/63, sectors = 78165360, start = 0
>
>udma5, writecache enabled (8mb cache)
>
>- - kernel options :
>$ grep -i reiser .config
>CONFIG_REISER4_FS=y
># CONFIG_REISER4_FS_SYSCALL is not set
>CONFIG_REISER4_LARGE_KEY=y
># CONFIG_REISER4_CHECK is not set
>CONFIG_REISER4_USE_EFLUSH=y
># CONFIG_REISER4_COPY_ON_CAPTURE is not set
># CONFIG_REISER4_BADBLOCKS is not set
>CONFIG_REISERFS_FS=y
># CONFIG_REISERFS_CHECK is not set
># CONFIG_REISERFS_PROC_INFO is not set
>
>- --
>Mihai RUSU Email: dizzy@roedu.net
>GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
> "Linux is obsolete" -- AST
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.2.4 (GNU/Linux)
>
>iD8DBQFATfB5PZzOzrZY/1QRAnu5AJ9ZdGt9BE8iPz9iB/QrnzYOr2XIiwCcCgfb
>lY0VWu16eyCMiQceZP1d708=
>=l+SO
>-----END PGP SIGNATURE-----
>
>
>
>
--
Hans
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 17:11 ` Hans Reiser
@ 2004-03-09 17:18 ` Mihai RUSU
2004-03-09 17:24 ` Hans Reiser
0 siblings, 1 reply; 10+ messages in thread
From: Mihai RUSU @ 2004-03-09 17:18 UTC (permalink / raw)
To: Hans Reiser; +Cc: reiserfs-list
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tue, 9 Mar 2004, Hans Reiser wrote:
> >V3:
> >Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> > -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> >Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
> >dizzywork 1G 13459 71 62922 20 26061 6 15687 83 54593 8 191.0 0
> > ------Sequential Create------ --------Random Create--------
> > -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> >files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >256:64000:16000/256 1287 27 165 2 11382 77 1230 25 63 1 253 2
> >
> >V4:
> >Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> > -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> >Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
> >dizzywork 1G 16926 87 52469 15 26474 11 16759 85 55349 9 166.6 1
> > ------Sequential Create------ --------Random Create--------
> > -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> >files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >256:64000:16000/256 863 40 1243 16 4219 47 873 46 81 1 3980 48
> >
> >
>
> >Interesting issues:
> >- - see that the "sequential writes" test using write with a 8192 bytes
> >
> >
> which of the above is this, the sequential output test?
This one:
62922 for v3 and 52469 for v4
> >buffer is lower with v4 than with v3 (and it was like that in all tests)
> >- - both reiserfs v3 and v4 have almost same number of "little" files
> >created in both sequential or random order but... reiser4 is much slower
> >than v3 here
> are little files = 64k?
In my oppinion yes (I consider not a small file after 256 kbytes or so).
I dont think that "definitions" matter here, but numbers do :) Im trying
to test a setup as close as it can be to my real setup (huge FS partitions
with millions of files, web files with 32kbytes average size).
> I need to take the time to read through bonnie++ so that I can really
> understand it, sigh, not today...
That would be good couse Im not sure I completly understand it myself heh.
> It does seem clear though that V4 consumes too much CPU for creates and
> writes, and that we are CPU bound for this task. After we release it
> Zam will tweak it more, and I will review the write code again. He is
> tweaking the linux profiler to collect better statistics on parent callers.
>
> Thanks for gathering this data.
You're welcomed.
- --
Mihai RUSU Email: dizzy@roedu.net
GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
"Linux is obsolete" -- AST
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFATfx9PZzOzrZY/1QRAlCqAJ9dVdk9tDbrRTmBn+oEP0NHL1RaOQCgy+ag
MHOxu6aUewDIpg1X4Jq3Y6k=
=I8Ya
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 17:18 ` Mihai RUSU
@ 2004-03-09 17:24 ` Hans Reiser
2004-03-09 17:32 ` Mihai RUSU
0 siblings, 1 reply; 10+ messages in thread
From: Hans Reiser @ 2004-03-09 17:24 UTC (permalink / raw)
To: Mihai RUSU; +Cc: reiserfs-list
Mihai RUSU wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On Tue, 9 Mar 2004, Hans Reiser wrote:
>
>
>
>>>V3:
>>>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
>>> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>>>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>>>dizzywork 1G 13459 71 62922 20 26061 6 15687 83 54593 8 191.0 0
>>> ------Sequential Create------ --------Random Create--------
>>> -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>>>files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>>>256:64000:16000/256 1287 27 165 2 11382 77 1230 25 63 1 253 2
>>>
>>>V4:
>>>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
>>> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>>>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>>>dizzywork 1G 16926 87 52469 15 26474 11 16759 85 55349 9 166.6 1
>>> ------Sequential Create------ --------Random Create--------
>>> -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>>>files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>>>256:64000:16000/256 863 40 1243 16 4219 47 873 46 81 1 3980 48
>>>
>>>
>>>
>>>
>>>Interesting issues:
>>>- - see that the "sequential writes" test using write with a 8192 bytes
>>>
>>>
>>>
>>>
>>which of the above is this, the sequential output test?
>>
>>
>
>This one:
>62922 for v3 and 52469 for v4
>
>
what is the amount of ram vs. the amount written? If you write 10x the
ram, does the performance loss go away?
>
>
>>>buffer is lower with v4 than with v3 (and it was like that in all tests)
>>>- - both reiserfs v3 and v4 have almost same number of "little" files
>>>created in both sequential or random order but... reiser4 is much slower
>>>than v3 here
>>>
>>>
>
>
>
>>are little files = 64k?
>>
>>
>
>In my oppinion yes (I consider not a small file after 256 kbytes or so).
>I dont think that "definitions" matter here, but numbers do :) Im trying
>to test a setup as close as it can be to my real setup (huge FS partitions
>with millions of files, web files with 32kbytes average size).
>
>
>
>>I need to take the time to read through bonnie++ so that I can really
>>understand it, sigh, not today...
>>
>>
>
>That would be good couse Im not sure I completly understand it myself heh.
>
>
>
>>It does seem clear though that V4 consumes too much CPU for creates and
>>writes, and that we are CPU bound for this task. After we release it
>>Zam will tweak it more, and I will review the write code again. He is
>>tweaking the linux profiler to collect better statistics on parent callers.
>>
>>Thanks for gathering this data.
>>
>>
>
>You're welcomed.
>
>- --
>Mihai RUSU Email: dizzy@roedu.net
>GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
> "Linux is obsolete" -- AST
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.2.4 (GNU/Linux)
>
>iD8DBQFATfx9PZzOzrZY/1QRAlCqAJ9dVdk9tDbrRTmBn+oEP0NHL1RaOQCgy+ag
>MHOxu6aUewDIpg1X4Jq3Y6k=
>=I8Ya
>-----END PGP SIGNATURE-----
>
>
>
>
--
Hans
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 17:24 ` Hans Reiser
@ 2004-03-09 17:32 ` Mihai RUSU
2004-03-09 17:38 ` Hans Reiser
0 siblings, 1 reply; 10+ messages in thread
From: Mihai RUSU @ 2004-03-09 17:32 UTC (permalink / raw)
To: Hans Reiser; +Cc: reiserfs-list
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tue, 9 Mar 2004, Hans Reiser wrote:
> >This one:
> >62922 for v3 and 52469 for v4
> >
> >
> what is the amount of ram vs. the amount written? If you write 10x the
> ram, does the performance loss go away?
bonnie++ writes by default in the "block" tests double of memory (so it
makes sure the cache has a small impact). I have 512 mb of ram and the
ammount of data written was 1gb. I can force it to write 10x ...
considering what I said here do you still want that ?
- --
Mihai RUSU Email: dizzy@roedu.net
GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
"Linux is obsolete" -- AST
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFATf+iPZzOzrZY/1QRAhIQAJ0W4Y7gvIMp0TnpXXXE9kubMgrr/wCePHCM
LBpuALzT+Mun64EcyAMd4bc=
=wQqP
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 17:32 ` Mihai RUSU
@ 2004-03-09 17:38 ` Hans Reiser
2004-03-09 21:14 ` Mihai RUSU
0 siblings, 1 reply; 10+ messages in thread
From: Hans Reiser @ 2004-03-09 17:38 UTC (permalink / raw)
To: Mihai RUSU; +Cc: reiserfs-list
Mihai RUSU wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On Tue, 9 Mar 2004, Hans Reiser wrote:
>
>
>
>>>This one:
>>>62922 for v3 and 52469 for v4
>>>
>>>
>>>
>>>
>>what is the amount of ram vs. the amount written? If you write 10x the
>>ram, does the performance loss go away?
>>
>>
>
>bonnie++ writes by default in the "block" tests double of memory (so it
>makes sure the cache has a small impact). I have 512 mb of ram and the
>ammount of data written was 1gb. I can force it to write 10x ...
>considering what I said here do you still want that ?
>
>- --
>Mihai RUSU Email: dizzy@roedu.net
>GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
> "Linux is obsolete" -- AST
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.2.4 (GNU/Linux)
>
>iD8DBQFATf+iPZzOzrZY/1QRAhIQAJ0W4Y7gvIMp0TnpXXXE9kubMgrr/wCePHCM
>LBpuALzT+Mun64EcyAMd4bc=
>=wQqP
>-----END PGP SIGNATURE-----
>
>
>
>
Yes, it would be interesting. I think maybe we need to tweak reiser4
more for benchmarks near to ram size.
--
Hans
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 17:38 ` Hans Reiser
@ 2004-03-09 21:14 ` Mihai RUSU
2004-03-10 6:33 ` Hans Reiser
0 siblings, 1 reply; 10+ messages in thread
From: Mihai RUSU @ 2004-03-09 21:14 UTC (permalink / raw)
To: Hans Reiser; +Cc: reiserfs-list
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tue, 9 Mar 2004, Hans Reiser wrote:
> Yes, it would be interesting. I think maybe we need to tweak reiser4
> more for benchmarks near to ram size.
Hi again
I have repeated the "block" read/write tests asking bonnie++ to write 5 GB
of data (10x my RAM size).
reiser v3:
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
dizzywork 5G 17340 89 46074 15 22140 5 16473 85 46145 6 102.8 0
dizzywork,5G,17340,89,46074,15,22140,5,16473,85,46145,6,102.8,0,,,,,,,,,,,,,
reiser v4:
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
dizzywork 5G 17491 90 48452 14 23563 11 17426 89 47952 8 110.9 0
dizzywork,5G,17491,90,48452,14,23563,11,17426,89,47952,8,110.9,0,,,,,,,,,,,,,
You were right. The values now seem almost the same (well, reiser4 seems a
little bit better).
Should I repeat the "many small files" tests too with lots more files ?
> --
> Hans
- --
Mihai RUSU Email: dizzy@roedu.net
GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
"Linux is obsolete" -- AST
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFATjObPZzOzrZY/1QRAnOMAJ4glX0TXNnuz3hbKQiR4SGijXvE1QCePYOd
HPRGgqPvWMuUb/UYIzUGtLE=
=Gyjj
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-09 21:14 ` Mihai RUSU
@ 2004-03-10 6:33 ` Hans Reiser
2004-03-11 8:33 ` Mihai RUSU
2004-03-25 23:35 ` Mike Fedyk
0 siblings, 2 replies; 10+ messages in thread
From: Hans Reiser @ 2004-03-10 6:33 UTC (permalink / raw)
To: Mihai RUSU; +Cc: reiserfs-list, Yann Dupont
Mihai RUSU wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On Tue, 9 Mar 2004, Hans Reiser wrote:
>
>
>
>>Yes, it would be interesting. I think maybe we need to tweak reiser4
>>more for benchmarks near to ram size.
>>
>>
>
>Hi again
>
>I have repeated the "block" read/write tests asking bonnie++ to write 5 GB
>of data (10x my RAM size).
>
>reiser v3:
>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>dizzywork 5G 17340 89 46074 15 22140 5 16473 85 46145 6 102.8 0
>dizzywork,5G,17340,89,46074,15,22140,5,16473,85,46145,6,102.8,0,,,,,,,,,,,,,
>
>reiser v4:
>Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
>dizzywork 5G 17491 90 48452 14 23563 11 17426 89 47952 8 110.9 0
>dizzywork,5G,17491,90,48452,14,23563,11,17426,89,47952,8,110.9,0,,,,,,,,,,,,,
>
>You were right. The values now seem almost the same (well, reiser4 seems a
>little bit better).
>
>Should I repeat the "many small files" tests too with lots more files ?
>
>
>
>>--
>>Hans
>>
>>
>
>- --
>Mihai RUSU Email: dizzy@roedu.net
>GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
> "Linux is obsolete" -- AST
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.2.4 (GNU/Linux)
>
>iD8DBQFATjObPZzOzrZY/1QRAnOMAJ4glX0TXNnuz3hbKQiR4SGijXvE1QCePYOd
>HPRGgqPvWMuUb/UYIzUGtLE=
>=Gyjj
>-----END PGP SIGNATURE-----
>
>
>
>
Sure. Thanks for repeating it.
Yann, I think you will get the same effect if you repeat with 10x larger
fileset. Reiser4 is not tuned for size of ram benchmarks, and I don't
think it should be.
--
Hans
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-10 6:33 ` Hans Reiser
@ 2004-03-11 8:33 ` Mihai RUSU
2004-03-25 23:35 ` Mike Fedyk
1 sibling, 0 replies; 10+ messages in thread
From: Mihai RUSU @ 2004-03-11 8:33 UTC (permalink / raw)
To: Hans Reiser; +Cc: reiserfs-list, Yann Dupont
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Wed, 10 Mar 2004, Hans Reiser wrote:
> Sure. Thanks for repeating it.
>
> Yann, I think you will get the same effect if you repeat with 10x larger
> fileset. Reiser4 is not tuned for size of ram benchmarks, and I don't
> think it should be.
Hi again
I repeated the "many small files" test with a double ammount of files (I
couldnt do more couse my test partition whould fill up). I have 512MB of
memory and in this tests df reported a 20GB usage of the partition so I
say its pretty plenty more than the memory :)
reiser v3:
Version 1.03 ------Sequential Create------ --------Random Create--------
dizzywork -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
512:64000:16000/512 1108 22 157 1 10093 67 1045 21 44 0 130 1
dizzywork,,,,,,,,,,,,,,512:64000:16000/512,1108,22,157,1,10093,67,1045,21,44,0,130,1
reiser v4:
Version 1.03 ------Sequential Create------ --------Random Create--------
dizzywork -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
512:64000:16000/512 730 35 1113 15 3882 46 767 42 45 1 821 12
dizzywork,,,,,,,,,,,,,,512:64000:16000/512,730,35,1113,15,3882,46,767,42,45,1,82
The results seem fimilar with the first "many files" test (maybe because
there too I was written much more than memory, ie 10GB).
Interesting observations:
- - v4 seems to be much slower than v3 at the "create files" test
- - v4 also is much slower at deleting many files sequentially but its much
faster than v3 at deleting them random (much closer to real life so this
is a gain for v4);
- - v4 also is a lot faster than v3 when reading many "small" files
sequentially althought I would love to see some increase in the "random
read" case where its closer to real-life case
> --
> Hans
- --
Mihai RUSU Email: dizzy@roedu.net
GPG : http://dizzy.roedu.net/dizzy-gpg.txt WWW: http://dizzy.roedu.net
"Linux is obsolete" -- AST
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFAUCRlPZzOzrZY/1QRAqJXAKC2zeMIpcMi8HW8xGAILKeH8Lp06wCcDASA
nCw0Pvfg3v4Br6g/aAmJTkI=
=joC5
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: reiserfs v3 vs v4 bonnie++ benchmarks
2004-03-10 6:33 ` Hans Reiser
2004-03-11 8:33 ` Mihai RUSU
@ 2004-03-25 23:35 ` Mike Fedyk
1 sibling, 0 replies; 10+ messages in thread
From: Mike Fedyk @ 2004-03-25 23:35 UTC (permalink / raw)
To: Hans Reiser; +Cc: Mihai RUSU, reiserfs-list, Yann Dupont
Hans Reiser wrote:
> Yann, I think you will get the same effect if you repeat with 10x larger
> fileset. Reiser4 is not tuned for size of ram benchmarks, and I don't
> think it should be.
Hans, first of all, can you please trim the areas of a message you quote
(especially after the area you're quoting if you don't write anything
below...)? It would make reading your replies much easier.
And second, the tuning isn't important now, especially since reiser4
hasn't made it into -mm yet, but it should definitely be on the to-do
list, just way down below stability, and "get into -mm".
Mike
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2004-03-25 23:35 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-03-09 16:27 reiserfs v3 vs v4 bonnie++ benchmarks Mihai RUSU
2004-03-09 17:11 ` Hans Reiser
2004-03-09 17:18 ` Mihai RUSU
2004-03-09 17:24 ` Hans Reiser
2004-03-09 17:32 ` Mihai RUSU
2004-03-09 17:38 ` Hans Reiser
2004-03-09 21:14 ` Mihai RUSU
2004-03-10 6:33 ` Hans Reiser
2004-03-11 8:33 ` Mihai RUSU
2004-03-25 23:35 ` Mike Fedyk
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.