From: Con Kolivas <kernel@kolivas.org>
To: linux kernel mailing list <linux-kernel@vger.kernel.org>
Cc: Andrew Morton <akpm@osdl.org>
Subject: [BENCHMARK] reaim 2x,4x,8x with various SMP balancing patches
Date: Sun, 14 Sep 2003 19:44:10 +1000 [thread overview]
Message-ID: <200309141944.45097.kernel@kolivas.org> (raw)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
With the help of the OSDL hardware I've thrown a bootload of reaim benchmarks
at these various patches because of the confusion regarding their usefulness.
2.6.0-test5 is the baseline, and the 3 unique patches CMT,A3,BT were added by
themselves and all together to 2.6.0-test5 for comparison. Finally the full
2.6.0-test5-mm1 patch was compared. The A3 patch was modified slightly to
overcome the one minor magnitude error for fairness. Looking at the url for
each benchmark and examining the "Graph - Jobs per minute" is useful to
assess where the performance hovers.
Legend
260t5 - vanilla 2.6.0-test5
CMT - sched-CAN_MIGRATE_TASK-fix.patch
A3 - sched-2.6.0-test2-mm2-A3.patch
BT - sched-balance-tuning.patch
all - CMT + A3 + BT
mm1 - 2.6.0-test5-mm1
2 CPU:
Kernel Throughput URL
260t5 1337.11 http://khack.osdl.org/stp/279474/
CMT 1321.92 http://khack.osdl.org/stp/279796/
A3 1304.82 http://khack.osdl.org/stp/279799
BT 1326.27 http://khack.osdl.org/stp/279802
All 1337.79 http://khack.osdl.org/stp/279805/
mm1 1316.95 http://khack.osdl.org/stp/279588/
4 CPU:
Kernel Throughput URL
260t5 5406.68 http://khack.osdl.org/stp/279883
CMT
A3 5099.14 http://khack.osdl.org/stp/279800
BT 5721.79 http://khack.osdl.org/stp/279803
All 4919.36 http://khack.osdl.org/stp/279806
mm1 5360.32 http://khack.osdl.org/stp/279887
8 CPU:
Kernel Throughput URL
260t5 8812.21 http://khack.osdl.org/stp/279448/
CMT 8794.14 http://khack.osdl.org/stp/279798
A3 7084.15 http://khack.osdl.org/stp/279801
BT 8615.13 http://khack.osdl.org/stp/279804
All 7629.14 http://khack.osdl.org/stp/279919
mm1 8478.10 http://khack.osdl.org/stp/279562/
4xCMT seemed to not successfully finish.
In summary, it appears that all the patches cause detriment to performance,
except the combination of BT with 4 cpus. The A3 patch is particularly
detrimental, but it is reassuring to see that the extra patches in mm1 (which
includes O1int patches) recover a lot of that performance, even with the
other balancing patches.
Con
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)
iD8DBQE/ZDhtZUg7+tp6mRURAmJAAJ4zMYHg7iSuVHi7SpiS5hkmoaQl9wCggrcB
SSH6dbJthgABVnEIn4K/z3M=
=Jz7G
-----END PGP SIGNATURE-----
reply other threads:[~2003-09-14 9:36 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=200309141944.45097.kernel@kolivas.org \
--to=kernel@kolivas.org \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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.