From: Hugh Dickins <hugh@veritas.com>
To: Andrew Morton <akpm@digeo.com>
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: lmbench on anobjrmap
Date: Mon, 24 Mar 2003 17:35:18 +0000 (GMT) [thread overview]
Message-ID: <Pine.LNX.4.44.0303241731540.2456-100000@localhost.localdomain> (raw)
For the historical record...
lmbench comparison of 2.5.65-mm3 (which included anonymous extension
of objrmap) against 2.5.65-mm3 without it (those patches backed out).
fork and exec tests show 9% benefit from anobjrmap
(looks like selct TCP does significant forking too).
Both machines booted with mem=512M, but the 2*HT*P4 configured
HIGHMEM4G, and the 2*PIII configured HIGHMEM64G and HIGHPTE.
Processor, Processes - times in microseconds - smaller is better
----------------------------------------------------------------
Host OS Mhz null null open selct sig sig fork exec sh
call I/O stat clos TCP inst hndl proc proc proc
--------- ------------- ---- ---- ---- ---- ---- ----- ---- ---- ---- ---- ----
2*HT*P4 with anobjrmap 1800 0.73 1.22 6.01 7.85 39.8 1.41 4.75 405. 1287 4246
2*HT*P4 with anobjrmap 1800 0.73 1.22 6.01 7.88 39.8 1.43 4.70 406. 1287 4172
2*HT*P4 with anobjrmap 1800 0.77 1.22 6.03 7.83 39.8 1.45 4.63 407. 1269 4192
2*HT*P4 with anobjrmap 1800 0.72 1.21 6.58 7.92 40.3 1.44 4.68 403. 1241 4229
2*HT*P4 with anobjrmap 1800 0.72 1.21 5.95 7.84 43.2 1.43 4.77 404. 1242 3989
2*HT*P4 with anobjrmap 1800 0.72 1.22 6.10 7.88 39.8 1.44 4.68 435. 1254 4025
2*HT*P4 w/o anobjrmap 1800 0.73 1.22 6.19 8.02 40.2 1.43 4.64 455. 1336 4133
2*HT*P4 w/o anobjrmap 1800 0.73 1.23 6.18 7.96 40.0 1.53 4.63 462. 1340 4131
2*HT*P4 w/o anobjrmap 1800 0.73 1.22 6.21 8.04 40.4 1.45 4.60 441. 1326 4161
2*HT*P4 w/o anobjrmap 1800 0.78 1.22 6.12 7.90 40.2 1.44 4.93 447. 1344 4151
2*HT*P4 w/o anobjrmap 1800 0.72 1.21 6.08 7.87 42.5 1.46 4.79 444. 1325 4140
2*HT*P4 w/o anobjrmap 1800 0.78 1.22 6.11 7.88 40.3 1.46 4.51 446. 1314 4157
2*PIII with anobjrmap 794 0.46 0.84 27.8 31.2 37.0 1.27 5.50 539. 2164 6959
2*PIII with anobjrmap 794 0.46 0.85 28.0 31.3 35.7 1.26 5.33 540. 2048 6895
2*PIII with anobjrmap 794 0.46 0.84 27.8 31.2 35.5 1.27 5.26 503. 2150 6962
2*PIII with anobjrmap 794 0.46 0.81 27.8 30.8 35.4 1.26 5.50 547. 2076 6761
2*PIII with anobjrmap 794 0.46 0.84 27.8 31.0 36.7 1.26 5.53 553. 2235 6931
2*PIII with anobjrmap 794 0.46 0.83 28.0 30.9 37.8 1.27 5.55 536. 2073 6842
2*PIII w/o anobjrmap 794 0.46 0.83 28.0 31.2 38.5 1.26 5.31 570. 2323 7053
2*PIII w/o anobjrmap 794 0.46 0.82 27.8 31.3 36.8 1.27 5.56 631. 2321 7062
2*PIII w/o anobjrmap 794 0.46 0.83 27.8 31.3 38.9 1.27 5.36 655. 2349 7145
2*PIII w/o anobjrmap 794 0.46 0.81 27.7 31.2 38.3 1.27 5.57 564. 2311 7011
2*PIII w/o anobjrmap 794 0.46 0.80 27.9 31.5 37.4 1.27 5.36 611. 2332 7081
2*PIII w/o anobjrmap 794 0.46 0.82 27.8 31.2 35.4 1.26 5.31 595. 2332 7084
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
reply other threads:[~2003-03-24 17:35 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=Pine.LNX.4.44.0303241731540.2456-100000@localhost.localdomain \
--to=hugh@veritas.com \
--cc=akpm@digeo.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox