Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 4 Oct 1999 15:44:18 -0200 (EDT)
From:      Fabio Cesar Gozzo <fabio@iqm.unicamp.br>
To:        smp@FreeBSD.ORG
Cc:        fabio@iqm.unicamp.br
Subject:   SMP improvements in current
Message-ID:  <199910041744.PAA12343@thomson.iqm.unicamp.br>

next in thread | raw e-mail | index | archive | help
I just built another SMP box (ASUS P2B-DS, 2x550, 512 MB RAM)
to use with FP intensive non-threaded programs. 
In a normal situation I run 2 jobs at once and top looks like:

  PID USERNAME PRI NICE  SIZE    RES STATE  C   TIME   WCPU    CPU COMMAND
 8495 fabio     97   0   118M 14644K RUN    0  40:30 98.97% 98.97% l502.exe
95064 fabio     97   0   270M 26700K CPU1   1 249:30 98.93% 98.93% l502.exe
 8558 fabio     28   0  1572K   828K CPU0   1   0:00  1.54%  0.15% top

IIRC, a simple CPU affinity was implemented in -current and, if I
understood it correctly, it could improve the calculations by attaching
the program to one CPU, and, most important, to one cache.
Is that correct ? Are there other SMP improvements in -current
that would speed up this kind of programs ?

Thank you.

PS: before someone asks, yes, I'm aware of the problems -current
can cause on a production machine. I have already use SMP under
-current in pre-3.0 times.

-- 
**************************************************
Fabio Gozzo                   fabio@iqm.unicamp.br
State University of Campinas               UNICAMP
Chemistry Institute  http://thomson.iqm.unicamp.br
**************************************************


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199910041744.PAA12343>