Hey all!
I found out why, even after 'make update', I still didn't get the right values... After some investigation of my system configuration with Marc, it seems that my kernel configuration had the
CONFIG_HIGH_RES_TIMERS
option was disbled. After recompiling my kernel with it, Marc's patch using the nanosleep calls now works fine on my box. I just hope that this flag is set by default on the regular linux distros hehe!
Thanks to all for the help!
David
David St-Hilaire wrote:
vasil wrote:
Interpreted code shows average n about 9064 (peak value was 9193), compiled code - 9210 (peak value was 9226).
Hi all!
Sorry for the late response...
So, even after upgrading to the latest gambit-c version:
dave@david-laptop ~/sources/gambit-sources $ git log | head -n 3 commit fd8910a5c8649c037be72ec3349ffe535f0a7cd1 Author: Marc Feeley feeley@iro.umontreal.ca Date: Wed Jan 28 11:33:16 2009 -0500
I still get the bad performances...
dave@david-laptop ~/sources/gambit-sources $ gsi/gsi ~/toto.scm 102 dave@david-laptop ~/sources/gambit-sources $ gsc/gsc ~/toto.scm dave@david-laptop ~/sources/gambit-sources $ gsi/gsi ~/toto.o1 102
The toto.scm file is the example provided by Marc, with a (declare (not safe)) declartion added.
I'm using the gentoo kernel 2.6.28-gentoo-r1. Vasil, what kernel did you used?
Thank you for the help!
David