Den 2009-11-25 17:52 skrev Bob Friesenhahn:
On Wed, 25 Nov 2009, Peter Rosin wrote:
Today I have tested the pr-msvc-support branch on Linux
without noticing any regressions from master in the testsuite.
I have used debian testing (squeeze) since that's where
I found the required autotools versions.
I think that people are quite interested to hear about build performance
differences. Can you time execution of the test suite for both HEAD and
pr-msvc-support branch and see what execution impact there may be?
master:
$ time make -k check
*snip*
real 25m4.150s
user 13m3.800s
sys 10m51.980s
pr-msvc-support:
$ time make -k check
*snip*
real 23m49.380s
user 12m15.370s
sys 10m32.390s
master again:
$ time make -k check
*snip*
real 23m35.390s
user 12m9.290s
sys 10m24.720s
This is in a VM on an XP box with some programs started but otherwise
"doing nothing"...
Cheers,
Peter
--
They are in the crowd with the answer before the question.
> Why do you dislike Jeopardy?
