optimize msleep with PCATCH flag
David Xu
davidxu at freebsd.org
Sat May 20 03:12:31 UTC 2006
On Saturday 20 May 2006 09:49, Julian Elischer wrote:
> looks like you need to do a lot more samples to get a better error
> value.. (8 may not be enough)
Is 40 enough ? this time, the system runs in single user mode, so
any noise is eliminated.
x plain_PCATCH.st
+ PCATCH_optimized.st
+--------------------------------------------------------------------------+
| + |
| + + |
| + + |
| + + x x |
| + ++++ + x xx |
| ++++++++++ + xx x xxxxx x |
|+ +++++++++++++ + x xxx+xxxxxxxxx xxx x x xxx xx x x x|
| |____A____| |_______M____A___________| |
+--------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 40 31199 38404 32691.5 33376.775 1740.5576
+ 40 28245 31794 29300.5 29380.8 649.64192
Difference at 95.0% confidence
-3995.98 +/- 584.857
-11.9723% +/- 1.75229%
(Student's t, pooled s = 1313.69)
More information about the freebsd-threads
mailing list