Re: problem with "sched: revert back to per-rq vruntime"?
From: Jayson King
Date: Tue Dec 30 2008 - 14:35:56 EST
Here's some times using bash, for example. Let me know if there are some
other tests I can do.
It shows on vanilla 2.6.28, a huge increase in latency (about 37X to
39X) starting bash -i when the CPU is loaded with 100% nice tasks,
relative to idle CPU. But after reverting
f9c0b0950d5fd8c8c5af39bc061f27ea8fddcac3 the latency increase is not as
great (about 2X).
# running bash -i on vanilla 2.6.28,
# with idle CPU
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.020s
user 0m0.008s
sys 0m0.014s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.020s
user 0m0.014s
sys 0m0.005s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.020s
user 0m0.008s
sys 0m0.009s
# running bash -i on vanilla 2.6.28,
# with 100% nice CPU usage
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.753s
user 0m0.011s
sys 0m0.011s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.752s
user 0m0.012s
sys 0m0.012s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.786s
user 0m0.010s
sys 0m0.008s
# running 2.6.28 reverted f9c0b0950d5fd8c8c5af39bc061f27ea8fddcac3,
# with idle CPU
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.021s
user 0m0.012s
sys 0m0.007s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.020s
user 0m0.012s
sys 0m0.011s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.020s
user 0m0.011s
sys 0m0.012s
# running 2.6.28 reverted f9c0b0950d5fd8c8c5af39bc061f27ea8fddcac3,
# with 100% nice CPU usage
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.043s
user 0m0.010s
sys 0m0.009s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.046s
user 0m0.011s
sys 0m0.007s
[jayson@atlas tmp]$ time bash -i -c 'exit'
exit
real 0m0.047s
user 0m0.011s
sys 0m0.006s
Jayson
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/