I was thinking of adding decision rates to cpudrv kstat. Powertop may be better because it would use dtrace probes that do not effect performance when not in use.
Regards, Bill On 12/04/08 16:58, Raj, Ashok wrote: > Should we add some sort of logging facility to powertop (maybe it already > does?) to get history or log generated with decisions to switch > pstate/cstate/intr rate or other parameters. This will help analyze running > different workloads and see how our PAD scheduler was good/bad with choices > made? > > > > >> -----Original Message----- >> From: tesla-dev-bounces at opensolaris.org [mailto:tesla-dev-bounces at >> opensolaris.org] On Behalf Of Bill >> Holler >> Sent: Thursday, December 04, 2008 12:03 PM >> To: tesla-dev at opensolaris.org >> Subject: [tesla-dev] Some experiments with C-state throttleing based on >> number of active CPUs >> >> Hi Tesla Dev, >> >> I have been experimenting not going into C3 or C2 when >> the number of non-idle CPUs in the cpu-partion exceeds >> a threshold. This is an attempt to regain the high load performance. >> These numbers are on a 2-socket system. >> >> With C3 threshold = 40% active and C2 threshold = 60% active >> the libmicro fork_1000 benchmark completes in 80/100 the time. >> >> With C3 threshold = 20% active and C2 threshold = 30% active >> the libmicro fork_1000 benchmark completes in 50/100 the time. >> These numbers are still about 80/100 slower than with c-states >> totally disabled. >> >> >> I am also going to experiment with cpu idle/wakeup rate. >> I suspect this may be more important for performance than >> the number of active cpus. >> >> Regards, >> Bill >> _______________________________________________ >> tesla-dev mailing list >> tesla-dev at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/tesla-dev >>
