John Martin wrote: > bill holler wrote: >>>> >>>> PowerTOP should not be the case that causes 100% highest p-state >>>> residency in idle. We have this bug >>>> "6818514 Event based CPUPM too easily shifts into high gear" >>>> >>>> fixed and putback since build 117 by rev#. 9802. It looks like it >>>> does not >>>> fix the problem. >>> I believe it does, at least for a i7 920 based system: >>> >>> OpenSolaris PowerTOP version 1.2 >>> >>> C-states (idle power) Avg Residency P-states (frequencies) >>> C0 (cpu running) (0.6%) 1596 Mhz 100.0% >>> C1 3.9ms (99.4%) 1729 Mhz 0.0% >> >> >> Hi, >> >> It looks like C-states are disabled in BIOS on this i7-920? >> You definitely want C-state enabled on i7 for power savings. > The C-states are set to "Auto" in the SBIOS, for which the comment > says something about deciding based on the capabilities of the CPU > and the motherboard (eVGA X58 SLI). Latest SBIOS is installed. > > The SBIOS also has explicit options for enabling deep C-states beyond > just C1: (Auto, C1, C1E, C3, C6). However, when I enable any of > these I still only > get C1 in the powertop output. I have also added "cpu_deep_idle > default" to /etc/power.conf > but that didn't change the behavior.
Sorry for hijacking this p-state thread. :-b What does "kstat cstate" show? These show what Solaris read from ACPI _CST. I suspect there will be no C2/C3 kstats. The BIOS setting sound correct.... Solaris enables deeper c-state by default on i7 if available. Regards, Bill
