X86_ENERGY_PERF_POLICY(8) System Manager's Manual X86_ENERGY_PERF_POLICY(8)NAME
x86_energy_perf_policy - read or write MSR_IA32_ENERGY_PERF_BIAS
SYNOPSIS
x86_energy_perf_policy [-c cpu] [-v] -r
x86_energy_perf_policy [-c cpu] [-v] 'performance'
x86_energy_perf_policy [-c cpu] [-v] 'normal'
x86_energy_perf_policy [-c cpu] [-v] 'powersave'
x86_energy_perf_policy [-c cpu] [-v] n
DESCRIPTION
x86_energy_perf_policy allows software to convey its policy for the relative importance of performance versus energy savings to the proces-
sor.
The processor uses this information in model-specific ways when it must select trade-offs between performance and energy efficiency.
This policy hint does not supersede Processor Performance states (P-states) or CPU Idle power states (C-states), but allows software to
have influence where it would otherwise be unable to express a preference.
For example, this setting may tell the hardware how aggressively or conservatively to control frequency in the "turbo range" above the
explicitly OS-controlled P-state frequency range. It may also tell the hardware how aggressively is should enter the OS requested C-
states.
Support for this feature is indicated by CPUID.06H.ECX.bit3 per the Intel Architectures Software Developer's Manual.
Options
-c limits operation to a single CPU. The default is to operate on all CPUs. Note that MSR_IA32_ENERGY_PERF_BIAS is defined per logical
processor, but that the initial implementations of the MSR were shared among all processors in each package.
-v increases verbosity. By default x86_energy_perf_policy is silent.
-r is for "read-only" mode - the unchanged state is read and displayed.
performance Set a policy where performance is paramount. The processor will be unwilling to sacrifice any performance for the sake of
energy saving. This is the hardware default.
normal Set a policy with a normal balance between performance and energy efficiency. The processor will tolerate minor performance compro-
mise for potentially significant energy savings. This reasonable default for most desktops and servers.
powersave Set a policy where the processor can accept a measurable performance hit to maximize energy efficiency.
n Set MSR_IA32_ENERGY_PERF_BIAS to the specified number. The range of valid numbers is 0-15, where 0 is maximum performance and 15 is max-
imum energy efficiency.
NOTES
x86_energy_perf_policy runs only as root.
FILES
/dev/cpu/*/msr
SEE ALSO msr(4)AUTHORS
Written by Len Brown <len.brown@intel.com>
X86_ENERGY_PERF_POLICY(8)
Check Out this Related Man Page
PMCCONTROL(8) BSD System Manager's Manual PMCCONTROL(8)NAME
pmccontrol -- control hardware performance monitoring counters
SYNOPSIS
pmccontrol [-c cpu | -d pmc | -e pmc] ...
pmccontrol -l
pmccontrol -L
pmccontrol -s
DESCRIPTION
The pmccontrol utility controls the operation of the system's hardware performance monitoring counters.
OPTIONS
The pmccontrol utility processes options in command line order, so later options modify the effect of earlier ones. The following options
are available:
-c cpu Subsequent enable and disable options affect the CPU denoted by argument cpu. The argument cpu is a number denoting a CPU in the
system, or ``*'', denoting all unhalted CPUs in the system.
-d pmc Disable PMC number pmc on the CPU specified by -c, preventing it from being used till subsequently re-enabled. The argument pmc is a
number denoting a specific PMC, or ``*'' denoting all the PMCs on the specified CPU.
Only idle PMCs may be disabled.
-e pmc Enable PMC number pmc, on the CPU specified by -c, allowing it to be used in the future. The argument pmc is a number denoting a
specific PMC, or ``*'' denoting all the PMCs on the specified CPU. If PMC pmc is already enabled, this option has no effect.
-l List available hardware performance counters and their current disposition.
-L List available hardware performance counter classes and their supported event names.
-s Print driver statistics maintained by hwpmc(4).
EXAMPLES
To disable all PMCs on all CPUs, use the command:
pmccontrol -d*
To enable all PMCs on all CPUs, use:
pmccontrol -e*
To disable PMCs 0 and 1 on CPU 2, use:
pmccontrol -c2 -d0 -d1
To disable PMC 0 of CPU 0 only, and enable all other PMCS on all other CPUs, use:
pmccontrol -c* -e* -c0 -d0
DIAGNOSTICS
The pmccontrol utility exits 0 on success, and >0 if an error occurs.
SEE ALSO pmc(3), pmclog(3), hwpmc(4), pmcstat(8), sysctl(8)HISTORY
The pmccontrol utility first appeared in FreeBSD 6.0.
AUTHORS
Joseph Koshy <jkoshy@FreeBSD.org>
BSD November 9, 2008 BSD