On 28 March 2013 23:54, Jacob Shin <jacob.shin@xxxxxxx> wrote: > This patchset adds AMD specific powersave bias function to the ondemand > governor; which can be used to help ondemand governor make more power conscious > frequency change decisions based on feedback from hardware (availble on AMD > Family 16h and above). > > Hardware feedback tells software how "sensitive" to frequency changes the > workloads are. CPU-bound workloads will be more sensitive -- they will > perform better as frequency increases. Memory/IO-bound workloads will be less > sensitive -- they will not necessarily perform better as frequnecy increases. > > This patchset was compared against ondemand governor without powersave bias > and did not show any performance degradation on CPU-bound workloads such as > kernbench and unixbench. While saving power on Memory-bound workloads such as > stream. > > This applies to linux-pm's linux-next branch, on top of Viresh's 'Implement > per policy instance of governor' V4 patchset: > > https://lkml.org/lkml/2013/3/27/348 > > V2: > * Added proper include files to amd_freq_sensitivity.c > * Only register powersave_bias_target function pointer and not the entire > od_ops. > > Jacob Shin (2): > cpufreq: ondemand: allow custom powersave_bias_target function to be > registered > cpufreq: AMD "frequency sensitivity feedback" powersave bias for > ondemand governor > > arch/x86/include/uapi/asm/msr-index.h | 1 + > drivers/cpufreq/Kconfig.x86 | 10 +++ > drivers/cpufreq/Makefile | 1 + > drivers/cpufreq/amd_freq_sensitivity.c | 147 ++++++++++++++++++++++++++++++++ > drivers/cpufreq/cpufreq_governor.h | 3 + > drivers/cpufreq/cpufreq_ondemand.c | 22 ++++- > 6 files changed, 181 insertions(+), 3 deletions(-) > create mode 100644 drivers/cpufreq/amd_freq_sensitivity.c Acked-by: Viresh Kumar <viresh.kumar@xxxxxxxxxx> -- To unsubscribe from this list: send the line "unsubscribe cpufreq" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
