Message ID | 14777287.JCcGWNJJiE@work (mailing list archive) |
---|---|
State | Changes Requested, archived |
Delegated to: | Shuah Khan |
Headers | show |
Series | cpupower: Introduce powercap module | expand |
On 11/17/22 02:48, Thomas Renninger wrote: > This CPU power monitor shows the power consumption > as exposed by the powercap subsystem, cmp with: > Documentation/power/powercap/powercap.rst > > cpupower monitor -m RAPL > CPU| pack | core | unco > 0|6853926|967832|442381 > 8|6853926|967832|442381 > 1|6853926|967832|442381 > 9|6853926|967832|442381 > > Unfortunately RAPL domains cannot be directly mapped to the corresponding > CPU socket/package, core it belongs to. > Not sure this is possible at all with the current data exposed from the kernel. > > Still it can be worthful information for developers trying to optimize > power consumption of workloads or their system in general. > > Signed-off-by: Thomas Renninger <trenn@suse.de> > CC: Zhang Rui <rui.zhang@intel.com> > CC: Shuah Khan <skhan@linuxfoundation.org> > --- > tools/power/cpupower/Makefile | 1 + > .../utils/idle_monitor/cpupower-monitor.c | 7 +- > .../utils/idle_monitor/idle_monitors.def | 1 + > .../utils/idle_monitor/rapl_monitor.c | 147 ++++++++++++++++++ > 4 files changed, 153 insertions(+), 3 deletions(-) > create mode 100644 tools/power/cpupower/utils/idle_monitor/rapl_monitor.c > > diff --git a/tools/power/cpupower/Makefile b/tools/power/cpupower/Makefile > index 9fd3b309b3a6..59bfa05dec5d 100644 > --- a/tools/power/cpupower/Makefile > +++ b/tools/power/cpupower/Makefile > @@ -131,6 +131,7 @@ UTIL_OBJS = utils/helpers/amd.o utils/helpers/msr.o \ > utils/idle_monitor/hsw_ext_idle.o \ > utils/idle_monitor/amd_fam14h_idle.o utils/idle_monitor/cpuidle_sysfs.o \ > utils/idle_monitor/mperf_monitor.o utils/idle_monitor/cpupower-monitor.o \ > + utils/idle_monitor/rapl_monitor.o \ > utils/cpupower.o utils/cpufreq-info.o utils/cpufreq-set.o \ > utils/cpupower-set.o utils/cpupower-info.o utils/cpuidle-info.o \ > utils/cpuidle-set.o utils/powercap-info.o > diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c > index 7c77045fef52..075e766ff1f3 100644 > --- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c > +++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c > @@ -459,9 +459,10 @@ int cmd_monitor(int argc, char **argv) > print_results(1, cpu); > } > > - for (num = 0; num < avail_monitors; num++) > - monitors[num]->unregister(); > - > + for (num = 0; num < avail_monitors; num++) { > + if (monitors[num]->unregister) > + monitors[num]->unregister(); > + } > cpu_topology_release(cpu_top); > return 0; > } > diff --git a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def > index 0d6ba4dbb9c7..7c926e90c87e 100644 > --- a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def > +++ b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def > @@ -4,5 +4,6 @@ DEF(intel_nhm) > DEF(intel_snb) > DEF(intel_hsw_ext) > DEF(mperf) > +DEF(rapl) > #endif > DEF(cpuidle_sysfs) > diff --git a/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c b/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c > new file mode 100644 > index 000000000000..1acadf5ef90e > --- /dev/null > +++ b/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c > @@ -0,0 +1,147 @@ > +/* > + * (C) 2016 Thomas Renninger <trenn@suse.com> > + * > + * Licensed under the terms of the GNU GPL License version 2. Missing SPDX - please add > + * > + */ > + > +#if defined(__i386__) || defined(__x86_64__) I am seeing checpatch errors in this one as well in addition to missing SPDX. Running checkpatch will show you some of the errors. thanks, -- Shuah
diff --git a/tools/power/cpupower/Makefile b/tools/power/cpupower/Makefile index 9fd3b309b3a6..59bfa05dec5d 100644 --- a/tools/power/cpupower/Makefile +++ b/tools/power/cpupower/Makefile @@ -131,6 +131,7 @@ UTIL_OBJS = utils/helpers/amd.o utils/helpers/msr.o \ utils/idle_monitor/hsw_ext_idle.o \ utils/idle_monitor/amd_fam14h_idle.o utils/idle_monitor/cpuidle_sysfs.o \ utils/idle_monitor/mperf_monitor.o utils/idle_monitor/cpupower-monitor.o \ + utils/idle_monitor/rapl_monitor.o \ utils/cpupower.o utils/cpufreq-info.o utils/cpufreq-set.o \ utils/cpupower-set.o utils/cpupower-info.o utils/cpuidle-info.o \ utils/cpuidle-set.o utils/powercap-info.o diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c index 7c77045fef52..075e766ff1f3 100644 --- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c +++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c @@ -459,9 +459,10 @@ int cmd_monitor(int argc, char **argv) print_results(1, cpu); } - for (num = 0; num < avail_monitors; num++) - monitors[num]->unregister(); - + for (num = 0; num < avail_monitors; num++) { + if (monitors[num]->unregister) + monitors[num]->unregister(); + } cpu_topology_release(cpu_top); return 0; } diff --git a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def index 0d6ba4dbb9c7..7c926e90c87e 100644 --- a/tools/power/cpupower/utils/idle_monitor/idle_monitors.def +++ b/tools/power/cpupower/utils/idle_monitor/idle_monitors.def @@ -4,5 +4,6 @@ DEF(intel_nhm) DEF(intel_snb) DEF(intel_hsw_ext) DEF(mperf) +DEF(rapl) #endif DEF(cpuidle_sysfs) diff --git a/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c b/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c new file mode 100644 index 000000000000..1acadf5ef90e --- /dev/null +++ b/tools/power/cpupower/utils/idle_monitor/rapl_monitor.c @@ -0,0 +1,147 @@ +/* + * (C) 2016 Thomas Renninger <trenn@suse.com> + * + * Licensed under the terms of the GNU GPL License version 2. + * + */ + +#if defined(__i386__) || defined(__x86_64__) + +#include <stdio.h> +#include <stdlib.h> +#include <stdint.h> +#include <time.h> +#include <string.h> + +#include <pci/pci.h> + +#include "idle_monitor/cpupower-monitor.h" +#include "helpers/helpers.h" +#include "powercap.h" + +#define MAX_RAPL_ZONES 10 + +int rapl_zone_count = 0; +cstate_t rapl_zones[MAX_RAPL_ZONES]; +struct powercap_zone *rapl_zones_pt[MAX_RAPL_ZONES] = { 0 }; + +unsigned long long rapl_zone_previous_count[MAX_RAPL_ZONES]; +unsigned long long rapl_zone_current_count[MAX_RAPL_ZONES]; +unsigned long long rapl_max_count; + +static int rapl_get_count_uj(unsigned int id, unsigned long long *count, + unsigned int cpu) +{ + if (rapl_zones_pt[id] == NULL) + /* error */ + return -1; + + *count = rapl_zone_current_count[id] - rapl_zone_previous_count[id]; + + return 0; +} + +static int powercap_count_zones(struct powercap_zone *zone) +{ + uint64_t val; + int uj; + if (rapl_zone_count >= MAX_RAPL_ZONES) + return -1; + + if (!zone->has_energy_uj) + return 0; + + printf("%s\n", zone->sys_name); + uj = powercap_get_energy_uj(zone, &val); + printf("%d\n", uj); + + strncpy(rapl_zones[rapl_zone_count].name, zone->name, CSTATE_NAME_LEN - 1); + strcpy(rapl_zones[rapl_zone_count].desc, ""); + rapl_zones[rapl_zone_count].id = rapl_zone_count; + rapl_zones[rapl_zone_count].range = RANGE_MACHINE; + rapl_zones[rapl_zone_count].get_count = rapl_get_count_uj; + rapl_zones_pt[rapl_zone_count] = zone; + rapl_zone_count++; + + return 0; +} + +static int rapl_start(void) +{ + int i, ret; + uint64_t uj_val; + + for (i = 0; i < rapl_zone_count; i++) { + ret = powercap_get_energy_uj(rapl_zones_pt[i], &uj_val); + if (ret) + return ret; + rapl_zone_previous_count[i] = uj_val; + } + + return 0; +} + +static int rapl_stop(void) +{ + int i; + uint64_t uj_val; + + for (i = 0; i < rapl_zone_count; i++) { + int ret; + ret = powercap_get_energy_uj(rapl_zones_pt[i], &uj_val); + if (ret) + return ret; + rapl_zone_current_count[i] = uj_val; + if (rapl_max_count < uj_val) + rapl_max_count = uj_val - rapl_zone_previous_count[i]; + } + return 0; +} + +struct cpuidle_monitor *rapl_register(void) +{ + struct powercap_zone *root_zone; + char line[MAX_LINE_LEN] = ""; + int ret, val; + + ret = powercap_get_driver(line, MAX_LINE_LEN); + if (ret < 0) { + dprint("No powercapping driver loaded\n"); + return NULL; + } + + dprint("Driver: %s\n", line); + ret = powercap_get_enabled(&val); + if (ret < 0) + return NULL; + if (!val) { + dprint("Powercapping is disabled\n"); + return NULL; + } + + dprint("Powercap domain hierarchy:\n\n"); + root_zone = powercap_init_zones(); + + if (root_zone == NULL) { + dprint("No powercap info found\n"); + return NULL; + } + + powercap_walk_zones(root_zone, powercap_count_zones); + rapl_monitor.hw_states_num = rapl_zone_count; + + return &rapl_monitor; +} + +struct cpuidle_monitor rapl_monitor = { + .name = "RAPL", + .hw_states = rapl_zones, + .hw_states_num = 0, + .start = rapl_start, + .stop = rapl_stop, + .do_register = rapl_register, + .flags.needs_root = 0, + .overflow_s = 60 * 60 * 24 * 100, /* To be implemented */ +}; + +#endif
This CPU power monitor shows the power consumption as exposed by the powercap subsystem, cmp with: Documentation/power/powercap/powercap.rst cpupower monitor -m RAPL CPU| pack | core | unco 0|6853926|967832|442381 8|6853926|967832|442381 1|6853926|967832|442381 9|6853926|967832|442381 Unfortunately RAPL domains cannot be directly mapped to the corresponding CPU socket/package, core it belongs to. Not sure this is possible at all with the current data exposed from the kernel. Still it can be worthful information for developers trying to optimize power consumption of workloads or their system in general. Signed-off-by: Thomas Renninger <trenn@suse.de> CC: Zhang Rui <rui.zhang@intel.com> CC: Shuah Khan <skhan@linuxfoundation.org> --- tools/power/cpupower/Makefile | 1 + .../utils/idle_monitor/cpupower-monitor.c | 7 +- .../utils/idle_monitor/idle_monitors.def | 1 + .../utils/idle_monitor/rapl_monitor.c | 147 ++++++++++++++++++ 4 files changed, 153 insertions(+), 3 deletions(-) create mode 100644 tools/power/cpupower/utils/idle_monitor/rapl_monitor.c