What's stopping my CPU from scaling down? [closed]
1
My CPU scales from 800MHz to 4.8GHz, obviously in a vastly different power envelope. I have an applet that tells me what speed it's running at on my panel and today I've noticed that it's running pretty close to the top end all the time. Even when it's essentially idle.
$ uptime
22:05:34 up 1 day, 5:37, 1 user, load average: 0.00, 0.00, 0.00I'm running on the "powersave" governor (intel_pstate driver).
So —short of closing random processes until it tweaks down— how do I go about finding what's stopping my CPU from underclocking?
Some follow up from the comments.
It's important to note that this system has scaled down in recent history, while on 18.04 (which is what I'm still using). There haven't been any major upgrades, save whatever Nvidia do to their driver. This is something new, perhaps just something this session. That's kind of my point of this question. Even if it fixes it, I don't just want to reboot. I want to find what's causing it so that I can find what causes it tomorrow, etc.
$ grep "model name" /proc/cpuinfo
model name : Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz # 8 times
$ sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
turbostat version 17.06.23 - Len Brown <lenb@kernel.org>
CPUID(0): GenuineIntel 22 CPUID levels; family:model:stepping 0x6:9e:9 (6:158:9)
CPUID(1): SSE3 MONITOR - EIST TM2 TSC MSR ACPI-TM TM
CPUID(6): APERF, TURBO, DTS, PTM, HWP, HWPnotify, HWPwindow, HWPepp, No-HWPpkg, EPB
cpu0: MSR_IA32_MISC_ENABLE: 0x00850089 (TCC EIST No-MWAIT PREFETCH TURBO)
CPUID(7): SGX
cpu0: MSR_IA32_FEATURE_CONTROL: 0x00000005 (Locked )
CPUID(0x15): eax_crystal: 2 ebx_tsc: 350 ecx_crystal_hz: 0
TSC: 4200 MHz (24000000 Hz * 350 / 2 / 1000000)
CPUID(0x16): base_mhz: 4200 max_mhz: 4800 bus_mhz: 100
cpu0: MSR_MISC_PWR_MGMT: 0x00401cc0 (ENable-EIST_Coordination DISable-EPB DISable-OOB)
RAPL: 2881 sec. Joule Counter Range, at 91 Watts
cpu0: MSR_PLATFORM_INFO: 0x80838f1012a00
8 * 100.0 = 800.0 MHz max efficiency frequency
42 * 100.0 = 4200.0 MHz base frequency
cpu0: MSR_IA32_POWER_CTL: 0x003c005f (C1E auto-promotion: ENabled)
cpu0: MSR_TURBO_RATIO_LIMIT: 0x30303030
48 * 100.0 = 4800.0 MHz max turbo 4 active cores
48 * 100.0 = 4800.0 MHz max turbo 3 active cores
48 * 100.0 = 4800.0 MHz max turbo 2 active cores
48 * 100.0 = 4800.0 MHz max turbo 1 active cores
cpu0: MSR_CONFIG_TDP_NOMINAL: 0x0000002a (base_ratio=42)
cpu0: MSR_CONFIG_TDP_LEVEL_1: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_LEVEL_2: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_CONTROL: 0x80000000 ( lock=1)
cpu0: MSR_TURBO_ACTIVATION_RATIO: 0x00000000 (MAX_NON_TURBO_RATIO=0 lock=0)
cpu0: MSR_PKG_CST_CONFIG_CONTROL: 0x1e000000 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, UNlocked: pkg-cstate-limit=0: pc0)
cpu0: cpufreq driver: intel_pstate
cpu0: cpufreq governor: powersave
cpufreq intel_pstate no_turbo: 0
cpu0: MSR_MISC_FEATURE_CONTROL: 0x00000000 (L2-Prefetch L2-Prefetch-pair L1-Prefetch L1-IP-Prefetch)
cpu0: MSR_PM_ENABLE: 0x00000001 (HWP)
cpu0: MSR_HWP_CAPABILITIES: 0x010a2a30 (high 48 guar 42 eff 10 low 1)
cpu0: MSR_HWP_REQUEST: 0x80003008 (min 8 max 48 des 0 epp 0x80 window 0x0 pkg 0x0)
cpu0: MSR_HWP_INTERRUPT: 0x00000000 (Dis_Guaranteed_Perf_Change, Dis_Excursion_Min)
cpu0: MSR_HWP_STATUS: 0x00000004 (No-Guaranteed_Perf_Change, No-Excursion_Min)
cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced)
cpu0: MSR_RAPL_POWER_UNIT: 0x000a0e03 (0.125000 Watts, 0.000061 Joules, 0.000977 sec.)
cpu0: MSR_PKG_POWER_INFO: 0x000002d8 (91 W TDP, RAPL 0 - 0 W, 0.000000 sec.)
cpu0: MSR_PKG_POWER_LIMIT: 0x42ffff001bffff (UNlocked)
cpu0: PKG Limit #1: ENabled (4095.875000 Watts, 8.000000 sec, clamp ENabled)
cpu0: PKG Limit #2: ENabled (4095.875000 Watts, 0.002441* sec, clamp DISabled)
cpu0: MSR_DRAM_POWER_LIMIT: 0x5400de00000000 (UNlocked)
cpu0: DRAM Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP0_POLICY: 0
cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP1_POLICY: 0
cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x0064140d (100 C)
cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x88440800 (32 C)
cpu0: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x00000003 (100 C, 100 C)
cpu0: MSR_PKGC3_IRTL: 0x0000884e (valid, 79872 ns)
cpu0: MSR_PKGC6_IRTL: 0x00008876 (valid, 120832 ns)
cpu0: MSR_PKGC7_IRTL: 0x00008894 (valid, 151552 ns)
cpu0: MSR_PKGC8_IRTL: 0x000088fa (valid, 256000 ns)
cpu0: MSR_PKGC9_IRTL: 0x0000894c (valid, 339968 ns)
cpu0: MSR_PKGC10_IRTL: 0x00008bf2 (valid, 1034240 ns)
Busy% Bzy_MHz PkgTmp PkgWatt
0.21 4359 27 5.43
0.22 4509 27 5.41
And the busiest section of a perf report:
Children Self Command Shared Object Symbol
+ 51.83% 0.00% swapper [kernel.kallsyms] [k] secondary_startup_64
+ 51.75% 0.17% swapper [kernel.kallsyms] [k] cpu_startup_entry
+ 51.36% 0.54% swapper [kernel.kallsyms] [k] do_idle
+ 47.36% 0.00% swapper [kernel.kallsyms] [k] start_secondary
+ 45.49% 0.08% swapper [kernel.kallsyms] [k] call_cpuidle
+ 45.42% 0.01% swapper [kernel.kallsyms] [k] cpuidle_enter
+ 45.10% 0.47% swapper [kernel.kallsyms] [k] cpuidle_enter_state
+ 39.05% 39.05% swapper [kernel.kallsyms] [k] intel_idle
+ 7.41% 7.41% swapper [kernel.kallsyms] [k] _nv031472rm
+ 4.66% 0.06% swapper [kernel.kallsyms] [k] apic_timer_interrupt
+ 4.56% 0.04% swapper [kernel.kallsyms] [k] smp_apic_timer_interrupt
I've no idea how to interpret this. The top results for this tell me it's about paging to swap but I'm fairly sure (8GB RAM free, and it still happens after I turn swap off completely) it's not that at all.
The above was run just sitting at a gnome desktop with most of my extensions turned off, browser closed. Idle.
Found a similar recent question where —for some reason or other— they were stuck on performance with a borked minimum speed. That does not appear to be the case here.
$ cd /sys/devices/system/cpu/cpu0/cpufreq && paste <(ls *) <(cat *)
affected_cpus 0
cpuinfo_max_freq 4800000
cpuinfo_min_freq 800000
cpuinfo_transition_latency 0
energy_performance_available_preferences default performance balance_performance balance_power power
energy_performance_preference balance_performance
related_cpus 0
scaling_available_governors performance powersave
scaling_cur_freq 4787854
scaling_driver intel_pstate
scaling_governor powersave
scaling_max_freq 4800000
scaling_min_freq 800000
scaling_setspeed <unsupported>
power-management cpu governor
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
closed as off-topic by Oli♦ Jan 11 at 9:25
This question appears to be off-topic. The users who voted to close gave this specific reason:
- "This describes a problem that can't be reproduced, that seemingly went away on its own or was only relevant to a very specific period of time. It's off-topic as it's unlikely to help future readers." – Oli
If this question can be reworded to fit the rules in the help center, please edit the question.
|
show 4 more comments
1
My CPU scales from 800MHz to 4.8GHz, obviously in a vastly different power envelope. I have an applet that tells me what speed it's running at on my panel and today I've noticed that it's running pretty close to the top end all the time. Even when it's essentially idle.
$ uptime
22:05:34 up 1 day, 5:37, 1 user, load average: 0.00, 0.00, 0.00I'm running on the "powersave" governor (intel_pstate driver).
So —short of closing random processes until it tweaks down— how do I go about finding what's stopping my CPU from underclocking?
Some follow up from the comments.
It's important to note that this system has scaled down in recent history, while on 18.04 (which is what I'm still using). There haven't been any major upgrades, save whatever Nvidia do to their driver. This is something new, perhaps just something this session. That's kind of my point of this question. Even if it fixes it, I don't just want to reboot. I want to find what's causing it so that I can find what causes it tomorrow, etc.
$ grep "model name" /proc/cpuinfo
model name : Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz # 8 times
$ sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
turbostat version 17.06.23 - Len Brown <lenb@kernel.org>
CPUID(0): GenuineIntel 22 CPUID levels; family:model:stepping 0x6:9e:9 (6:158:9)
CPUID(1): SSE3 MONITOR - EIST TM2 TSC MSR ACPI-TM TM
CPUID(6): APERF, TURBO, DTS, PTM, HWP, HWPnotify, HWPwindow, HWPepp, No-HWPpkg, EPB
cpu0: MSR_IA32_MISC_ENABLE: 0x00850089 (TCC EIST No-MWAIT PREFETCH TURBO)
CPUID(7): SGX
cpu0: MSR_IA32_FEATURE_CONTROL: 0x00000005 (Locked )
CPUID(0x15): eax_crystal: 2 ebx_tsc: 350 ecx_crystal_hz: 0
TSC: 4200 MHz (24000000 Hz * 350 / 2 / 1000000)
CPUID(0x16): base_mhz: 4200 max_mhz: 4800 bus_mhz: 100
cpu0: MSR_MISC_PWR_MGMT: 0x00401cc0 (ENable-EIST_Coordination DISable-EPB DISable-OOB)
RAPL: 2881 sec. Joule Counter Range, at 91 Watts
cpu0: MSR_PLATFORM_INFO: 0x80838f1012a00
8 * 100.0 = 800.0 MHz max efficiency frequency
42 * 100.0 = 4200.0 MHz base frequency
cpu0: MSR_IA32_POWER_CTL: 0x003c005f (C1E auto-promotion: ENabled)
cpu0: MSR_TURBO_RATIO_LIMIT: 0x30303030
48 * 100.0 = 4800.0 MHz max turbo 4 active cores
48 * 100.0 = 4800.0 MHz max turbo 3 active cores
48 * 100.0 = 4800.0 MHz max turbo 2 active cores
48 * 100.0 = 4800.0 MHz max turbo 1 active cores
cpu0: MSR_CONFIG_TDP_NOMINAL: 0x0000002a (base_ratio=42)
cpu0: MSR_CONFIG_TDP_LEVEL_1: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_LEVEL_2: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_CONTROL: 0x80000000 ( lock=1)
cpu0: MSR_TURBO_ACTIVATION_RATIO: 0x00000000 (MAX_NON_TURBO_RATIO=0 lock=0)
cpu0: MSR_PKG_CST_CONFIG_CONTROL: 0x1e000000 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, UNlocked: pkg-cstate-limit=0: pc0)
cpu0: cpufreq driver: intel_pstate
cpu0: cpufreq governor: powersave
cpufreq intel_pstate no_turbo: 0
cpu0: MSR_MISC_FEATURE_CONTROL: 0x00000000 (L2-Prefetch L2-Prefetch-pair L1-Prefetch L1-IP-Prefetch)
cpu0: MSR_PM_ENABLE: 0x00000001 (HWP)
cpu0: MSR_HWP_CAPABILITIES: 0x010a2a30 (high 48 guar 42 eff 10 low 1)
cpu0: MSR_HWP_REQUEST: 0x80003008 (min 8 max 48 des 0 epp 0x80 window 0x0 pkg 0x0)
cpu0: MSR_HWP_INTERRUPT: 0x00000000 (Dis_Guaranteed_Perf_Change, Dis_Excursion_Min)
cpu0: MSR_HWP_STATUS: 0x00000004 (No-Guaranteed_Perf_Change, No-Excursion_Min)
cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced)
cpu0: MSR_RAPL_POWER_UNIT: 0x000a0e03 (0.125000 Watts, 0.000061 Joules, 0.000977 sec.)
cpu0: MSR_PKG_POWER_INFO: 0x000002d8 (91 W TDP, RAPL 0 - 0 W, 0.000000 sec.)
cpu0: MSR_PKG_POWER_LIMIT: 0x42ffff001bffff (UNlocked)
cpu0: PKG Limit #1: ENabled (4095.875000 Watts, 8.000000 sec, clamp ENabled)
cpu0: PKG Limit #2: ENabled (4095.875000 Watts, 0.002441* sec, clamp DISabled)
cpu0: MSR_DRAM_POWER_LIMIT: 0x5400de00000000 (UNlocked)
cpu0: DRAM Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP0_POLICY: 0
cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP1_POLICY: 0
cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x0064140d (100 C)
cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x88440800 (32 C)
cpu0: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x00000003 (100 C, 100 C)
cpu0: MSR_PKGC3_IRTL: 0x0000884e (valid, 79872 ns)
cpu0: MSR_PKGC6_IRTL: 0x00008876 (valid, 120832 ns)
cpu0: MSR_PKGC7_IRTL: 0x00008894 (valid, 151552 ns)
cpu0: MSR_PKGC8_IRTL: 0x000088fa (valid, 256000 ns)
cpu0: MSR_PKGC9_IRTL: 0x0000894c (valid, 339968 ns)
cpu0: MSR_PKGC10_IRTL: 0x00008bf2 (valid, 1034240 ns)
Busy% Bzy_MHz PkgTmp PkgWatt
0.21 4359 27 5.43
0.22 4509 27 5.41
And the busiest section of a perf report:
Children Self Command Shared Object Symbol
+ 51.83% 0.00% swapper [kernel.kallsyms] [k] secondary_startup_64
+ 51.75% 0.17% swapper [kernel.kallsyms] [k] cpu_startup_entry
+ 51.36% 0.54% swapper [kernel.kallsyms] [k] do_idle
+ 47.36% 0.00% swapper [kernel.kallsyms] [k] start_secondary
+ 45.49% 0.08% swapper [kernel.kallsyms] [k] call_cpuidle
+ 45.42% 0.01% swapper [kernel.kallsyms] [k] cpuidle_enter
+ 45.10% 0.47% swapper [kernel.kallsyms] [k] cpuidle_enter_state
+ 39.05% 39.05% swapper [kernel.kallsyms] [k] intel_idle
+ 7.41% 7.41% swapper [kernel.kallsyms] [k] _nv031472rm
+ 4.66% 0.06% swapper [kernel.kallsyms] [k] apic_timer_interrupt
+ 4.56% 0.04% swapper [kernel.kallsyms] [k] smp_apic_timer_interrupt
I've no idea how to interpret this. The top results for this tell me it's about paging to swap but I'm fairly sure (8GB RAM free, and it still happens after I turn swap off completely) it's not that at all.
The above was run just sitting at a gnome desktop with most of my extensions turned off, browser closed. Idle.
Found a similar recent question where —for some reason or other— they were stuck on performance with a borked minimum speed. That does not appear to be the case here.
$ cd /sys/devices/system/cpu/cpu0/cpufreq && paste <(ls *) <(cat *)
affected_cpus 0
cpuinfo_max_freq 4800000
cpuinfo_min_freq 800000
cpuinfo_transition_latency 0
energy_performance_available_preferences default performance balance_performance balance_power power
energy_performance_preference balance_performance
related_cpus 0
scaling_available_governors performance powersave
scaling_cur_freq 4787854
scaling_driver intel_pstate
scaling_governor powersave
scaling_max_freq 4800000
scaling_min_freq 800000
scaling_setspeed <unsupported>
power-management cpu governor
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
closed as off-topic by Oli♦ Jan 11 at 9:25
This question appears to be off-topic. The users who voted to close gave this specific reason:
- "This describes a problem that can't be reproduced, that seemingly went away on its own or was only relevant to a very specific period of time. It's off-topic as it's unlikely to help future readers." – Oli
If this question can be reworded to fit the rules in the help center, please edit the question.
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do:sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
– Doug Smythies
Jan 9 at 7:46
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and theperfoutput doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.
– Oli♦
Jan 9 at 10:32
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24
|
show 4 more comments
1
1
1
My CPU scales from 800MHz to 4.8GHz, obviously in a vastly different power envelope. I have an applet that tells me what speed it's running at on my panel and today I've noticed that it's running pretty close to the top end all the time. Even when it's essentially idle.
$ uptime
22:05:34 up 1 day, 5:37, 1 user, load average: 0.00, 0.00, 0.00I'm running on the "powersave" governor (intel_pstate driver).
So —short of closing random processes until it tweaks down— how do I go about finding what's stopping my CPU from underclocking?
Some follow up from the comments.
It's important to note that this system has scaled down in recent history, while on 18.04 (which is what I'm still using). There haven't been any major upgrades, save whatever Nvidia do to their driver. This is something new, perhaps just something this session. That's kind of my point of this question. Even if it fixes it, I don't just want to reboot. I want to find what's causing it so that I can find what causes it tomorrow, etc.
$ grep "model name" /proc/cpuinfo
model name : Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz # 8 times
$ sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
turbostat version 17.06.23 - Len Brown <lenb@kernel.org>
CPUID(0): GenuineIntel 22 CPUID levels; family:model:stepping 0x6:9e:9 (6:158:9)
CPUID(1): SSE3 MONITOR - EIST TM2 TSC MSR ACPI-TM TM
CPUID(6): APERF, TURBO, DTS, PTM, HWP, HWPnotify, HWPwindow, HWPepp, No-HWPpkg, EPB
cpu0: MSR_IA32_MISC_ENABLE: 0x00850089 (TCC EIST No-MWAIT PREFETCH TURBO)
CPUID(7): SGX
cpu0: MSR_IA32_FEATURE_CONTROL: 0x00000005 (Locked )
CPUID(0x15): eax_crystal: 2 ebx_tsc: 350 ecx_crystal_hz: 0
TSC: 4200 MHz (24000000 Hz * 350 / 2 / 1000000)
CPUID(0x16): base_mhz: 4200 max_mhz: 4800 bus_mhz: 100
cpu0: MSR_MISC_PWR_MGMT: 0x00401cc0 (ENable-EIST_Coordination DISable-EPB DISable-OOB)
RAPL: 2881 sec. Joule Counter Range, at 91 Watts
cpu0: MSR_PLATFORM_INFO: 0x80838f1012a00
8 * 100.0 = 800.0 MHz max efficiency frequency
42 * 100.0 = 4200.0 MHz base frequency
cpu0: MSR_IA32_POWER_CTL: 0x003c005f (C1E auto-promotion: ENabled)
cpu0: MSR_TURBO_RATIO_LIMIT: 0x30303030
48 * 100.0 = 4800.0 MHz max turbo 4 active cores
48 * 100.0 = 4800.0 MHz max turbo 3 active cores
48 * 100.0 = 4800.0 MHz max turbo 2 active cores
48 * 100.0 = 4800.0 MHz max turbo 1 active cores
cpu0: MSR_CONFIG_TDP_NOMINAL: 0x0000002a (base_ratio=42)
cpu0: MSR_CONFIG_TDP_LEVEL_1: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_LEVEL_2: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_CONTROL: 0x80000000 ( lock=1)
cpu0: MSR_TURBO_ACTIVATION_RATIO: 0x00000000 (MAX_NON_TURBO_RATIO=0 lock=0)
cpu0: MSR_PKG_CST_CONFIG_CONTROL: 0x1e000000 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, UNlocked: pkg-cstate-limit=0: pc0)
cpu0: cpufreq driver: intel_pstate
cpu0: cpufreq governor: powersave
cpufreq intel_pstate no_turbo: 0
cpu0: MSR_MISC_FEATURE_CONTROL: 0x00000000 (L2-Prefetch L2-Prefetch-pair L1-Prefetch L1-IP-Prefetch)
cpu0: MSR_PM_ENABLE: 0x00000001 (HWP)
cpu0: MSR_HWP_CAPABILITIES: 0x010a2a30 (high 48 guar 42 eff 10 low 1)
cpu0: MSR_HWP_REQUEST: 0x80003008 (min 8 max 48 des 0 epp 0x80 window 0x0 pkg 0x0)
cpu0: MSR_HWP_INTERRUPT: 0x00000000 (Dis_Guaranteed_Perf_Change, Dis_Excursion_Min)
cpu0: MSR_HWP_STATUS: 0x00000004 (No-Guaranteed_Perf_Change, No-Excursion_Min)
cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced)
cpu0: MSR_RAPL_POWER_UNIT: 0x000a0e03 (0.125000 Watts, 0.000061 Joules, 0.000977 sec.)
cpu0: MSR_PKG_POWER_INFO: 0x000002d8 (91 W TDP, RAPL 0 - 0 W, 0.000000 sec.)
cpu0: MSR_PKG_POWER_LIMIT: 0x42ffff001bffff (UNlocked)
cpu0: PKG Limit #1: ENabled (4095.875000 Watts, 8.000000 sec, clamp ENabled)
cpu0: PKG Limit #2: ENabled (4095.875000 Watts, 0.002441* sec, clamp DISabled)
cpu0: MSR_DRAM_POWER_LIMIT: 0x5400de00000000 (UNlocked)
cpu0: DRAM Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP0_POLICY: 0
cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP1_POLICY: 0
cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x0064140d (100 C)
cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x88440800 (32 C)
cpu0: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x00000003 (100 C, 100 C)
cpu0: MSR_PKGC3_IRTL: 0x0000884e (valid, 79872 ns)
cpu0: MSR_PKGC6_IRTL: 0x00008876 (valid, 120832 ns)
cpu0: MSR_PKGC7_IRTL: 0x00008894 (valid, 151552 ns)
cpu0: MSR_PKGC8_IRTL: 0x000088fa (valid, 256000 ns)
cpu0: MSR_PKGC9_IRTL: 0x0000894c (valid, 339968 ns)
cpu0: MSR_PKGC10_IRTL: 0x00008bf2 (valid, 1034240 ns)
Busy% Bzy_MHz PkgTmp PkgWatt
0.21 4359 27 5.43
0.22 4509 27 5.41
And the busiest section of a perf report:
Children Self Command Shared Object Symbol
+ 51.83% 0.00% swapper [kernel.kallsyms] [k] secondary_startup_64
+ 51.75% 0.17% swapper [kernel.kallsyms] [k] cpu_startup_entry
+ 51.36% 0.54% swapper [kernel.kallsyms] [k] do_idle
+ 47.36% 0.00% swapper [kernel.kallsyms] [k] start_secondary
+ 45.49% 0.08% swapper [kernel.kallsyms] [k] call_cpuidle
+ 45.42% 0.01% swapper [kernel.kallsyms] [k] cpuidle_enter
+ 45.10% 0.47% swapper [kernel.kallsyms] [k] cpuidle_enter_state
+ 39.05% 39.05% swapper [kernel.kallsyms] [k] intel_idle
+ 7.41% 7.41% swapper [kernel.kallsyms] [k] _nv031472rm
+ 4.66% 0.06% swapper [kernel.kallsyms] [k] apic_timer_interrupt
+ 4.56% 0.04% swapper [kernel.kallsyms] [k] smp_apic_timer_interrupt
I've no idea how to interpret this. The top results for this tell me it's about paging to swap but I'm fairly sure (8GB RAM free, and it still happens after I turn swap off completely) it's not that at all.
The above was run just sitting at a gnome desktop with most of my extensions turned off, browser closed. Idle.
Found a similar recent question where —for some reason or other— they were stuck on performance with a borked minimum speed. That does not appear to be the case here.
$ cd /sys/devices/system/cpu/cpu0/cpufreq && paste <(ls *) <(cat *)
affected_cpus 0
cpuinfo_max_freq 4800000
cpuinfo_min_freq 800000
cpuinfo_transition_latency 0
energy_performance_available_preferences default performance balance_performance balance_power power
energy_performance_preference balance_performance
related_cpus 0
scaling_available_governors performance powersave
scaling_cur_freq 4787854
scaling_driver intel_pstate
scaling_governor powersave
scaling_max_freq 4800000
scaling_min_freq 800000
scaling_setspeed <unsupported>
power-management cpu governor
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
My CPU scales from 800MHz to 4.8GHz, obviously in a vastly different power envelope. I have an applet that tells me what speed it's running at on my panel and today I've noticed that it's running pretty close to the top end all the time. Even when it's essentially idle.
$ uptime
22:05:34 up 1 day, 5:37, 1 user, load average: 0.00, 0.00, 0.00I'm running on the "powersave" governor (intel_pstate driver).
So —short of closing random processes until it tweaks down— how do I go about finding what's stopping my CPU from underclocking?
Some follow up from the comments.
It's important to note that this system has scaled down in recent history, while on 18.04 (which is what I'm still using). There haven't been any major upgrades, save whatever Nvidia do to their driver. This is something new, perhaps just something this session. That's kind of my point of this question. Even if it fixes it, I don't just want to reboot. I want to find what's causing it so that I can find what causes it tomorrow, etc.
$ grep "model name" /proc/cpuinfo
model name : Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz # 8 times
$ sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
turbostat version 17.06.23 - Len Brown <lenb@kernel.org>
CPUID(0): GenuineIntel 22 CPUID levels; family:model:stepping 0x6:9e:9 (6:158:9)
CPUID(1): SSE3 MONITOR - EIST TM2 TSC MSR ACPI-TM TM
CPUID(6): APERF, TURBO, DTS, PTM, HWP, HWPnotify, HWPwindow, HWPepp, No-HWPpkg, EPB
cpu0: MSR_IA32_MISC_ENABLE: 0x00850089 (TCC EIST No-MWAIT PREFETCH TURBO)
CPUID(7): SGX
cpu0: MSR_IA32_FEATURE_CONTROL: 0x00000005 (Locked )
CPUID(0x15): eax_crystal: 2 ebx_tsc: 350 ecx_crystal_hz: 0
TSC: 4200 MHz (24000000 Hz * 350 / 2 / 1000000)
CPUID(0x16): base_mhz: 4200 max_mhz: 4800 bus_mhz: 100
cpu0: MSR_MISC_PWR_MGMT: 0x00401cc0 (ENable-EIST_Coordination DISable-EPB DISable-OOB)
RAPL: 2881 sec. Joule Counter Range, at 91 Watts
cpu0: MSR_PLATFORM_INFO: 0x80838f1012a00
8 * 100.0 = 800.0 MHz max efficiency frequency
42 * 100.0 = 4200.0 MHz base frequency
cpu0: MSR_IA32_POWER_CTL: 0x003c005f (C1E auto-promotion: ENabled)
cpu0: MSR_TURBO_RATIO_LIMIT: 0x30303030
48 * 100.0 = 4800.0 MHz max turbo 4 active cores
48 * 100.0 = 4800.0 MHz max turbo 3 active cores
48 * 100.0 = 4800.0 MHz max turbo 2 active cores
48 * 100.0 = 4800.0 MHz max turbo 1 active cores
cpu0: MSR_CONFIG_TDP_NOMINAL: 0x0000002a (base_ratio=42)
cpu0: MSR_CONFIG_TDP_LEVEL_1: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_LEVEL_2: 0x00000000 ()
cpu0: MSR_CONFIG_TDP_CONTROL: 0x80000000 ( lock=1)
cpu0: MSR_TURBO_ACTIVATION_RATIO: 0x00000000 (MAX_NON_TURBO_RATIO=0 lock=0)
cpu0: MSR_PKG_CST_CONFIG_CONTROL: 0x1e000000 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, UNlocked: pkg-cstate-limit=0: pc0)
cpu0: cpufreq driver: intel_pstate
cpu0: cpufreq governor: powersave
cpufreq intel_pstate no_turbo: 0
cpu0: MSR_MISC_FEATURE_CONTROL: 0x00000000 (L2-Prefetch L2-Prefetch-pair L1-Prefetch L1-IP-Prefetch)
cpu0: MSR_PM_ENABLE: 0x00000001 (HWP)
cpu0: MSR_HWP_CAPABILITIES: 0x010a2a30 (high 48 guar 42 eff 10 low 1)
cpu0: MSR_HWP_REQUEST: 0x80003008 (min 8 max 48 des 0 epp 0x80 window 0x0 pkg 0x0)
cpu0: MSR_HWP_INTERRUPT: 0x00000000 (Dis_Guaranteed_Perf_Change, Dis_Excursion_Min)
cpu0: MSR_HWP_STATUS: 0x00000004 (No-Guaranteed_Perf_Change, No-Excursion_Min)
cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced)
cpu0: MSR_RAPL_POWER_UNIT: 0x000a0e03 (0.125000 Watts, 0.000061 Joules, 0.000977 sec.)
cpu0: MSR_PKG_POWER_INFO: 0x000002d8 (91 W TDP, RAPL 0 - 0 W, 0.000000 sec.)
cpu0: MSR_PKG_POWER_LIMIT: 0x42ffff001bffff (UNlocked)
cpu0: PKG Limit #1: ENabled (4095.875000 Watts, 8.000000 sec, clamp ENabled)
cpu0: PKG Limit #2: ENabled (4095.875000 Watts, 0.002441* sec, clamp DISabled)
cpu0: MSR_DRAM_POWER_LIMIT: 0x5400de00000000 (UNlocked)
cpu0: DRAM Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP0_POLICY: 0
cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_PP1_POLICY: 0
cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked)
cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x0064140d (100 C)
cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x88440800 (32 C)
cpu0: MSR_IA32_PACKAGE_THERM_INTERRUPT: 0x00000003 (100 C, 100 C)
cpu0: MSR_PKGC3_IRTL: 0x0000884e (valid, 79872 ns)
cpu0: MSR_PKGC6_IRTL: 0x00008876 (valid, 120832 ns)
cpu0: MSR_PKGC7_IRTL: 0x00008894 (valid, 151552 ns)
cpu0: MSR_PKGC8_IRTL: 0x000088fa (valid, 256000 ns)
cpu0: MSR_PKGC9_IRTL: 0x0000894c (valid, 339968 ns)
cpu0: MSR_PKGC10_IRTL: 0x00008bf2 (valid, 1034240 ns)
Busy% Bzy_MHz PkgTmp PkgWatt
0.21 4359 27 5.43
0.22 4509 27 5.41
And the busiest section of a perf report:
Children Self Command Shared Object Symbol
+ 51.83% 0.00% swapper [kernel.kallsyms] [k] secondary_startup_64
+ 51.75% 0.17% swapper [kernel.kallsyms] [k] cpu_startup_entry
+ 51.36% 0.54% swapper [kernel.kallsyms] [k] do_idle
+ 47.36% 0.00% swapper [kernel.kallsyms] [k] start_secondary
+ 45.49% 0.08% swapper [kernel.kallsyms] [k] call_cpuidle
+ 45.42% 0.01% swapper [kernel.kallsyms] [k] cpuidle_enter
+ 45.10% 0.47% swapper [kernel.kallsyms] [k] cpuidle_enter_state
+ 39.05% 39.05% swapper [kernel.kallsyms] [k] intel_idle
+ 7.41% 7.41% swapper [kernel.kallsyms] [k] _nv031472rm
+ 4.66% 0.06% swapper [kernel.kallsyms] [k] apic_timer_interrupt
+ 4.56% 0.04% swapper [kernel.kallsyms] [k] smp_apic_timer_interrupt
I've no idea how to interpret this. The top results for this tell me it's about paging to swap but I'm fairly sure (8GB RAM free, and it still happens after I turn swap off completely) it's not that at all.
The above was run just sitting at a gnome desktop with most of my extensions turned off, browser closed. Idle.
Found a similar recent question where —for some reason or other— they were stuck on performance with a borked minimum speed. That does not appear to be the case here.
$ cd /sys/devices/system/cpu/cpu0/cpufreq && paste <(ls *) <(cat *)
affected_cpus 0
cpuinfo_max_freq 4800000
cpuinfo_min_freq 800000
cpuinfo_transition_latency 0
energy_performance_available_preferences default performance balance_performance balance_power power
energy_performance_preference balance_performance
related_cpus 0
scaling_available_governors performance powersave
scaling_cur_freq 4787854
scaling_driver intel_pstate
scaling_governor powersave
scaling_max_freq 4800000
scaling_min_freq 800000
scaling_setspeed <unsupported>
power-management cpu governor
power-management cpu governor
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
edited Jan 9 at 14:41
Oli
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
asked Jan 8 at 22:11
Oli♦Oli
221k86561763
221k86561763
closed as off-topic by Oli♦ Jan 11 at 9:25
This question appears to be off-topic. The users who voted to close gave this specific reason:
- "This describes a problem that can't be reproduced, that seemingly went away on its own or was only relevant to a very specific period of time. It's off-topic as it's unlikely to help future readers." – Oli
If this question can be reworded to fit the rules in the help center, please edit the question.
closed as off-topic by Oli♦ Jan 11 at 9:25
This question appears to be off-topic. The users who voted to close gave this specific reason:
- "This describes a problem that can't be reproduced, that seemingly went away on its own or was only relevant to a very specific period of time. It's off-topic as it's unlikely to help future readers." – Oli
If this question can be reworded to fit the rules in the help center, please edit the question.
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do:sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
– Doug Smythies
Jan 9 at 7:46
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and theperfoutput doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.
– Oli♦
Jan 9 at 10:32
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24
|
show 4 more comments
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do:sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15
– Doug Smythies
Jan 9 at 7:46
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and theperfoutput doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.
– Oli♦
Jan 9 at 10:32
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:
grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do: sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15– Doug Smythies
Jan 9 at 7:46
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:
grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do: sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15– Doug Smythies
Jan 9 at 7:46
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and the
perf output doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.– Oli♦
Jan 9 at 10:32
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and the
perf output doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.– Oli♦
Jan 9 at 10:32
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24
|
show 4 more comments
0
active
oldest
votes
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
Possible bug: bugs.launchpad.net/ubuntu/+source/linux/+bug/1579278
– Terrance
Jan 9 at 4:59
What is the sample rate of your applet? Try no more than 1 sample every 15 seconds. What Intel processor model? do:
grep "model name" /proc/cpuinfo. turbostat (linux-tools-common) is a suggested place to start.Do:sudo turbostat --Summary --show Busy%,Bzy_MHz,PkgTmp,PkgWatt --interval 15– Doug Smythies
Jan 9 at 7:46
I've added some more data. Doug, the sample rate was 500ms but turning it down to 15000ms did nothing to slow it down and even turning it off entirely did nothing to improve things. I've added turbostat output. Terrance, I played around with the governor (switched to performance from powersave) but that didn't help and the
perfoutput doesn't make much sense to me though it may well be related to that bug. The thing that makes me think it isn't related to that is that this has been scaling pretty well up to now.– Oli♦
Jan 9 at 10:32
Yeah, that's the stupid thing about bugs. They appear when we least expect them. askubuntu.com/q/947884/231142 is an example of a bug that appeared when my system was working just fine. It took a format and reinstallation that fixed it. I guess you could try LiveCD/USB with persistence and try to duplicate the best you can of your environment to see if that makes any difference. This is only just thoughts.
– Terrance
Jan 9 at 14:22
I have similar problems with Kernel 4.14.78. started last year some time.
– WinEunuuchs2Unix
Jan 9 at 14:24