3 turbostat \- Report processor frequency and idle statistics
12 .RB [ "\--interval seconds" ]
14 \fBturbostat \fP reports processor topology, frequency,
15 idle power-state statistics, temperature and power on X86 processors.
16 There are two ways to invoke turbostat.
17 The first method is to supply a
18 \fBcommand\fP, which is forked and statistics are printed
20 The second method is to omit the command,
21 and turbostat displays statistics every 5 seconds.
22 The 5-second interval can be changed using the --interval option.
24 Some information is not available on older processors.
26 Options can be specified with a single or double '-', and only as much of the option
27 name as necessary to disambiguate it from others is necessary. Note that options are case-sensitive.
28 \fB--Counter MSR#\fP shows the delta of the specified 64-bit MSR counter.
30 \fB--counter MSR#\fP shows the delta of the specified 32-bit MSR counter.
32 \fB--Dump\fP displays the raw counter values.
34 \fB--debug\fP displays additional system configuration information. Invoking this parameter
35 more than once may also enable internal turbostat debug information.
37 \fB--interval seconds\fP overrides the default 5-second measurement interval.
39 \fB--help\fP displays usage for the most common parameters.
41 \fB--Joules\fP displays energy in Joules, rather than dividing Joules by time to print power in Watts.
43 \fB--MSR MSR#\fP shows the specified 64-bit MSR value.
45 \fB--msr MSR#\fP shows the specified 32-bit MSR value.
47 \fB--Package\fP limits output to the system summary plus the 1st thread in each Package.
49 \fB--processor\fP limits output to the system summary plus the 1st thread in each processor of each package. Ie. it skips hyper-threaded siblings.
51 \fB--Summary\fP limits output to a 1-line System Summary for each interval.
53 \fB--TCC temperature\fP sets the Thermal Control Circuit temperature for systems which do not export that value. This is used for making sense of the Digital Thermal Sensor outputs, as they return degrees Celsius below the TCC activation temperature.
55 \fB--version\fP displays the version.
57 The \fBcommand\fP parameter forks \fBcommand\fP, and upon its exit,
58 displays the statistics gathered since it was forked.
60 .SH DEFAULT FIELD DESCRIPTIONS
62 \fBCPU\fP Linux CPU (logical processor) number. Yes, it is okay that on many systems the CPUs are not listed in numerical order -- for efficiency reasons, turbostat runs in topology order, so HT siblings appear together.
63 \fBAVG_MHz\fP number of cycles executed divided by time elapsed.
64 \fB%Busy\fP percent of the interval that the CPU retired instructions, aka. % of time in "C0" state.
65 \fBBzy_MHz\fP average clock rate while the CPU was busy (in "c0" state).
66 \fBTSC_MHz\fP average MHz that the TSC ran during the entire interval.
69 .SH DEBUG FIELD DESCRIPTIONS
71 \fBPackage\fP processor package number.
72 \fBCore\fP processor core number.
73 Note that multiple CPUs per core indicate support for Intel(R) Hyper-Threading Technology (HT).
74 \fBCPU%c1, CPU%c3, CPU%c6, CPU%c7\fP show the percentage residency in hardware core idle states.
75 \fBCoreTmp\fP Degrees Celsius reported by the per-core Digital Thermal Sensor.
76 \fBPkgTtmp\fP Degrees Celsius reported by the per-package Package Thermal Monitor.
77 \fBPkg%pc2, Pkg%pc3, Pkg%pc6, Pkg%pc7\fP percentage residency in hardware package idle states.
78 \fBPkgWatt\fP Watts consumed by the whole package.
79 \fBCorWatt\fP Watts consumed by the core part of the package.
80 \fBGFXWatt\fP Watts consumed by the Graphics part of the package -- available only on client processors.
81 \fBRAMWatt\fP Watts consumed by the DRAM DIMMS -- available only on server processors.
82 \fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package.
83 \fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM.
87 Without any parameters, turbostat displays statistics ever 5 seconds.
88 (override interval with "-i sec" option, or specify a command
89 for turbostat to fork).
91 [root@hsw]# ./turbostat
92 CPU Avg_MHz %Busy Bzy_MHz TSC_MHz
95 4 3897 99.99 3898 3498
105 The "--debug" option prints additional system information before measurements:
107 The first row of statistics is a summary for the entire system.
108 For residency % columns, the summary is a weighted average.
109 For Temperature columns, the summary is the column maximum.
110 For Watts columns, the summary is a system total.
111 Subsequent rows show per-CPU statistics.
113 turbostat version 4.1 10-Feb, 2015 - Len Brown <lenb@kernel.org>
114 CPUID(0): GenuineIntel 13 CPUID levels; family:model:stepping 0x6:3c:3 (6:60:3)
115 CPUID(6): APERF, DTS, PTM, EPB
116 RAPL: 3121 sec. Joule Counter Range, at 84 Watts
117 cpu0: MSR_NHM_PLATFORM_INFO: 0x80838f3012300
118 8 * 100 = 800 MHz max efficiency
119 35 * 100 = 3500 MHz TSC frequency
120 cpu0: MSR_IA32_POWER_CTL: 0x0004005d (C1E auto-promotion: DISabled)
121 cpu0: MSR_NHM_SNB_PKG_CST_CFG_CTL: 0x1e000400 (UNdemote-C3, UNdemote-C1, demote-C3, demote-C1, UNlocked: pkg-cstate-limit=0: pc0)
122 cpu0: MSR_NHM_TURBO_RATIO_LIMIT: 0x25262727
123 37 * 100 = 3700 MHz max turbo 4 active cores
124 38 * 100 = 3800 MHz max turbo 3 active cores
125 39 * 100 = 3900 MHz max turbo 2 active cores
126 39 * 100 = 3900 MHz max turbo 1 active cores
127 cpu0: MSR_IA32_ENERGY_PERF_BIAS: 0x00000006 (balanced)
128 cpu0: MSR_CORE_PERF_LIMIT_REASONS, 0x31200000 (Active: ) (Logged: Auto-HWP, Amps, MultiCoreTurbo, Transitions, )
129 cpu0: MSR_GFX_PERF_LIMIT_REASONS, 0x00000000 (Active: ) (Logged: )
130 cpu0: MSR_RING_PERF_LIMIT_REASONS, 0x0d000000 (Active: ) (Logged: Amps, PkgPwrL1, PkgPwrL2, )
131 cpu0: MSR_RAPL_POWER_UNIT: 0x000a0e03 (0.125000 Watts, 0.000061 Joules, 0.000977 sec.)
132 cpu0: MSR_PKG_POWER_INFO: 0x000002a0 (84 W TDP, RAPL 0 - 0 W, 0.000000 sec.)
133 cpu0: MSR_PKG_POWER_LIMIT: 0x428348001a82a0 (UNlocked)
134 cpu0: PKG Limit #1: ENabled (84.000000 Watts, 8.000000 sec, clamp DISabled)
135 cpu0: PKG Limit #2: ENabled (105.000000 Watts, 0.002441* sec, clamp DISabled)
136 cpu0: MSR_PP0_POLICY: 0
137 cpu0: MSR_PP0_POWER_LIMIT: 0x00000000 (UNlocked)
138 cpu0: Cores Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
139 cpu0: MSR_PP1_POLICY: 0
140 cpu0: MSR_PP1_POWER_LIMIT: 0x00000000 (UNlocked)
141 cpu0: GFX Limit: DISabled (0.000000 Watts, 0.000977 sec, clamp DISabled)
142 cpu0: MSR_IA32_TEMPERATURE_TARGET: 0x00641400 (100 C)
143 cpu0: MSR_IA32_PACKAGE_THERM_STATUS: 0x88340800 (48 C)
144 cpu0: MSR_IA32_THERM_STATUS: 0x88340000 (48 C +/- 1)
145 cpu1: MSR_IA32_THERM_STATUS: 0x88440000 (32 C +/- 1)
146 cpu2: MSR_IA32_THERM_STATUS: 0x88450000 (31 C +/- 1)
147 cpu3: MSR_IA32_THERM_STATUS: 0x88490000 (27 C +/- 1)
148 Core CPU Avg_MHz %Busy Bzy_MHz TSC_MHz SMI CPU%c1 CPU%c3 CPU%c6 CPU%c7 CoreTmp PkgTmp PkgWatt CorWatt GFXWatt
149 - - 493 12.64 3898 3498 0 12.64 0.00 0.00 74.72 47 47 21.62 13.74 0.00
150 0 0 4 0.11 3894 3498 0 99.89 0.00 0.00 0.00 47 47 21.62 13.74 0.00
151 0 4 3897 99.98 3898 3498 0 0.02
152 1 1 7 0.17 3887 3498 0 0.04 0.00 0.00 99.79 32
153 1 5 0 0.00 3885 3498 0 0.21
154 2 2 29 0.76 3895 3498 0 0.10 0.01 0.01 99.13 32
155 2 6 2 0.06 3896 3498 0 0.80
156 3 3 1 0.02 3832 3498 0 0.03 0.00 0.00 99.95 28
157 3 7 0 0.00 3879 3498 0 0.04
161 The \fBmax efficiency\fP frequency, a.k.a. Low Frequency Mode, is the frequency
162 available at the minimum package voltage. The \fBTSC frequency\fP is the base
163 frequency of the processor -- this should match the brand string
164 in /proc/cpuinfo. This base frequency
165 should be sustainable on all CPUs indefinitely, given nominal power and cooling.
166 The remaining rows show what maximum turbo frequency is possible
167 depending on the number of idle cores. Note that not all information is
168 available on all processors.
170 The --debug option adds additional columns to the measurement ouput, including CPU idle power-state residency processor temperature sensor readinds.
171 See the field definitions above.
173 If turbostat is invoked with a command, it will fork that command
174 and output the statistics gathered when the command exits.
175 eg. Here a cycle soaker is run on 1 CPU (see %c0) for a few seconds
176 until ^C while the other CPUs are mostly idle:
179 root@hsw: turbostat cat /dev/zero > /dev/null
181 CPU Avg_MHz %Busy Bzy_MHz TSC_MHz
182 - 482 12.51 3854 3498
186 5 3854 99.98 3855 3498
194 Above the cycle soaker drives cpu5 up its 3.9 GHz turbo limit.
195 The first row shows the average MHz and %Busy across all the processors in the system.
197 Note that the Avg_MHz column reflects the total number of cycles executed
198 divided by the measurement interval. If the %Busy column is 100%,
199 then the processor was running at that speed the entire interval.
200 The Avg_MHz multiplied by the %Busy results in the Bzy_MHz --
201 which is the average frequency while the processor was executing --
202 not including any non-busy idle time.
208 Alternatively, non-root users can be enabled to run turbostat this way:
210 # setcap cap_sys_rawio=ep ./turbostat
212 # chmod +r /dev/cpu/*/msr
215 reads hardware counters, but doesn't write them.
216 So it will not interfere with the OS or other programs, including
217 multiple invocations of itself.
220 may work poorly on Linux-2.6.20 through 2.6.29,
221 as \fBacpi-cpufreq \fPperiodically cleared the APERF and MPERF MSRs
224 AVG_MHz = APERF_delta/measurement_interval. This is the actual
225 number of elapsed cycles divided by the entire sample interval --
226 including idle time. Note that this calculation is resilient
227 to systems lacking a non-stop TSC.
229 TSC_MHz = TSC_delta/measurement_interval.
230 On a system with an invariant TSC, this value will be constant
231 and will closely match the base frequency value shown
232 in the brand string in /proc/cpuinfo. On a system where
233 the TSC stops in idle, TSC_MHz will drop
234 below the processor's base frequency.
236 %Busy = MPERF_delta/TSC_delta
238 Bzy_MHz = TSC_delta/APERF_delta/MPERF_delta/measurement_interval
240 Note that these calculations depend on TSC_delta, so they
241 are not reliable during intervals when TSC_MHz is not running at the base frequency.
243 Turbostat data collection is not atomic.
244 Extremely short measurement intervals (much less than 1 second),
245 or system activity that prevents turbostat from being able
246 to run on all CPUS to quickly collect data, will result in
247 inconsistent results.
249 The APERF, MPERF MSRs are defined to count non-halted cycles.
250 Although it is not guaranteed by the architecture, turbostat assumes
251 that they count at TSC rate, which is true on all processors tested to date.
254 "Intel® Turbo Boost Technology
255 in Intel® Core™ Microarchitecture (Nehalem) Based Processors"
256 http://download.intel.com/design/processor/applnots/320354.pdf
258 "Intel® 64 and IA-32 Architectures Software Developer's Manual
259 Volume 3B: System Programming Guide"
260 http://www.intel.com/products/processor/manuals/
273 Written by Len Brown <len.brown@intel.com>