I have two problems with my CPU Intel Core i5-3210M on my laptop Lenovo B590. First, I cannot see all tipical governors availables such as ondemand, userspace, conservative... only I see powersave and performance with intel_pstate driver enabled. I see this result of cpupower frequency-info with pstate driver: "analyzing CPU 0: driver: intel_pstate CPUs which run at the same hardware frequency: 0 CPUs which need to have their frequency coordinated by software: 0 maximum transition latency: 0.97 ms. hardware limits: 1.20 GHz - 3.10 GHz available cpufreq governors: performance, powersave current policy: frequency should be within 1.20 GHz and 3.10 GHz. The governor "powersave" may decide which speed to use within this range. current CPU frequency is 2.77 GHz (asserted by call to hardware). boost state support: Supported: yes Active: yes 2900 MHz max turbo 4 active cores 2900 MHz max turbo 3 active cores 2900 MHz max turbo 2 active cores 3100 MHz max turbo 1 active cores" I can see on each cpu* file cpuinfo_max_freq with 3100000 values. I guess is available set frequency into 3.1Ghz. ... "# cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_max_freq 3100000 " Other way, turbostat display only 2.89 frecuencies of all cpus compiling a linux kernel for test porporsing with make -j5. This is display: "cor CPU %c0 GHz TSC SMI %c1 %c3 %c6 %c7 CTMP PTMP %pc2 %pc3 %pc6 %pc7 Pkg_W Cor_W GFX_W 99.78 2.89 2.49 0 0.22 0.00 0.00 0.00 61 61 0.00 0.00 0.00 0.00 18.24 13.54 1.64 0 0 99.80 2.89 2.49 0 0.20 0.00 0.00 0.00 60 61 0.00 0.00 0.00 0.00 18.24 13.54 1.64 0 1 99.78 2.89 2.49 0 0.22 1 2 99.92 2.89 2.49 0 0.08 0.00 0.00 0.00 61 1 3 99.62 2.89 2.49 0 0.38 " You can see how does not increase into 3.1Ghz per core. It's does not works properly Turbo Boost. Anyway intel_pstate is loaded correctly: " [ 0.852750] Intel pstate controlling: cpu 0 [ 0.852762] Intel pstate controlling: cpu 1 [ 0.852777] Intel pstate controlling: cpu 2 [ 0.852789] Intel pstate controlling: cpu 3 " Respect governors with intel_pstate=disabled at kernel command line boot. I can see all kernel's governors if I boot with intel_pstate=disable. That's using old ACPI driver called acpi_cpufreq. This is result of cpupower frequency-info: "analyzing CPU 0: driver: acpi-cpufreq CPUs which run at the same hardware frequency: 0 CPUs which need to have their frequency coordinated by software: 0 maximum transition latency: 10.0 us. hardware limits: 1.20 GHz - 2.50 GHz available frequency steps: 2.50 GHz, 2.50 GHz, 2.40 GHz, 2.30 GHz, 2.20 GHz, 2.10 GHz, 2.00 GHz, 1.90 GHz, 1.80 GHz, 1.70 GHz, 1.60 GHz, 1.50 GHz, 1.40 GHz, 1.30 GHz, 1.20 GHz available cpufreq governors: conservative, userspace, powersave, ondemand, performance current policy: frequency should be within 1.20 GHz and 2.50 GHz. The governor "ondemand" may decide which speed to use within this range. current CPU frequency is 1.20 GHz (asserted by call to hardware). boost state support: Supported: yes Active: yes 2900 MHz max turbo 4 active cores 2900 MHz max turbo 3 active cores 2900 MHz max turbo 2 active cores 3100 MHz max turbo 1 active cores " However, I saw on all cpuinfo_max_freq reported: "# cat /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_max_freq 2501000" It's not same that pstate that shows 3100000. However, includes switch governor into performance. Does not increases into 3.1Ghz when I was compiling a Linux kernel to test. This is same results that use intel_pstate: "cor CPU %c0 GHz TSC SMI %c1 %c3 %c6 %c7 CTMP PTMP %pc2 %pc3 %pc6 %pc7 Pkg_W Cor_W GFX_W 99.57 2.89 2.49 0 0.43 0.00 0.00 0.00 59 59 0.00 0.00 0.00 0.00 18.37 13.38 1.94 0 0 99.80 2.89 2.49 0 0.20 0.00 0.00 0.00 59 59 0.00 0.00 0.00 0.00 18.37 13.38 1.94 0 1 99.25 2.89 2.49 0 0.75 1 2 99.61 2.89 2.49 0 0.39 0.00 0.00 0.00 58 1 3 99.61 2.89 2.49 0 0.39 " It's likes bug on intel_pstate driver. At Lenovo BIOS Intel Speed Step and Turbo Boost It's enabled by default.
2900 MHz max turbo 4 active cores intel_pstate has an internal governor. If you are using intel_pstate you use its internal governor which has two modes performance which is equivalent to the performance governor and powersave roughly equal to ondemand but is better on power. When a scaling driver with an internal governor is being used powersave and performance are the only options You have a 2.5Ghz processor and are getting 2.9Ghz in both cases which *is* in the turbo range.
Yes I am getting 2.90 but does not set 3.1Ghz. For example, In this bug https://bugzilla.kernel.org/show_bug.cgi?id=66581 got problems that does not increases more of 2.8Ghz of kernel when max frecuencies is 3.8Ghz. I don't understand why is not increases into 3.1Ghz.If not Intel should be tell the truth about Turbo boost and frequencies. By other hand, I end compile a kernel over 30 minutes more less. However, my friend does have an Intel Core 2 P8600 takes 10 minutes more less. How is a CPU old with 2 cores can compile a kernel into 10 minutes more less than my new CPU It's new model. I think is a bug. Some users of Gentoo with i5 processors tells me that 5-6 minutes more less should be end compiling a kernel process. :/
The msr-tools package from https://01.org/msr-tools/downloads makes it easy to see what P state is being requested for each core with the command: sudo rdmsr -a 0x199 The requested P state is in the upper eight bits. The P state intel_pstate requests is just that a request in the turbo range, the CPU itself decides what frequency will be used while keeping the processor in it thermal envelop.
This command reports: rdmsr -a 0x199 rdmsr: invalid option -- 'a' Usage: rdmsr [options] regno --help -h Print this help --version -V Print current version --hexadecimal -x Hexadecimal output (lower case) --capital-hex -X Hexadecimal output (upper case) --decimal -d Signed decimal output --unsigned -u Unsigned decimal output --octal -o Octal output --c-language -c Format output as a C language constant --zero-pad -0 Output leading zeroes --raw -r Raw binary output --processor # -p Select processor number (default 0) --bitfield h:l -f Output bits [h:l] only msr-tools-1.1.2-8.fc20.x86_64 I don't understand this. How can an Intel Core 2 duo more older than my Intel i5 processor It's passed 3 generations of releases built. Be more powerful than it? Some conclusions: - 1º: Intel is lying. I should as good citizen I am report them for fraud. - 2º: Intel_pstate does not working correctly.
(In reply to Álvaro Castillo from comment #4) > This command reports: > > > msr-tools-1.1.2-8.fc20.x86_64 > The current version of msr-tools is v1.3 which the link above points to. > I don't understand this. > > How can an Intel Core 2 duo more older than my Intel i5 processor It's > passed 3 generations of releases built. Be more powerful than it? > Are both systems configured the same? memory, disk, laptop/desktop/server? Are you building with the same .config on both systems? > Some conclusions: > - 1º: Intel is lying. I should as good citizen I am report them for fraud. > - 2º: Intel_pstate does not working correctly. So you are getting the highest turbo frequency available when four cores are active. I fail to see how this is an intel_pstate issue. BTW you got the same results with ondemand.