Skip to contentSkip to conversation

pwoertop frequencies not ordered descending

DISCUSSION Toralf Förster 2 participants 2 messages Jan 26, 2015

lore
Share & tools
lore
#0 lore raw

With the ACPI driver (and CPU governor "ondemand") the output of powertop looked always fine. However with the P-State driver I do now have this output at 2 hardened 64 bit Gentoo Linux systems:

PowerTOP 2.6 Overview Idle stats Frequency stats Device stats Tunables

            Package |             Core    |            CPU 0       CPU 1
                    |                     | Actual    2.6 GHz     2.7 GHz
Idle         0.3%   | Idle        63.2%   | Idle        73.9%       75.0%
2.50 GHz     0.1%   | 2.50 GHz     0.1%   | 2.50 GHz     0.1%        0.0%
2.60 GHz     9.0%   | 2.60 GHz     8.9%   | 2.60 GHz     8.9%        0.3%
2.90 GHz    74.5%   | 2.90 GHz    13.4%   | 2.90 GHz     5.4%       11.3%
2.71 GHz    13.5%   | 2.71 GHz    12.9%   | 2.71 GHz    11.2%       12.3%
 800 MHz     0.2%   |  800 MHz     0.1%   | 2.81 GHz     0.5%        0.1%
2.81 GHz     2.3%   | 2.81 GHz     1.1%   | 1200 MHz     0.0%        0.9%
 900 MHz     0.0%   | 1000 MHz     0.1%   | 1100 MHz     0.0%        0.1%
1000 MHz     0.1%   | 1200 MHz     0.0%   | 1000 MHz     0.0%        0.0%
1200 MHz     0.0%   |  900 MHz     0.0%   |  800 MHz     0.0%        0.0%
1100 MHz     0.0%   | 1100 MHz     0.0%   |  900 MHz     0.0%        0.0%
1300 MHz     0.0%   | 1500 MHz     0.0%   |                          0.0%
1500 MHz     0.0%   | 1400 MHz     0.0%   |                          0.0%
1400 MHz     0.0%   |                     |
                    |             Core    |            CPU 2       CPU 3
                    |                     | Actual    2.8 GHz     2.8 GHz
                    | Idle         6.6%   | Idle        23.1%       24.9%
                    | 2.50 GHz     0.1%   | 2.50 GHz     0.0%        0.1%
                    | 2.60 GHz     8.9%   | 2.60 GHz     0.1%        8.9%
                    | 2.90 GHz    69.1%   | 2.90 GHz    62.9%       62.5%

Is this (partly ?) related to the kernel or just an issue of powertop ?

Signature · 2 lines
Toralf
pgp key: 7B1A 07F4 EC82 0F90 D4C2  8936 872A E508 0076 E94E
#1 lore raw
Cc Kristen
Quoted 34 lines · Toralf Förster

With the ACPI driver (and CPU governor "ondemand") the output of powertop looked always fine. However with the P-State driver I do now have this output at 2 hardened 64 bit Gentoo Linux systems:

PowerTOP 2.6 Overview Idle stats Frequency stats Device stats Tunables

            Package |             Core    |            CPU 0       CPU 1
                    |                     | Actual    2.6 GHz     2.7 GHz
Idle         0.3%   | Idle        63.2%   | Idle        73.9%       75.0%
2.50 GHz     0.1%   | 2.50 GHz     0.1%   | 2.50 GHz     0.1%        0.0%
2.60 GHz     9.0%   | 2.60 GHz     8.9%   | 2.60 GHz     8.9%        0.3%
2.90 GHz    74.5%   | 2.90 GHz    13.4%   | 2.90 GHz     5.4%       11.3%
2.71 GHz    13.5%   | 2.71 GHz    12.9%   | 2.71 GHz    11.2%       12.3%
 800 MHz     0.2%   |  800 MHz     0.1%   | 2.81 GHz     0.5%        0.1%
2.81 GHz     2.3%   | 2.81 GHz     1.1%   | 1200 MHz     0.0%        0.9%
 900 MHz     0.0%   | 1000 MHz     0.1%   | 1100 MHz     0.0%        0.1%
1000 MHz     0.1%   | 1200 MHz     0.0%   | 1000 MHz     0.0%        0.0%
1200 MHz     0.0%   |  900 MHz     0.0%   |  800 MHz     0.0%        0.0%
1100 MHz     0.0%   | 1100 MHz     0.0%   |  900 MHz     0.0%        0.0%
1300 MHz     0.0%   | 1500 MHz     0.0%   |                          0.0%
1500 MHz     0.0%   | 1400 MHz     0.0%   |                          0.0%
1400 MHz     0.0%   |                     |
                    |             Core    |            CPU 2       CPU 3
                    |                     | Actual    2.8 GHz     2.8 GHz
                    | Idle         6.6%   | Idle        23.1%       24.9%
                    | 2.50 GHz     0.1%   | 2.50 GHz     0.0%        0.1%
                    | 2.60 GHz     8.9%   | 2.60 GHz     0.1%        8.9%
                    | 2.90 GHz    69.1%   | 2.90 GHz    62.9%       62.5%

Is this (partly ?) related to the kernel or just an issue of powertop ?

Signature · 2 lines
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.