Timeline for Get all available frequency steps
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 19, 2018 at 18:34 | history | edited | WinEunuuchs2Unix | CC BY-SA 4.0 |
Correct loop control.
|
| Aug 19, 2018 at 17:21 | comment | added | WinEunuuchs2Unix | @ilkkachu I've added a script to list available frequencies based on linear methodology. Perhaps it might shed some clues. | |
| Aug 19, 2018 at 17:19 | history | edited | WinEunuuchs2Unix | CC BY-SA 4.0 |
Add script
|
| Aug 19, 2018 at 16:32 | comment | added | WinEunuuchs2Unix |
@ilkkachu not sure. Using the cpuinfo alias above on the cpufreq directory last two readings are 1699.941 and 1007.489 Mhz. As OP noted though when you set a frequency Intel automatically finds the closest "real" one.
|
|
| Aug 19, 2018 at 16:30 | comment | added | ilkkachu |
mmh. That wouldn't match the steps of 96.4 either. I wonder if it averages the frequency over some (even short) period of time. I seem to remember seeing /proc/cpuinfo do that.
|
|
| Aug 19, 2018 at 16:28 | comment | added | WinEunuuchs2Unix | I agree about 100 being a nice round number. However my conky shows MHz at 1148, 1147, 1919, 1706, 1111, 2127, 1795 as I sit here watching display and typing this comment. So round numbers are not the way it works. | |
| Aug 19, 2018 at 16:26 | comment | added | ilkkachu |
Hmm, (3500-800)/27 would be 100 (MHz) exactly, which is a nice round number, so it seems likely that 28 would be the full number of steps (counting both ends). That is, you'd have 28 steps with frequency f(n) = 800 + n*100, where n is in 0..27.
|
|
| Aug 19, 2018 at 16:08 | history | answered | WinEunuuchs2Unix | CC BY-SA 4.0 |