#!/bin/bash
#
# /etc/rc.local: Local multi-user startup script.
#
cpufreq-set -c 0 -g conservative
cpufreq-set -c 1 -g conservative
It may be that the cpufreq monitor in gkrellm may be reporting the wrong frequency? when I do a
cat /sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_cur_freq
(on CPU0 or CPU1) even though gkrellm is telling me that cpufreq is set at 1830MHz, this readout is telling me 1000MHz. When I open up a bunch of programs and redo that cat, it says 1830.
So perhaps it is working correctly afterall, it's just the sensor in gkrellm or conky is not reporting correctly. thepizzaking, I appreciate the input and I'll look into this further.
Yes, I've noticed that at times too, though conky and the gnome applet tell me my cpu is at 2.13GHz, it's really running at 800MHz.
]]>cat /sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_cur_freq
(on CPU0 or CPU1) even though gkrellm is telling me that cpufreq is set at 1830MHz, this readout is telling me 1000MHz. When I open up a bunch of programs and redo that cat, it says 1830.
So perhaps it is working correctly afterall, it's just the sensor in gkrellm or conky is not reporting correctly. thepizzaking, I appreciate the input and I'll look into this further.
]]>Powersaved, powernowd, and cpufreqd all do throttle my cpu, but it seems to be very sensitive. While the system is completely idle, my 1.8GHz C2D (using speedstep_centrino) will be at 900MHz. 50%, perfect, I think to myself. Yet when there is *any* load whatsoever, even a constant 3% load, my CPU is at 1.8GHz. The powersaved cpufreq utility documentation is talking about 50% jumps in CPU load until it increases cpu frequency... but that is not the case for me.
Is there some sort of tolerance setting somewhere that I'm missing?
]]>