Re: p4-clockmod doing the Right Thing[tm]?

[Ian Morgan]

On Tue, 19 Nov 2002, Ian Morgan wrote:

> ... the p4-clockmod module loads OK. Thx.

OK.. this is strange, maybe.

# cat /proc/cpufreq
          minimum CPU frequency  -  maximum CPU frequency  -  policy
CPU  0       300685 kHz ( 12 %)  -    2405487 kHz (100 %)  -  performance

# cat /proc/sys/cpu/0/*
2405487
2405487
300685

# echo 300685 >proc/sys/cpu/0/speed
[box stalls for about 25-30 seconds]

# cat /proc/sys/cpu/0/*
300685
2405487
300685

# echo 2405487 >proc/sys/cpu/0/speed
[box stalls for about 25-30 seconds]

# cat /proc/sys/cpu/0/*
2706165
2405487
300685

# cat /proc/cpufreq
          minimum CPU frequency  -  maximum CPU frequency  -  policy
CPU  0     2405487 kHz ( 100 %)  -    2706165 kHz (112 %)  -  powersave

?!?! Should the P4 really be able to go to 112%, and why would it do that
after I told it to go back to 100% ?

Any idea why the box would stall (at least no kb/mouse input is accepted,
not sure about other interrupts) for such a long time when changing the
speed?

On the other hand, some short (<1 min) cpu-bound processes for benchmarking
each took the same amount of time to run when the clock was at 12%, 100%,
and 112%, so I'm not sure the clock was really being changed at all.

Regards,
Ian Morgan

-- 
-------------------------------------------------------------------
 Ian E. Morgan          Vice President & C.O.O.       Webcon, Inc.
 [e-mail]          PGP: #2DA40D07           www.webcon.ca
    *  Customized Linux network solutions for your business  *
-------------------------------------------------------------------