--On Sunday, 11 May 2003 4:48 p.m. -0400 Valdis.Kletnieks@vt.edu wrote:
> On Sun, 11 May 2003 23:28:25 +1200, Andrew McGregor
> <andrew@indranet.co.nz> said:
>
>> cpufreq and speedstep don't work on Dell P3 laptops anyway, and the
>> *internal power supplies* of the i8x00 series make wierd noises when APM
>> tries to idle the CPU. The board will do this anyway, without making
>> noise, so linux need not.
>
> Dell Latitude C840 (1.6G Pentium4 Mobile) has the "power supplies buzz at
> 1Khz on APM idle" symptom too. I haven't checked the ACPI side of the
> fence yet, nor have I gotten brave enough to try the cpufreq and
> speedstep stuff.
AFAIK Speedstep should always work on P4M systems, so it's worth a try.
>
> Even *more* bizarre, there's "something odd" done by the seti@home client
> (which usually causes 100% CPU use and thus silence) several minutes into
> a workunit that causes the noise to change frequencies - it will start
> down around 500hz, sweep up to 1Khz (taking about 2 seconds to do so),
> and repeat (so the buzzing is exhibiting a sawtooth wave). I'm not
> seeing any paging or swapping or I/O, so I'm wondering if it's some code
> walking through a large array with strides 1/2/4/8/16 (like an FFT)
> causing different cache hit ratios, and thus different power consumption
> patterns while it's stuck on a L1/L2 cache miss.....
>
That's pretty bizzarre, I agree :-)
Probably it's an autocorrelation with a sliding offset. I'd guess that the
reason the buzz starts at 500Hz and sweeps up is that if the frequency is
below 500Hz you don't hear it. Maybe the RAM starts sucking more power as
the cache hit ratio drops...
Andrew
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu May 15 2003 - 22:00:38 EST