Hello,
Per: http://kolbusa.livejournal.com/71066.html
Is it necessary to do this--?
Is there a way to verify whether turbo boost is being enabled when one
core is at 100% cpu?
Justin.
I realize the BIOS is supposed to take care of it, but it would be nice if
there was an easy way to verify it is working.
On Tue, 3 Nov 2009, Justin Piszcz wrote:
> Hello,
>
> Per: http://kolbusa.livejournal.com/71066.html
>
> Is it necessary to do this--?
>
> Is there a way to verify whether turbo boost is being enabled when one core
> is at 100% cpu?
>
> Justin.
>
On Tue, 2009-11-03 at 16:00 -0500, Justin Piszcz wrote:
> I realize the BIOS is supposed to take care of it, but it would be nice if
> there was an easy way to verify it is working.
>
> On Tue, 3 Nov 2009, Justin Piszcz wrote:
>
> > Hello,
> >
> > Per: http://kolbusa.livejournal.com/71066.html
> >
> > Is it necessary to do this--?
> >
> > Is there a way to verify whether turbo boost is being enabled when one core
> > is at 100% cpu?
> >
> > Justin.
Assuming that the cpufreq functionality is enabled in your kernel, a
simple
cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_cur_freq
Should return the frequencies that each cpu core is currently running
at. I'm using a core i7 920 (2.67 ghz nominal) and have seen it hit 2.80
ghz on occasion.
None of the extra kernel options mentioned in that blog post should be
required, assuming that you are using an up-to-date kernel. I would
expect this to work correctly in the latest Ubuntu or Fedora, for
example.
For the record, I have all of the bios power saving modes enabled (and
turbo mode, of course), and am using the kernel 'ondemand' governor with
no extra userspace applications or configuration.
My understanding of how it works is that when the cpu load is OK, the
bios adjusts the values for the top couple of speedstep settings (use
cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_available_frequencies
to see them), and the kernel will automatically take advantage of them,
according to the cpufreq governor you select.
--
Calvin Walton <[email protected]>