[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Fedora Core 5 Test Update: kernel-2.6.17-1.2187_FC5



On Mon, 2006-09-11 at 18:57 -0600, Kevin Fenzi wrote:
> >>>>> "David" == David Nielsen <david lovesunix net> writes:
> 
> David> man, 11 09 2006 kl. 19:27 -0500, skrev Jay Cliburn:
> >> David Nielsen wrote: > man, 11 09 2006 kl. 19:45 -0400, skrev Dave
> >> Jones:
> >> >> On Mon, Sep 11, 2006 at 06:15:28PM -0500, Gilbert Sebenste wrote:
> >> >>  > Dave Jones wrote:
> >> >>  >
> >> >> > >>  
> >> ---------------------------------------------------------------------
> >> >> > >> Fedora Test Update Notification >> > >> FEDORA-2006-967 >>
> >> > >> 2006-09-11 >> > >>
> >> ---------------------------------------------------------------------
> >> >>  > >>
> >> >> > >> Product : Fedora Core 5 >> > >> Name : kernel >> > >>
> >> Version : 2.6.17 >> > >> Release : 1.2187_FC5
> >> >>  >
> >> >> > Worksforme. Nicely done. I take it 2.6.18 is not going to be
> >> released for >> > a while?
> >> >>
> >> >> I'm hoping to get a 2.6.18rc6 based kernel into updates-testing
> >> soon >> (and then rebase to .18 for updates-final when that gets
> >> released), but >> as usual, xen is throwing a spanner in the works.
> >> >
> >> > if you do that I believe we would be likely to hit this one:
> >> >
> >> > https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=205305
> >> >
> >> > which is currently breaking cpu scaling for me on Development
> >>
> >> Hmmm...  I have an AMD64x2 3800+ that scales up to 2 GHz just fine
> >> under the current rawhide kernel.
> 
> David> Very odd, I and a few others are seeing this, maybe something
> David> is loading wrong modules on certain machines. I had to manually
> David> recompile the kernel to make it stop locking it at the lowest
> David> setting (although I managed to kill scaling as well but I'll
> David> take broken scaling at 2.2GHz over broken scaling at 1GHz)
> 
> Amusingly enough, I just got one of these today.
> (AMD Athlon(tm) 64 X2 Dual Core Processor 4200+)
> 
> Mine also seems to stick at 1ghz on both cores instead of ramping up
> to 2.2ghz (as the cpus are able to).
> 
> It looks to me like a bug in the 'ondemand' cpu govener...
> If I simply change to performance:
> 
> echo "performance" > /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
> 
> It goes to 2.2 and stays there. ondemand seems to never raise from the
> lowest possible speed here.
> 
> David> - David
> 
> kevin

I've been noticing under rawhide with my single core Intel Pentium M
that the cpufreq applet reports On Demand and that it only seems to swap
between 600 MHz and 2000 MHz with no steps in between, 30% to 100% to
30% to 100% back to 30%. It used to ramp up and down based on demand but
now just jumps to either extreme. Maybe it's more efficient this way?

tjb
--
=======================================================================
| Thomas Baker                                  email: tjb unh edu    |
| Systems Programmer                                                  |
| Research Computing Center                     voice: (603) 862-4490 |
| University of New Hampshire                     fax: (603) 862-1761 |
| 332 Morse Hall                                                      |
| Durham, NH 03824 USA              http://wintermute.sr.unh.edu/~tjb |
=======================================================================


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]