Re: cpufreq probs dual core intel



Ok, a little more info...

plugged in

leroy# sysctl hw.acpi.acline
hw.acpi.acline: 1

unplugged
leroy# sysctl hw.acpi.acline
hw.acpi.acline: 0

and from /var/log/messages
Jun 27 07:18:46 leroy power_profile: changed to 'economy'
Jun 27 07:18:57 leroy power_profile: changed to 'performance'

When unplugged the screen dims, the machine seems to slow down, yet heres
the problem...
plugged in

leroy# sysctl dev.cpu.0.freq
dev.cpu.0.freq: 2000

unplugged

leroy# sysctl dev.cpu.0.freq
dev.cpu.0.freq: 2000

I can change the freq using sysctl but it doesn't stick, no crash or
anything though.

Whats odd to me is the temp the cpu is running at.
unplugged
dev.cpu.0.temperature: 54
dev.cpu.1.temperature: 55
plugged in
dev.cpu.0.temperature: 61
dev.cpu.1.temperature: 62

Everything seems to be working as it should, just that pesky cpu freq

Oh by the way, the laptop is a Dell Latitude D820, and I did find out after
testing acpi sleep while X/gnome is running it works! And another side
note... with FreeBSD 7.2 I get approx 2 hours battery life using wifi,
windows gives 1.5 hrs, OpenSolaris 1 hour, have not left the laptop
unplugged with current long enough to test.





On Fri, Jun 26, 2009 at 9:09 PM, Kevin Oberman <oberman@xxxxxx> wrote:

Date: Fri, 26 Jun 2009 19:24:52 -0700
From: Ron Freidel <rfreidel@xxxxxxxxx>
Sender: owner-freebsd-current@xxxxxxxxxxx

Hi all,

I suppose I should state I am only posting after lots of googling.

I upgraded to current from 7.2, I just couldn't wait too see if my laptop
would acpi sleep, it seems to just fine by the way, not well tested yet
though.

My prob is with cpufreq, it seems to be locked at max speed, nothing I do
seems to change the freq it is running at.

CPU: Intel(R) Core(TM)2 CPU T7200 @ 2.00GHz (1997.34-MHz
K8-class
CPU)

leroy# dmesg | grep cpu
cpu0 (BSP): APIC ID: 0
cpu1 (AP): APIC ID: 1
cpu0: <ACPI CPU> on acpi0
coretemp0: <CPU On-Die Thermal Sensors> on cpu0
est0: <Enhanced SpeedStep Frequency Control> on cpu0
p4tcc0: <CPU Frequency Thermal Control> on cpu0
cpu1: <ACPI CPU> on acpi0
coretemp1: <CPU On-Die Thermal Sensors> on cpu1
est1: <Enhanced SpeedStep Frequency Control> on cpu1
p4tcc1: <CPU Frequency Thermal Control> on cpu1

It looks to this feeble mind like it should be working.
Where else do I look?

leroy# ps ax | grep powerd
1164 ?? Ss 0:00.97 /usr/sbin/powerd -a adaptive -b adaptive

Hope this is enough info...

Not quite. The output of 'sysctl hw.acpi.acline' (on battery and on AC
power) and 'sysctl dev.apu.0' would help. Running powerd with the '-v'
option might provide a clue, too.

Everything you sent looks pretty normal, but it's not enough to tell much.
--
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman@xxxxxx Phone: +1 510 486-8634
Key fingerprint:059B 2DDF 031C 9BA3 14A4 EADA 927D EBB3 987B 3751




--
Ron
_______________________________________________
freebsd-current@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@xxxxxxxxxxx"



Relevant Pages

  • Re: cpufreq probs dual core intel
    ... leroy# sysctl hw.acpi.acline ... A test for the real clock ... The actual clock frequency does not ...
    (freebsd-current)
  • Re: cpufreq probs dual core intel
    ... leroy# sysctl hw.acpi.acline ... One final thought on this...run powerd from the command line with the ... Energy Sciences Network (ESnet) ...
    (freebsd-current)
  • Re: CONFIG_PREEMPT and server workloads
    ... I've been meaning to do another round of latency tuneups for ages, ... but the other CPU is hanging on the lock for ages ... ALSA drivers which would cause a dump_stackto be triggered if the audio ... of underruns we can just ask them to turn on the sysctl and we get a trace ...
    (Linux-Kernel)
  • Re: [PATCH] avoid cpu removal if busy revisited
    ... added sysctl ... cpu hot remove migrates all tasks on target cpu by force. ... +#ifdef CONFIG_HOTPLUG_CPU ...
    (Linux-Kernel)
  • Re: New optimized soreceive_stream() for TCP sockets, proof of concept
    ... >> those OSes do much better when CPU affinity ... You can slap a sched_bindinto the interrupt> handler the first time it runs and it should stick appropriately, and add> a sysctl to sched_bindfor a user process as a hack to test it out. ... Right now we have a facility to bind kernel threads to specific CPUs ), but not an API or mature facility to bind user threads. ...
    (freebsd-net)