Author |
Message |
knappster
|
Posted: Sat May 24, 2008 11:11 pm |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
I honestly don't know what the requirements are for powersaved to auto-scale your cpu. I think it is the AMD Cool 'n Quiet technology (or the Intel equivalent). I currently have it on my PC and my combined R5F27 BE/FE. My PC has this mobo/cpu combo:
ABIT KN9 AM2 NVIDIA nForce4 Ultra
AMD Athlon 64 X2 4600+ Windsor 2.4GHz Socket AM2 89W Dual-Core Processor Model ADA4600CUBOX
My combined BE/FE has this mobo/cpu combo:
ASUS M2NPV-VM
AMD Sempron 64 3400+ Manila 1.8GHz Socket AM2 62W Single-Core Processor Model SDA3400CWBOX
powersaved is available via apt-get without trying to upgrade any of my other packages, so I just went ahead and installed it by:
Code: apt-get install powersaved as root. When it tries to run after installing it complains because acpid is not running. This can be started with: Code: /etc/init.d/acpid start as root you can start powersaved with the command: Code: /etc/init.d/powersaved start as root It will still complain when you try to run any of the powersave commands. After some trial and error, I found that starting the hal daemon remedies this. You can start itwith the command: Code: /etc/init.d/hal start as root You can double check that it's using dynamic scaling by typing: Code: powersave -c as root. It should come back Code: DYNAMIC To make sure that it is working properly you should be able to type: Code: cat /proc/cpuinfo and see approx 1000 MHz when idle (both of my machines idle at 1000 MHz). then if you open another terminal and type glxgears and check Code: cat /proc/cpuinfo again it should be something higher than 1000 MHz. If you reboot, then powersaved will not run at startup because acpid and hald will not run on startup. After some trial and error (again) I was able to get it to autostart with the following sequence: Code: update-rc.d acpid defaults 80 update-rc.d hal defaults 80 update-rc.d -f powersaved remove update-rc.d powersaved defaults 99
The main difference I notice between my KM BE/FE and my desktop computer is that the desktop computer scales both frequency and core voltage. I see a pretty drastic change in temperature on both computers with no change in performance. Unfortunately I don't have a Kill-A-Watt, but I would expect there to be a pretty substantial difference in power consumption.
I also have acpi shutdown/mythwelcome setup on my combined BE/FE and I haven't yet tested that with powersaved, but I don't expect there to be any issues. I will report back here if there are.
|
|
Top |
|
|
cliffsjunk
|
Posted: Sun May 25, 2008 1:52 pm |
|
Joined: Fri Sep 15, 2006 12:16 pm
Posts: 292
|
I just tried this on my dual core intel 940 running R6F27 and it does not work.
'powersave -c' prints
Speedstepping is not supported.
Any ideas? I didn't find anything helpful in Google.
Cliff
_________________ R5F27 using R5F1 Nvidia drivers
HD-5500 analog from NTSC Sat Rx, with OTA DVB too
GeForce MX-440 SVideo tvout to a TV
Older dual core 3.4ghz Intel CPU
Asus P5PE-VM Motherboard
2 GB RAM
1 TB LVM2
VirtualBox
Samba
|
|
Top |
|
|
graysky
|
Posted: Mon May 26, 2008 4:22 am |
|
Joined: Wed Dec 10, 2003 8:31 pm
Posts: 1996
Location:
/dev/null
|
I thought speedstep was a BIOS-level thing that was independent of o/s? Is there a utility like CPU-Z for LINUX you can use to confirm that indeed your multiplier is dropping on idle?
_________________ Retired KM user (R4 - R6.04); friend to LH users.
|
|
Top |
|
|
knappster
|
Posted: Mon May 26, 2008 6:00 am |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
cliffsjunk wrote: I just tried this on my dual core intel 940 running R6F27 and it does not work.
'powersave -c' prints Speedstepping is not supported.
Any ideas? I didn't find anything helpful in Google.
Cliff Hey Cliff, I don't have an intel machine to test this on . My guess is that there's an option in bios that is not enabled. For AMD motherboards that support it, it's Cool 'n Quiet. For Intel I believe they call it "Enhanced Intel Speedstep Technology (EIST)". What motherboard are you using? graysky wrote: I thought speedstep was a BIOS-level thing that was independent of o/s? Is there a utility like CPU-Z for LINUX you can use to confirm that indeed your multiplier is dropping on idle?
Code: cat /proc/cpuinfo returns Code: processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 75 model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ stepping : 2 cpu MHz : 1000.000 cache size : 512 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8_legacy bogomips : 2001.51 TLB size : 1024 4K pages clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: ts fid vid ttp tm stc
processor : 1 vendor_id : AuthenticAMD cpu family : 15 model : 75 model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ stepping : 2 cpu MHz : 1000.000 cache size : 512 KB physical id : 0 siblings : 2 core id : 1 cpu cores : 2 fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8_legacy bogomips : 2001.51 TLB size : 1024 4K pages clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: ts fid vid ttp tm stc
when idle and Code: processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 75 model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ stepping : 2 cpu MHz : 2400.000 cache size : 512 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8_legacy bogomips : 4803.64 TLB size : 1024 4K pages clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: ts fid vid ttp tm stc
processor : 1 vendor_id : AuthenticAMD cpu family : 15 model : 75 model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4600+ stepping : 2 cpu MHz : 2400.000 cache size : 512 KB physical id : 0 siblings : 2 core id : 1 cpu cores : 2 fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8_legacy bogomips : 4803.64 TLB size : 1024 4K pages clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: ts fid vid ttp tm stc
when loaded.
Node that the CPU MHz went from 1000 when idle to 2400 when loaded. (There are 2 cpus showing because it's a dual core processor in my desktop).
Last edited by knappster on Mon May 26, 2008 6:23 am, edited 1 time in total.
|
|
Top |
|
|
graysky
|
Posted: Mon May 26, 2008 6:09 am |
|
Joined: Wed Dec 10, 2003 8:31 pm
Posts: 1996
Location:
/dev/null
|
/proc/info must be unique to your processor and above. My Athlon 1900+ has no such file/dir:
Code: $ cat /proc/info cat: /proc/info: No such file or directory I do have a /proc/cpuinfo which returns: Code: processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 6 model name : AMD Athlon(tm) MP 1900+ stepping : 2 cpu MHz : 1597.572 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mp mmxext 3dnowext 3dnow up ts bogomips : 3196.60
_________________ Retired KM user (R4 - R6.04); friend to LH users.
|
|
Top |
|
|
knappster
|
Posted: Mon May 26, 2008 6:32 am |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
Sorry, typo (fixed now)
Is your processor idle at the time? There are some options to manually change the frequency to preset values. For example:
Code: powersave -f should give you full speed on the cpuinfo. Code: powersave -l should give you a reduced speed on cpuinfo (1000 MHz). Code: powersave -A should put it back into dynamic mode. If you just want to see the current speed without any other details Code: powersave -r should provide that. When I run it, every now and then it shows a bogus speed, but usually it is correct.
|
|
Top |
|
|
cliffsjunk
|
Posted: Tue May 27, 2008 2:40 am |
|
Joined: Fri Sep 15, 2006 12:16 pm
Posts: 292
|
Hmm something fishy here...
The motherboard supports it on chips that have it, but it doesn't show that
option when my Pentium D 940 is plugged in.
/proc/cpuinfo always shows the max speed and does not show the est flag.
BTW that is R5F27 not R6F27.
_________________ R5F27 using R5F1 Nvidia drivers
HD-5500 analog from NTSC Sat Rx, with OTA DVB too
GeForce MX-440 SVideo tvout to a TV
Older dual core 3.4ghz Intel CPU
Asus P5PE-VM Motherboard
2 GB RAM
1 TB LVM2
VirtualBox
Samba
|
|
Top |
|
|
tjc
|
Posted: Tue May 27, 2008 6:36 am |
|
Joined: Thu Mar 25, 2004 11:00 am
Posts: 9551
Location:
Arlington, MA
|
cliffsjunk wrote: BTW that is R5F27 not R6F27.
Yeah, R6 anything would indicate ownership of a time machine.
Just in case you _do_ have a working time machine, the next time you take a jaunt into the future could you check some stock prices for me?
|
|
Top |
|
|
knappster
|
Posted: Tue May 27, 2008 6:49 am |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
cliffsjunk wrote: Hmm something fishy here...
The motherboard supports it on chips that have it, but it doesn't show that option when my Pentium D 940 is plugged in.
/proc/cpuinfo always shows the max speed and does not show the est flag.
BTW that is R5F27 not R6F27.
I'm not sure what to make of this. It may be worth waiting for R5.5 though, because Cecil said powersave would already be installed on it.
I also wanted to report that while my KM box ran perfectly well all day Sunday and Monday once I enabled powersaved, it has locked up twice in the past 12 hours. It is entirely possible that it is coincidence because the weather was very hot yesterday, which could have caused issues with the power supply, etc. For now I have disabled it, acpid, and hald and will see how it goes from here. I think this was the first lockup I've had since I set the noapic flag in lilo.
I hate when computers lockup without a clue as to why
|
|
Top |
|
|
cliffsjunk
|
Posted: Tue May 27, 2008 3:20 pm |
|
Joined: Fri Sep 15, 2006 12:16 pm
Posts: 292
|
I had a new motherboard that was flaky when hot and it turned
out to be that it was running real close to the BIOS shutdown
temp which I had set conservatively low during setup. All that I
had to do was bump the shutdown temp a little higher and the
problem went away.
If you suspect that it heat related, you can always put the PC
in a hot box for 24 hours. A cardboard box works well for this
and the heat from the PC is all that is needed. Run memtest
and/or Knoppix-running-a-video for 24 hours and monitor the
temp inside the box (but not touching the PC) with a simple
thermometer. If it passes this test at 10 degrees hotter than
it will ever be under normal circumstances (say 95 degrees
Fahrenheit), then you have a good system (at least you don't
have heat related problems). Open the cardboard box enough
to set it to this temp during the hot part of the day. If you
can get it to fail consistently, then the process of elimination
can point to which component is marginal.
Before you do this, make sure that you blow every spec of dust
out of the PC case (automotive air compressor is best, but a
leaf blower works in a pinch). About 50% of the heat problems
that I fix for friends are dust related. And verify that your PC
case has good circulation. Add an extra case fan or two if you
are uncertain about the PC case cooling. A marginally cooled case going into a hot box is a recipe for disaster.
You may need to up the BIOS shutdown temp while doing this
test.
Warning so that I can say I warned you!!! You need to be
cautious about letting it get too hot and don't run it this way
for weeks. Also, many modern hard drives keep a count of the
max temperature that it ran and this may void the hard drive
warrenty. If any of the hardware is too dear to you to bear a
flaky system becoming a boat anchor then do not do this.
At any rate...
good luck.
Cliff
_________________ R5F27 using R5F1 Nvidia drivers
HD-5500 analog from NTSC Sat Rx, with OTA DVB too
GeForce MX-440 SVideo tvout to a TV
Older dual core 3.4ghz Intel CPU
Asus P5PE-VM Motherboard
2 GB RAM
1 TB LVM2
VirtualBox
Samba
|
|
Top |
|
|
hartsu
|
Posted: Thu Jul 10, 2008 10:06 am |
|
Joined: Tue Apr 18, 2006 3:10 pm
Posts: 24
Location:
Finland SW
|
Hey,
I just install R5.5 ( auto install ) and i notice that powersaved not working anymore.
It works fine in R5F1 / knappster great instructions , same hardware .
Mobo ASUS M2NPV-VM ACPI BIOS Revision 1201 (09/20/2007)
Cpu AMD Athlon(tm) X2 Dual Core Processor BE-2400
code :
root@harpotv:/home/mythtv# powersave -c
liblazy (liblazy_dbus_send_method_call:97): Received error reply: Method "GetCPUFreqGovernor" with signature "" on interface "org.freedesktop.Hal.Device.CPUFreq" doesn't exist
Could not get current CPUFreq policy.
code:
root@harpotv:/home/mythtv# powersave -A
liblazy (liblazy_dbus_send_method_call:97): Received error reply: Method "SetCPUFreqGovernor" with signature "s" on interface "org.freedesktop.Hal.Device.CPUFreq" doesn't exist
Could not set CPUFreq policy to dynamic.
Any idea ? I want my cool@silent box back
|
|
Top |
|
|
knappster
|
Posted: Thu Jul 10, 2008 10:27 am |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
It was my understanding that Cecil was including powersave in R5.5. Did you have to install it?
Did you do an auto-install or auto-upgrade?
It appears to have something to do with the hal daemon, but I don't know what it is or what it does
|
|
Top |
|
|
hartsu
|
Posted: Thu Jul 10, 2008 12:14 pm |
|
Joined: Tue Apr 18, 2006 3:10 pm
Posts: 24
Location:
Finland SW
|
knappster wrote: It was my understanding that Cecil was including powersave in R5.5. Did you have to install it? Did you do an auto-install or auto-upgrade? It appears to have something to do with the hal daemon, but I don't know what it is or what it does
I do auto-install , i want to start clean table.
I try to install powersaved but apt-get install say that powersaved is already the newest version.
Maybe i have try to use Cpufrequtils because R5.5 newer kernel ?
wiki.archlinux.org write :
IMPORTANT
The powersaved package has been replaced by Pm-utils (hibernate/suspend) and Cpufrequtils (CPU frequency scaling). Please see those wikis for up-to-date information.
|
|
Top |
|
|
hartsu
|
Posted: Mon Jul 14, 2008 9:22 am |
|
Joined: Tue Apr 18, 2006 3:10 pm
Posts: 24
Location:
Finland SW
|
I try cpufrequtils , and its working .
Code: root@harpotv:/home/mythtv# cpufreq-info cpufrequtils 002: cpufreq-info (C) Dominik Brodowski 2004-2006 Report errors and bugs to linux@brodo.de, please. analyzing CPU 0: driver: powernow-k8 CPUs which need to switch frequency at the same time: 0 1 hardware limits: 1000 MHz - 2.30 GHz available frequency steps: 2.30 GHz, 2.20 GHz, 2.00 GHz, 1.80 GHz, 1000 MHz available cpufreq governors: userspace, powersave, ondemand, conservative, per formance current policy: frequency should be within 1000 MHz and 2.30 GHz. The governor "ondemand" may decide which speed to use within this range. current CPU frequency is 1000 MHz (asserted by call to hardware). analyzing CPU 1: driver: powernow-k8 CPUs which need to switch frequency at the same time: 0 1 hardware limits: 1000 MHz - 2.30 GHz available frequency steps: 2.30 GHz, 2.20 GHz, 2.00 GHz, 1.80 GHz, 1000 MHz available cpufreq governors: userspace, powersave, ondemand, conservative, per formance current policy: frequency should be within 1000 MHz and 2.30 GHz. The governor "ondemand" may decide which speed to use within this range. current CPU frequency is 1000 MHz (asserted by call to hardware).
Its not so smooth than powersaved , but i dont try all settings and different governors yet .
Cpu temperature and cpu-fan noise is OK again
|
|
Top |
|
|
knappster
|
Posted: Fri Jul 18, 2008 7:41 am |
|
Joined: Thu Sep 27, 2007 5:44 pm
Posts: 580
|
Any more progress with cpufrequtils hartsu? I am experiencing the same thing as you with powersave in 5.5. It was installed by default, but it spits out errors. I haven't tried cpufrequtils yet because I have bigger problems to deal with first
|
|
Top |
|
|