Re: [Clonezilla-live] Thermal Problems and a Solution
Hello, On 16 January 2013 01:42, Steven Shiau wrote: > Hi Andreas, > Thanks for sharing that. > Your solution is great. For sure it a great feature we should add in the > expert mode. Besides this, I think another option solution is to find a > way to know the temperature is too high then accordingly reduce the CPU > speed. This is something that should be handled by the kernel. On many laptops (and iMacs, ..) the cooling is simply insufficient and it is required as part of normal operation to lower the CPU and GPU frequency to keep the system running. There should be ACPI tables that tell the kernel to lower the CPU speed when a thermal zone reaches too high temperature. There might be a bug in the tables which Windows drivers work around by consulting their own tables (common), the thermal driver might not be loaded or the kernel does not act on the thermal zone info. Maybe an additional package is required to monitor and act on these thermal readings. It is also possible that a bios update would resolve this issue on some affected systems. Thanks Michal -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnnow-d2d ___ Clonezilla-live mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/clonezilla-live
Re: [Clonezilla-live] Thermal Problems and a Solution
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi Andreas, On 01/18/2013 06:29 AM, Andreas Winkelbauer wrote: > Hi Steven! > > It would of course be optimal to throttle the CPU clock according to the > current temperature, but this is exactly what the kernel attempts to do. > Unfortunately this approach may fail under heavy load for several > reasons (BIOS bugs, bad ACPI implementation, inaccurate sensors, > insufficient cooling either by design or through aging, dust, etc.). > > Therefore I think it is appropriate to give the user an (expert) option > to reduce the CPU clock. It would also be possible to monitor the kernel > messages for potential thermal problems (usually something like > "temperature above threshold") and to reduce the CPU clock automatically > if such messages show up. However, I am not sure if such a behavior > should be enabled by default (because even if the temperature is high it > might not always be necessary to permanently reduce the CPU clock, > thereby unnecessarily degrading the performance). > > For an expert option concerning the CPU clock, I have thought of the > following implementation possibilities: > > * Find available frequencies and let the user select one of them. > > * Let the user select 50%, 60%, ..., 100% of the maximum frequency. > > * Let the user enter the percentage of the maximum frequency to be used. > > I would prefer the second option, since it should be easy to understand > by the user and also simple to implement. What do you think? Agree. We will implement this in the future release. Thanks again for this idea. Steven. > > > Cheers, > Andreas > > > On 2013-01-16 01:42, Steven Shiau wrote: >> Hi Andreas, >> Thanks for sharing that. >> Your solution is great. For sure it a great feature we should add in the >> expert mode. Besides this, I think another option solution is to find a >> way to know the temperature is too high then accordingly reduce the CPU >> speed. >> We should discuss more about this. >> Thanks again. >> >> Steven. >> >> On 2013/1/16 上午 07:47, Andreas Winkelbauer wrote: >>> Dear all, >>> >>> several users have reported hardware failures with clonezilla due to >>> thermal problems, in particular with notebooks (see [1-3]). I also have >>> seen these issues with my HP 2540p laptop. It seems that the cooling >>> system of some notebooks is insufficient when the system is run at full >>> CPU load for an extended period of time (this can be minutes, hours, or >>> even days before a shut down to prevent overheating will occur). >>> >>> Since up to now no solution has been posted here, I'll explain how I >>> solved the thermal problems for me. A possible solution is to set the >>> CPU to a fixed, lower operating frequency using cpufreq-set from >>> cpufrequtils (which is installed by default). This effectively reduces >>> heat at the cost of a (relatively small) performance penalty. >>> Specifically, what I did was (for a system with 4 CPUs): >>> >>> $ sudo cpufreq-set -c 0 -f 1600MHz >>> $ sudo cpufreq-set -c 1 -f 1600MHz >>> $ sudo cpufreq-set -c 2 -f 1600MHz >>> $ sudo cpufreq-set -c 3 -f 1600MHz >>> >>> This worked for me and it should work with all reasonably modern CPUs. >>> The current settings and the available frequency steps can be read using >>> >>> $ cpufreq-info >>> >>> A few remarks: cpufreq-set automatically rounds the specified frequency >>> to the "nearest" available operating frequency (my laptop CPU, e.g., has >>> a minimum frequency of 1.2 GHz and a maximum frequency of 2.53 GHz with >>> 11 steps in total). However, I have observed that the rounding sometimes >>> does not work if -f specifies exactly one of the available frequencies >>> (then the next higher frequency might be used instead which seems to be >>> a bug of cpufreq-set). Using the above commands, the CPU runs at a fixed >>> frequency until changed again or until the next restart. Changes made by >>> cpufreq-set are not permanent, i.e., the changes are lost if the system >>> is rebooted. >>> >>> Another way to prevent overheating could be to use "cpulimit", but I >>> find it less suited for this particular use case and the corresponding >>> package is not installed by default. >>> >>> @Steven: I could provide a patch for clonezilla-live to include this in >>> the expert options. However, we should think about how to implement this >>> best. >>> >>> Cheers, >>> Andreas >>> >>> >>> [1] >>> http://sourceforge.net/tracker/?func=detail&aid=3090553&group_id=115473&atid=671650 >>> [2] >>> http://sourceforge.net/projects/clonezilla/forums/forum/663168/topic/3903931 >>> [3] >>> http://sourceforge.net/projects/clonezilla/forums/forum/394751/topic/3877156 > - -- Steven Shiau National Center for High-performance Computing, Taiwan. http://www.nchc.org.tw Public Key Server PGP Key ID: 4096R/47CF935C Fingerprint: 0240 1FEB 695D 7112 62F0 8796 11C1 12DA 47CF 935C -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with undefined - http:/
Re: [Clonezilla-live] Thermal Problems and a Solution
Hi Steven! It would of course be optimal to throttle the CPU clock according to the current temperature, but this is exactly what the kernel attempts to do. Unfortunately this approach may fail under heavy load for several reasons (BIOS bugs, bad ACPI implementation, inaccurate sensors, insufficient cooling either by design or through aging, dust, etc.). Therefore I think it is appropriate to give the user an (expert) option to reduce the CPU clock. It would also be possible to monitor the kernel messages for potential thermal problems (usually something like "temperature above threshold") and to reduce the CPU clock automatically if such messages show up. However, I am not sure if such a behavior should be enabled by default (because even if the temperature is high it might not always be necessary to permanently reduce the CPU clock, thereby unnecessarily degrading the performance). For an expert option concerning the CPU clock, I have thought of the following implementation possibilities: * Find available frequencies and let the user select one of them. * Let the user select 50%, 60%, ..., 100% of the maximum frequency. * Let the user enter the percentage of the maximum frequency to be used. I would prefer the second option, since it should be easy to understand by the user and also simple to implement. What do you think? Cheers, Andreas On 2013-01-16 01:42, Steven Shiau wrote: > Hi Andreas, > Thanks for sharing that. > Your solution is great. For sure it a great feature we should add in the > expert mode. Besides this, I think another option solution is to find a > way to know the temperature is too high then accordingly reduce the CPU > speed. > We should discuss more about this. > Thanks again. > > Steven. > > On 2013/1/16 上午 07:47, Andreas Winkelbauer wrote: >> Dear all, >> >> several users have reported hardware failures with clonezilla due to >> thermal problems, in particular with notebooks (see [1-3]). I also have >> seen these issues with my HP 2540p laptop. It seems that the cooling >> system of some notebooks is insufficient when the system is run at full >> CPU load for an extended period of time (this can be minutes, hours, or >> even days before a shut down to prevent overheating will occur). >> >> Since up to now no solution has been posted here, I'll explain how I >> solved the thermal problems for me. A possible solution is to set the >> CPU to a fixed, lower operating frequency using cpufreq-set from >> cpufrequtils (which is installed by default). This effectively reduces >> heat at the cost of a (relatively small) performance penalty. >> Specifically, what I did was (for a system with 4 CPUs): >> >> $ sudo cpufreq-set -c 0 -f 1600MHz >> $ sudo cpufreq-set -c 1 -f 1600MHz >> $ sudo cpufreq-set -c 2 -f 1600MHz >> $ sudo cpufreq-set -c 3 -f 1600MHz >> >> This worked for me and it should work with all reasonably modern CPUs. >> The current settings and the available frequency steps can be read using >> >> $ cpufreq-info >> >> A few remarks: cpufreq-set automatically rounds the specified frequency >> to the "nearest" available operating frequency (my laptop CPU, e.g., has >> a minimum frequency of 1.2 GHz and a maximum frequency of 2.53 GHz with >> 11 steps in total). However, I have observed that the rounding sometimes >> does not work if -f specifies exactly one of the available frequencies >> (then the next higher frequency might be used instead which seems to be >> a bug of cpufreq-set). Using the above commands, the CPU runs at a fixed >> frequency until changed again or until the next restart. Changes made by >> cpufreq-set are not permanent, i.e., the changes are lost if the system >> is rebooted. >> >> Another way to prevent overheating could be to use "cpulimit", but I >> find it less suited for this particular use case and the corresponding >> package is not installed by default. >> >> @Steven: I could provide a patch for clonezilla-live to include this in >> the expert options. However, we should think about how to implement this >> best. >> >> Cheers, >> Andreas >> >> >> [1] >> http://sourceforge.net/tracker/?func=detail&aid=3090553&group_id=115473&atid=671650 >> [2] >> http://sourceforge.net/projects/clonezilla/forums/forum/663168/topic/3903931 >> [3] >> http://sourceforge.net/projects/clonezilla/forums/forum/394751/topic/3877156 signature.asc Description: OpenPGP digital signature -- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnmore_122712___ Clonezilla-live mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/clonezilla-live
Re: [Clonezilla-live] Thermal Problems and a Solution
Hi Andreas, Thanks for sharing that. Your solution is great. For sure it a great feature we should add in the expert mode. Besides this, I think another option solution is to find a way to know the temperature is too high then accordingly reduce the CPU speed. We should discuss more about this. Thanks again. Steven. On 2013/1/16 上午 07:47, Andreas Winkelbauer wrote: > Dear all, > > several users have reported hardware failures with clonezilla due to > thermal problems, in particular with notebooks (see [1-3]). I also have > seen these issues with my HP 2540p laptop. It seems that the cooling > system of some notebooks is insufficient when the system is run at full > CPU load for an extended period of time (this can be minutes, hours, or > even days before a shut down to prevent overheating will occur). > > Since up to now no solution has been posted here, I'll explain how I > solved the thermal problems for me. A possible solution is to set the > CPU to a fixed, lower operating frequency using cpufreq-set from > cpufrequtils (which is installed by default). This effectively reduces > heat at the cost of a (relatively small) performance penalty. > Specifically, what I did was (for a system with 4 CPUs): > > $ sudo cpufreq-set -c 0 -f 1600MHz > $ sudo cpufreq-set -c 1 -f 1600MHz > $ sudo cpufreq-set -c 2 -f 1600MHz > $ sudo cpufreq-set -c 3 -f 1600MHz > > This worked for me and it should work with all reasonably modern CPUs. > The current settings and the available frequency steps can be read using > > $ cpufreq-info > > A few remarks: cpufreq-set automatically rounds the specified frequency > to the "nearest" available operating frequency (my laptop CPU, e.g., has > a minimum frequency of 1.2 GHz and a maximum frequency of 2.53 GHz with > 11 steps in total). However, I have observed that the rounding sometimes > does not work if -f specifies exactly one of the available frequencies > (then the next higher frequency might be used instead which seems to be > a bug of cpufreq-set). Using the above commands, the CPU runs at a fixed > frequency until changed again or until the next restart. Changes made by > cpufreq-set are not permanent, i.e., the changes are lost if the system > is rebooted. > > Another way to prevent overheating could be to use "cpulimit", but I > find it less suited for this particular use case and the corresponding > package is not installed by default. > > @Steven: I could provide a patch for clonezilla-live to include this in > the expert options. However, we should think about how to implement this > best. > > Cheers, > Andreas > > > [1] > http://sourceforge.net/tracker/?func=detail&aid=3090553&group_id=115473&atid=671650 > [2] > http://sourceforge.net/projects/clonezilla/forums/forum/663168/topic/3903931 > [3] > http://sourceforge.net/projects/clonezilla/forums/forum/394751/topic/3877156 > > > > > -- > Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery > and much more. Keep your Java skills current with LearnJavaNow - > 200+ hours of step-by-step video tutorials by Java experts. > SALE $49.99 this month only -- learn more at: > http://p.sf.net/sfu/learnmore_122612 > > > > ___ > Clonezilla-live mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/clonezilla-live > -- Steven Shiau National Center for High-performance Computing, Taiwan. http://www.nchc.org.tw Public Key Server PGP Key ID: 4096R/47CF935C Fingerprint: 0240 1FEB 695D 7112 62F0 8796 11C1 12DA 47CF 935C -- Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery and much more. Keep your Java skills current with LearnJavaNow - 200+ hours of step-by-step video tutorials by Java experts. SALE $49.99 this month only -- learn more at: http://p.sf.net/sfu/learnmore_122612 ___ Clonezilla-live mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/clonezilla-live
[Clonezilla-live] Thermal Problems and a Solution
Dear all, several users have reported hardware failures with clonezilla due to thermal problems, in particular with notebooks (see [1-3]). I also have seen these issues with my HP 2540p laptop. It seems that the cooling system of some notebooks is insufficient when the system is run at full CPU load for an extended period of time (this can be minutes, hours, or even days before a shut down to prevent overheating will occur). Since up to now no solution has been posted here, I'll explain how I solved the thermal problems for me. A possible solution is to set the CPU to a fixed, lower operating frequency using cpufreq-set from cpufrequtils (which is installed by default). This effectively reduces heat at the cost of a (relatively small) performance penalty. Specifically, what I did was (for a system with 4 CPUs): $ sudo cpufreq-set -c 0 -f 1600MHz $ sudo cpufreq-set -c 1 -f 1600MHz $ sudo cpufreq-set -c 2 -f 1600MHz $ sudo cpufreq-set -c 3 -f 1600MHz This worked for me and it should work with all reasonably modern CPUs. The current settings and the available frequency steps can be read using $ cpufreq-info A few remarks: cpufreq-set automatically rounds the specified frequency to the "nearest" available operating frequency (my laptop CPU, e.g., has a minimum frequency of 1.2 GHz and a maximum frequency of 2.53 GHz with 11 steps in total). However, I have observed that the rounding sometimes does not work if -f specifies exactly one of the available frequencies (then the next higher frequency might be used instead which seems to be a bug of cpufreq-set). Using the above commands, the CPU runs at a fixed frequency until changed again or until the next restart. Changes made by cpufreq-set are not permanent, i.e., the changes are lost if the system is rebooted. Another way to prevent overheating could be to use "cpulimit", but I find it less suited for this particular use case and the corresponding package is not installed by default. @Steven: I could provide a patch for clonezilla-live to include this in the expert options. However, we should think about how to implement this best. Cheers, Andreas [1] http://sourceforge.net/tracker/?func=detail&aid=3090553&group_id=115473&atid=671650 [2] http://sourceforge.net/projects/clonezilla/forums/forum/663168/topic/3903931 [3] http://sourceforge.net/projects/clonezilla/forums/forum/394751/topic/3877156 signature.asc Description: OpenPGP digital signature -- Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery and much more. Keep your Java skills current with LearnJavaNow - 200+ hours of step-by-step video tutorials by Java experts. SALE $49.99 this month only -- learn more at: http://p.sf.net/sfu/learnmore_122612 ___ Clonezilla-live mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/clonezilla-live
