Limit CPU load of a single process

Is it possible to limit the CPU load of a certain process to e.g. 70% leaving 30% for the idle process?
I have some CPU intensive simulations with hard-disk access running over night. Does not make a difference if the job takes 4 hours or 6 hours - I am sleeping anyway. But I'd like to be nice to my beloved MacMini and keep it cool. The room temperatures are pretty hot this summer anyway and I don't know if my MacMini will withstand several hours at 100% CPU load, hard disk access and ambient temps of 38°C (100°F). Just want to be on the save side.
Is there a shell command like 'limit -cpu <process_id> 0.70' ?
Or can I clock my macMini at half the usual clock, i.e. at 700 MHz?
What about a loop: (1) 'kill -s stop <process_id>' (2) wait 3 secs (4) 'kill -s cont <process_id>' (4) wait 7 sec. - How to program that loop in a shell? Which command to set the timing?
many thanks
Markus

Thank you for your answer, etresoft.
Yes, you are right, nice +20 and nice -20 does not make any difference, because the process I want to slow down is the "only" one running (beside residual 5% system load).
Your Mac will be fine even running full-bore.
I am not sure about that.
First, the fans get loud. Unfortunately, the Mac is in my bedroom and I would prefer a system load of 50-70%.
Secondly, I had a compressor job running over night for 6 hours. In the morning, I had a very weird file system corruption. I suspect a 'thermal issue'.
http://discussions.apple.com/message.jspa?messageID=2715556
Maybe I'll do to 'kill -s start/stop'-in-a-loop script.
Unfortunately, I never wrote a shell script.
Any help how to make a loop and how to pause the script for 3 secs?
Many thanks
Markus
MacMini 1.42 GHz Mac OS X (10.3.9)

Similar Messages

100% cpu load on emagent.exe process

i'm testing latest Oracle 10g (10.2.0.1) database on virtual machine (through vmware on winxp). i've got dualcore cpu, one core is dedicated for virtual OS. when i installed oracle database with default settings on it, everything worked fine. i tried to import my tablespaces, schemas and data into it and after a while, I noticed strange behavior. system was very very slow. i run sysinternals process explorer (something like extended task manager) and discovered, that there are 3 processes named "perl.exe" that are consuming all the cpu time constantly !! each one consumes cca 33% of cpu. all these procesess were created by process "emagent.exe" even if it wasn't accessed via explorer... can anybody give me a hand with this?

i needed to get it working, so i installed it on another virtual machine, and it seems to work now...
but back to this problem>
i turned it on and the problem is still here, BUT now the database instance doesn't even start :< when i want to start it manualy through web em, it needs to provide host OS user name and password. when i enter it, it will say "wrong password" (the user account is in administrator and dba group and have right password). i also tried to create new user account (winxp) with roles administrator and dba, but it doesn't help.
maybe it's newbie question, because i'm newbie as you could imagine ;>
i uploaded emagent log files here:
http://www.samopal.eu/emagent.log
http://www.samopal.eu/emagent.trc
if someone is interested.
these are old ones. i wanted to to clear logs and try to reproduce this overload and see clearly what exactly is the cause, but because i don't know how to start this db up, i can't...
i will try to do something with it tomorow and i will put there what am i discovered then...

Reading CPU load for individual process

Hi,
I am refering to an older post:
http://forums.ni.com/ni/board/message?board.id=170&message.id=156234&query.id=20131#M156234
How can I not only get "memory usage" but also "CPU load" for a particular process from "Simple Task Monitor.vi"? This smart little program is part of the LabVIEW examples. Unfortunately, I am a .net greenhorn and wasn't able to figure out how I could make the program giving me "CPU load". Where do I know from which inputs are available for a .net constructor node?
Thanks for any hint,
Peter

Hi,
I am refering to an older post:
http://forums.ni.com/ni/board/message?board.id=170&message.id=156234&query.id=20131#M156234
How can I not only get "memory usage" but also "CPU load" for a particular process from "Simple Task Monitor.vi"? This smart little program is part of the LabVIEW examples. Unfortunately, I am a .net greenhorn and wasn't able to figure out how I could make the program giving me "CPU load". Where do I know from which inputs are available for a .net constructor node?
Thanks for any hint,
Peter

Osx 10.9 Mail & Safari High cpu load

Hello,
A few weeks ago I updated to OSX 10.9. After working a few days with it, I run into several high CPU load issues.
2 processes: mail-webmateriaal and safari-webmateriaal (dutch).
Now I use Chrome instead of Safari which solves 50% of my problem, but it is not so easy to change to another mail-client...
Does anyone have a clue about the possible cause, but even more important, a solution?
Best regards,
Johan

jvdstoel,
Ditto. I have the exact same OS 10.9 issue on a Late 2008 iMac with a 3.06 GHz Intel Core 2 Duo that is maxed out at 4 GB RAM (seemed like a lot of RAM at the time).
Mail, Safari, and Firefox--but not Chrome--whether running any 1, 2, or all 3 programs together keeps the CPU running hard at and often above 100%. This occurs even when the programs a sitting idle from the user's perspective. Disabling all extensions on Safari and Firefox did not help. Nor did running Onyx to clear all catches system wide.
Calling on anyone with a virtual PhD in Maverics (or real Apple OS 10.9 street-smarts) for suggestions.
Best,
Pete

Is there a limit to number of File items loaded to a single portal page?

Hi,
Does anyone know if there is a limit to the number of File items that can be loaded to one single portal (10.1.4) content page?
thanks,
Bret

Hi Arnaud,
Don't recall that we ever investigated a threshold value where portal was not performing anymore. It also depends on other circumstances as well :
- what kind of caching is used on the page, i.e. pages cached on system level for a period of 24 hours will generally perform well, even if they have a large number of file items. The performance penalty is paid by the person who needs to generate the page from the back-end database.
- Are all files in a single region or are they divided over several regions. I seem to recall that we use a fetcher thread for each region.
My advice :
Built a prototype, add file items to a page and start load tests. This will allow you to determine how many files can be added to a page without an unacceptable loss of performance.
Regards,
Erik

J2EE Processes using 25% CPU load without load

Hi
We are running a Portal cluster with 2 cluster application nodes with each 2 J2EE server processes and a dispatcher.
We experiences that after some time the J2EE server processes can start using typical 25% CPU without any load on the portal - after some time the load typically increase to approx. 50% CPU (still without user load) and the only "solution" is to restart the processes.
Anyone you have experienced a similar behaviour?
BR
Tom

Are you doing load balancing too somewhere?
I though I read somewhere about some replication between nodes that goes on. If they transfer their information and drop it, that is fine. But they don't. On the next round they exchange the new info and info from the last time. It just grows. Look at you DB server, check to see if one of the DB's grows with the CPU load.
The other thing is logging. I am a log freak myself but, sometimes I get into trouble by logging too much. Next time your CPU loads check *.log and temp logs, see if there are any over 10 gig.
Not that I never had one,
no,
never.

Activity monitor showing iOS process CPU load?

Twice in the past two weeks my iPhone 4s (running latest 7.1.1) suddenly became very warm, and its battery went down rapidly even when on standby. If I plugged it in, it got even warmer and it charged very slowly.
In each case I was eventually able to recover, by stopping all apps, putting the phone in airplane mode, doing a couple of full restarts, and (this seemed to be important) turning the phone off completely until it cooled down completely. Because I did several things almost at once, I can't be sure which ones were effective.
This could be a hardware issue with the battery. The fact that turning the phone off completely seemed to be an essential part of the cure suggests that. It could also be a hardware issue related to network communication, cured by Airplane Mode.
It could also be related to beta software that I was running, though that software was not active either time. The testing profile was installed though, and could have been involved. After the second incident I removed the profile as a precaution (fortunately the GA version of the software had just come out).
To help track down the cause, if it happens again, I would like to find an iOS equivalent of Activity Monitor on my Mac. I've installed several free system status apps and I have read the description of several paid ones. The ones I have tried give me a list of running processes with their priorities, and total CPU load, but so far I have not found one that shows which processes are the CPU hogs. Maybe one of the paid ones does that, but their descriptions don't claim that feature. It t doesn't make sense to spend money without getting the feature I need. If it is a software issue, looking at total CPU load while starting and stopping apps may not be a good diagnostic since the culprit could be a background or system process.
Any suggestions?

And a process that is capable of running multiple threads of execution, can concurrently use more than 1 CPU, and thus consume greater than 100% of 1 CPU.

Slow Report Processing somtimes timout, but no cpu load

Hi,
after searching different forums, the help and here I finally have to post our problem.
Running a report from our web application is very slow (5 minutes), even the the timeout message
"Error The request timed out because there has been no reply from the server for 600,000 milliseconds. " for some reports occur.
Monitoring the database server shows me a hit at the cpu load, while processing the sql, after this i see some load at the server where BOE is installed, but only for a small amount of time (<1 minute).
After this no load on any server can be seen, but the viewing page in a browser is loading and loading and loading till the timeout appears.
Running the same reports in The CR Designer works, it takes some time but it works
I'm bit unsure now where to look next, a network problem I would exclude, having a 1 Gigabit network here.
Any one here with hints where to look next ?
Thanks in advance for your help
Regards
Thomas

Hi
You need to adjust the Windows Power Plane settings in order to improve the cooling system.
<Setting Method>
Changes in Operating System:
Control Panel --> Power Options --> Change plan settings --> Change advanced power settings --> TOSHIBA Power Saver Settings --> Cooling Method
Changes in Bios (Power Management):
Intel Turbo boost Technology
Changes in Operating System:
TOSHIBA HW Setup --> CPU Frequency mode
Try this settings:
Power plane (Toshiba power saver) Battery optimized
Intel Turbo Boost technology (BIOS) - Disable
CPU frequency Always low
From my experience this should reduce the noise.

'ubd' process huge cpu load

Since updating to 7.2 I have a weird issue with a process called ubd. It has a huge and constant cpu load of between 50-99%. This generates a lot of extra heat and fan noise to accompany that. Quitting the process or rebooting doesn't help.
I read somewhere that the process is iCloud related, but I tried disabling some iCloud services but that didn't seem to help either.
The console displays these lines all the time:
16/10/2011 12:14:35.122 ubd: createSelfSignedCertificate - _createCert returned -25295
and after about 6 tries it exits with code: 254 and starts again.
I guess this isn't really normal, does anone know what might be up or how to fix this?

hamster116 wrote:
There is a certificate (not a key) in my keychain called 'com.apple.ubiquity.peer-uuid", which was marked as untrusted. I opened it's info dialog and set it to "always trust".
Good eye. I noticed the extra keychain entries but didn't think to 'trust' the certficiate. Though this has not resolved the issue of ubd spiking CPU usage and more certificates/keys being created (no matter how many I "trust").
FWIW: you can tail your ~/Library/Logs/Ubiquity/{username}/ubiquity.log and see the certficiate errors every few seconds.
[ERROR]     5041c5de5e10 [11/10/17 16:35:56.047] 13116.main copy_mme_bag:156 copyPreferredMobileMeName failed
[ERROR]     5041c5dfac36 [11/10/17 16:35:56.047] 13116.main ubd_main:2030 null personid
[ERROR]     5041c70a0399 [11/10/17 16:35:56.067] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
[ERROR]     5041c79acafc [11/10/17 16:35:56.076] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
--[ERROR]     50422bafc893 [11/10/17 16:35:57.755] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
[ERROR]     5042825ac3de [11/10/17 16:35:59.209] 13116.main createSelfSignedCertificate:52 error: Error Domain=NSOSStatusErrorDomain Code=-25299 "The operation couldn’t be completed. (OSStatus error -25299.)" ( / The item already exists.)
[ERROR]     5042825ff022 [11/10/17 16:35:59.209] 13116.main create_new_identity_unsafe:989 failed to find an existing identity and failed to create a new one
[ERROR]     5042be133f3f [11/10/17 16:36:00.211] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
[ERROR]     5042e63362ea [11/10/17 16:36:00.884] 13116.main createSelfSignedCertificate:52 error: Error Domain=NSOSStatusErrorDomain Code=-25299 "The operation couldn’t be completed. (OSStatus error -25299.)" ( / The item already exists.)
[ERROR]     5042e634918f [11/10/17 16:36:00.884] 13116.main create_new_identity_unsafe:989 failed to find an existing identity and failed to create a new one
[ERROR]     504321ee3444 [11/10/17 16:36:01.886] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
[ERROR]     50440def58e3 [11/10/17 16:36:05.846] 13116.main createSelfSignedCertificate:52 error: Error Domain=NSOSStatusErrorDomain Code=-25299 "The operation couldn’t be completed. (OSStatus error -25299.)" ( / The item already exists.)
[ERROR]     50440df08e7a [11/10/17 16:36:05.846] 13116.main create_new_identity_unsafe:989 failed to find an existing identity and failed to create a new one
[ERROR]     504449a35eb8 [11/10/17 16:36:06.847] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching
[ERROR]     5044a1993f0e [11/10/17 16:36:08.323] 13116.main createSelfSignedCertificate:52 error: Error Domain=NSOSStatusErrorDomain Code=-25299 "The operation couldn’t be completed. (OSStatus error -25299.)" ( / The item already exists.)
[ERROR]     5044a19a6e77 [11/10/17 16:36:08.323] 13116.main create_new_identity_unsafe:989 failed to find an existing identity and failed to create a new one
[ERROR]     5044dd4c1e33 [11/10/17 16:36:09.325] 13116.main find_existing_identity_unsafe:963 error -25300 from SecItemCopyMatching

High CPU load with no proof from processes

I have a problem with my CPU load monitor.
In Activity Monitor, the processes that are showing do not reflect my actual CPU load.
At any given time the highest process shows 3% cpu load, but the percentages under the cpu tab show that the cpu is running at 50% (basically one full core processing SOMETHING).
Now I need to figure out what this "Something" is and kill it?
Anyone have any suggestions on how to find what bogging my system down?

hey, i know someone linked the other post i created, however ill give you the rundown quickly with out my babble (i was hyper at the time :p)
Go into your sharing control panel, services, turn OFF windows sharing. It seems for some reason teh netbios daemon is loading over and over again, most likely due to an error. it is restarting a few times a second creating the loads, in my acse atleast.
let me know if it works for you, if so im going to contact apple just to inform them since the tech I spoke to had no idea what it was.
good luck!

(Linux, SL510, NSLDEGE) No processes, 100% CPU-load reboot 1% CPU-load

I have a very big problem:
I bought a Lenovo ThinkPad SL510 NSLDEGE (2847-DEG) on this (german) website with no installed operating system. At first, I installed Ubuntu GNU/Linux 10.04 which worked excellently at the first view.
But then the big problem appeared the first time: suddenly my CPU-load grew from ca. 3% to 100% (because I started an editor or I opened a new firefox-tab) and I couldn't use my ThinkPad in any way, until the system crashed and I had to reboot. After the reboot all worked fine.
This problem appeares intermittent and I always see my last chance in rebooting the entire operating system. I've already tried different Linux-Distributions, but this did not solve my problem. And I've already tested my random access memory (RAM) with Memtest86.
Thus, my question is: Do you have an idea, what could cause this problem? What could be the reason for my CPU-crashes?
Greetings, Mekeor

I'm sure that this is not a Linux-problem!
Furthermore I wrote a bug report on launchpad (the ubuntu bug-report-website) here: https://bugs.launchpad.net/ubuntu/+bug/615792 . And here is my first forum-entry in the german ubuntu-forums (forum.ubuntuusers.de).
Besides this phenomenon, I think, is a specific characteristic of my Model (SL510, NSLDEGE)!
mekeor

Measuring CPU Load

Hi,
I have an X-serve with the following configuration:
Hardware Overview:
Model Name: Xserve
Model Identifier: Xserve1,1
Processor Name: Dual-Core Intel Xeon
Processor Speed: 3 GHz
Number Of Processors: 2
Total Number Of Cores: 4
I'm trying to figure out the best way to measure CPU load so I can graph it. When I run top I see the following:
Processes: 58 total, 4 running, 54 sleeping... 407 threads 14:10:53
Load Avg: 0.10, 0.08, 0.08 CPU usage: 0.72% user, 0.96% sys, 98.32% idle
The load average is low but my question is, what is a high load average for a 2 processor, 4 core box? I'm just wondering if I need to allow for a higher load average than 1.0 for a 2 processor, 4 core box. What would be the range of high CPU loads for my X-serve? Would it be double so anything hovering around 2.0 would be high or do I need to take into account the cores too when ascertaining the load average from top? Thanks for any help.

What constitutes a high load average depends a lot on your application.
The load average value indicates the number of active processes in the CPU queue. For your example of 0.10, that means that 1 CPU was busy for 10% of the time - not a lot, really.
Now, in the old days, a load average of 1.00 indicated the CPU was busy 100% of the time, and anything higher than that indicated an overloaded system.
In theory, since you have 4 cores you could think that a load average of 4.00 indicated all 4 cores were busy 100% of the time and that's your limit, but things are not so cut-and-dried in todays multi-processor, multi-core world.
That's only true if the applications you were running are 100% multi-threaded and able to scale perfectly across multiple processors. That's not always the case. You may find that any particular app only runs on a single core, so that core could be maxxed out while the other three were idle. You'd still have a load average of 1.00 yet your application is CPU-bound.
This is less of an issue if you're running multiple applications - the OS does a pretty good job of scheduling different applications across different cores, but that's why I say it depends on what your server is running. You know that if you have a number more than the number of cores/processors in your system that you have a problem, but you may also have a problem well before then.
So the only real way to tell is to profile your system - look at what you get for typical values when things are good and work out an acceptable differential from that.

How do you spread CPU load over more cores?

Hi
I recently bought a macbook pro quadcore i7 because my 2007 macbook couldn't cope with the demands of Logic when using it rewired with Ableton Live, and had lots of CPU overload messages.
I am really disappointed that having spent so much money to overcome this problem, it has only slightly improved the amount of overload messages. I read the article below about how to spread CPU power over the cores, now I have 8. I still can't figure out though how to control which channels are using which core and spreading the CPU load to cores 5-8 which don't seem to be being used. Any ideas?
Can I rewire from ableton to different cores?
In my project, the 2nd core keeps overloading whilst the others are down. Is it to do with aux channels, master channels, buses etc?
Or should I upload my memory to 8GB RAM?
Thanks
Steve
Here's the article:
About core activity
Logic breaks tasks down into threads that can be assigned to a single core for execution. It's not possible to split a single thread across multiple cores, although a single core may run multiple threads at the same time. This is one reason that you may sometimes see uneven load distributions across the available cores on your computer. If one thread is very processor-intensive, the core running that thread may show much more activity than the cores running less demanding tasks. This is the main reason that you may get "Core Audio Overload" warning messages even though some of the cores on your system may not show significant activity.
To monitor Logic as it uses the cores on your system, double-click the CPU meter in the Transport area.
A window displays a meter for each core.
The same concept applies to channel strips as well. Each channel strip—and all plug-ins it contains—is processed by one core. If you are using many processor-intensive plug-ins on a single channel strip, the core running it may show heavier activity. This is especially true if you are working at high sample rates (88.2 kHz or higher).
Redistributing core loads
If you find that a single core has much higher activity, use these strategies to redistribute the load.
In the example below, there is a single Software Instrument channel strip with plug-ins routed directly to Output 1-2. The channel strip for Output 1-2 has a large number of plug-ins.
In this example there is activity on two cores: the left core is handling the Software Instrument while the next core processes the Output 1-2 channel strip. You can use routing capabilities in Logic to distribute the load to more cores without changing the sound, as in the next image.
The resulting signal chain is Inst 1 > Aux 1 > Aux 2 > Out 1-2, as shown in this table:
Channel strip
Input
Output
Inst 1
Bus 1
Aux 1
Bus 1
Bus 2
Aux 2
Bus 2
Out 1-2
Also, the first four plug-ins have moved from Output 1-2 to Aux 1 and the remaining four plug-ins to Aux 2.
After adjusting the volumes to the respective channel strips appropriately, the result sounds identical, but now the load is distributed among four cores instead of two. You can use this concept to redistribute plug-ins from Audio channel strips, Software Instruments, Auxes, and Outputs (as in this example).
In the image above, there is still a very heavy load on a single core, even though it's using the routing described above. Notice the Inst 1 channel strip is selected. When a Software Instrument track is selected in the Arrange window, Logic enters "Live Input Mode". In this mode, Logic turns on every plug-in in the channel strip's signal path so the channel strip can be played live from a MIDI controller. All of this must be handled by a single thread and therefore just one core. It's a good idea to keep a track that uses an Audio channel strip selected in the Arrange window, unless you're actively playing or recording to a Software Instrument.

OS X Snow Leopard introduced Grand Central, where the OS will distribute the load better to more cores, however it takes time for software to adapt to these changes and make a rewrite of the codebase and test it out, Apple's included.

T500 CPU Load at 100 always

My T500 is always at 100% CPU load what is causing this and how do I stop it? My T61 used to do this also.
ThinkPad Twist S230U-
OS-Windows 8.1 Pro 64-bit
CPU-Intel Core i7 3517U @ 1.90GHz Ivy Bridge 22nm Technology
RAM-8.00GB Single-Channel DDR3 @ 665MHz (9-9-9-24)
Motherboard-LENOVO 3347CTO (CPU Socket - U3E1)
Graphics-Wide viewing angle & High density FlexView Display 1366x768 (1366x768@60Hz) Intel HD Graphics 4000 (Lenovo)
Storage-298GB Seagate ST320LT007-9ZV142 (SATA) 22GB SanDisk SSD U100 24GB (SSD) 29GB SDHC Card (SCSI)
Audio-Realtek High Definition Audio

Hi parmardian,
Please turn on Task Manager by "Crtl + Shift + Esc" key, then ou should able to see which process eats up your CPU power.
Fred
T61 + X201i

CPU load higher after applying in 147440 kernel on Oracle database servers

Hi all,
Just wanted to know if anyone else is seeing an Increased CPU load with higher peaks and for longer periods of time, especially on single thread processes on oracle database servers. Our Oracle Database versions vary from Oracle 10 to Oracle 11.1.0.7 11.2.0.2 10.2.0.4. The Hardware includes Sunfire V245's, V440 and T5240. We saw a slight performance hit when going to 147440-02 at the first of Oct, but no big deal. Then we applied 147440-04 in the first part of Nov. And of course we got CPU panics and Zfs file system stopped mounting basically a complete disaster. Oracle provided and IDR for the CPU panic and then later release 147440-06. Which resolved the CPU panics and zfs file system mounts. But now our load average on is 1 to 3 points higher during load, grid control is constantly complaining of 100% utilization. I have had a call open since the first of November with oracle/sun support and have provided guds performance stats on before and after Kernel change. They are still saying that its not an issue. It is so bad we had to roll back to a July kernel on one of the servers.

Just to let you all know what we have found. After 2 1/2 months of working with Oracle Kernel team, they have identified the Oracle Enterprise Manager (emagent) (Grid control Agent) was using nearly 60000 threads under the 147440 Kernel and only 1500 threads under an older kernel.
I googled the solution for the emagent, but I continue to press Oracle Kernel Engineers to work with their software division, to identify other products that might cause similar problems.
http://oracle-dba-quickanswers.blogspot.com/2011/10/emagent-consuming-very-high-cpu.html

Limit CPU load of a single process

Similar Messages

Maybe you are looking for