Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

Are there any apps I an install on client machines that will generate a report for me what app used what % of the CPU. We have some machines that are running, at times, VERY slowly. The machine will run really poorly then boom, back up to full speed. There isn't usually enough time to check Task Manager real quickly to see what is running, not to mention the majority of the time there are people using the computer that don't know what the Task Manager is. ;)

It would be really nice to take a look at some logs and see if, maybe, the anti virus is randomly taking alot of CPU for stretches of time. Or another application.

Thanks!

EDIT: This is for Windows XP. Sorry for the oversight. :)

share|improve this question
    
On what operating system? Clearly Windows from the Task Manager reference, but that version? –  John T Nov 19 '09 at 17:14
    
XP. Sorry for the confusion! –  bobber205 Nov 19 '09 at 17:37

4 Answers 4

Here are two free CPU meters that seem to fulfill your needs:

Spotlight on Windows

For a busy IT professional, it is nearly impossible to diagnose, troubleshoot, and resolve every component affecting Windows Operating System (OS) performance using manual methods. Without a clear view of I/O and system activity, you learn of performance issues only after problems erupt.

With its unique graphical view of the Windows OS internals, Quest® Spotlight® on Windows empowers you to quickly identify and eliminate bottlenecks in the Windows environment. Displaying the real-time flow of data within your Windows OS, Spotlight enables you to quickly identify and resolve performance problems.

SysMetrix (last version from 2006)

SysMetrix is a skinnable clock and metering application. Its purpose is to provide system metrics in a variety of interesting, useful, and cool ways. It can monitor and report on the hundreds of statistics.

share|improve this answer

This is going to sound a bit wishy-washy, because I don't have an XP system in front of me.

XP has a built in Performance Monitor. I can't recall how to get to it exactly, so you'll need to read the help - but somewhere under Computer Management or the Control Panel.

You can use that to create a graph that monitors dozens of performance items, including the processor time for each nominated process. You may need to spend some time creating graph lines for each process that might be responsible.

Before you start that, you might like to monitor, at an aggregate level, the CPU, page-faults per second, network traffic etc. I often find it isn't the CPU that is peaking, it is the virtual memory dumping to disk that causes my machine to freeze.

I don't have a solution for processes that pop up, consume all the CPU, and then disappear.

Finally, a common war story: Someone installed a CPU-intensive screen-saver on one of our file-servers once. Everyone complained that it was running slowly, but every time I rushed over to look at what it was doing in Performance Monitor, the busy task had just stopped and everything was back to normal... for another 20 minutes. Took a while before it dawned on me.

share|improve this answer

Here's the rest of the details for using perfmon/counter logs to track this down directly:

  • Right click on My Computer, select Manage
  • Expand System Tools-Performance Logs and Alerts-Counter Logs
  • Right click on it and select New Log Settings, enter a name and proceed
  • Click on the Add Counters button
  • Select local computer counters
  • Select the Process Object
  • Select All instances
  • Add at least % Processor Time
  • Close the add counters dialog
  • Change the interval to something more reasonable between 30sec-5min
  • Close and save the configuration, then start it by right clicking or using the toolbar

Personally I would start with a larger time interval and the overall CPU utilization (Processor Object), Disk Utilization, Networking, Some of the core Memory object counters and then add the Process counters that relate to whatever seems to be at the limits of capacity.

When you have captured the incidents you want to analyze you can analyze them manually by loading the files in perfmon (use View Log Data or Ctrl-L to load them).

If you have many machines to check, download PAL: http://pal.codeplex.com which has different pre-configured counter sets you can import into a Counter Logs configuration, as well as a large number of criteria that it uses to check different specific bottlenecks. More importantly the reports contain pretty specific background information on each type of counter/bottleneck.

share|improve this answer

You could install Nagios agents on Windows computers. More details here.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.