Two Identical Comps, different Commit Charge?

[lost0822]

At my work we have two identical dells comps. 2.2, 512 Ram....for some reason looking in the windows task mangaer under the performance tab one of the comps. Limit Commit Charge number is almost twice the amount of the other comp....why would this be? We feel this might be leading to a performance issue with the one comp. saving files slow, opening files slow....any suggestions?

 

[Nothinman]

Commit Charge is just the amount of used physical memory, of course they'll be different depending on what was run, opened, etc on each box.

 

[lost0822]

maybe i'm asking wrong here's another situation.
we have the same RAM in both machines, like i said the hardware is the same in both machines. We do CAD work here, if i open up a drawing on the faster comp. my cpu usage goes up to about 50% no big deal right, well on the other comp. when we open the same drawing that comp. usage goes up to 100%, basically it maxes out. why would that be?

 

[Nothinman]

Could be any number of things. Where are you opening the drawing from? Is one box setup with HT on and the other off?

 

[lost0822]

Originally posted by: Nothinman
Could be any number of things. Where are you opening the drawing from? Is one box setup with HT on and the other off?

I would doubt it, but how would i check if HT is on or off?
Even opening up AutoCad on the one comp. is slower then on the other.that goes for saving drawings as well.

 

[mikecel79]

Are the EXACT same programs loaded on each one? Are they from the same image? What about the AutoCAD version and service pack, are they the same?

Unless these are Xeon machines they won't have HT. The slowest P4 with HT was the 2.4C.

 

[Smilin]

Even the exact computers might not have the same commit charge. It depends on what is running on each of them. I would say it's very unlikely you would ever see this exactly the same.

 

[caledai]

Is it possible that there is any parasites (trojans, adware, spyware, malware, viruses, worms) on the computer, they can noticably affect the commit charge.

Also you mentioned a 'faster' machine. Is the machine just running faster or is there a difference in the hardware somewhere. The faster pc will not require as much ram to be used as it processes the data faster. Also it could be a setting anywhere on the computer - group policies, security policies, tweaks, different wallpapers, plug-ins, running services.

 

[kylef]

It really could be any number of things contributing to such a performance difference.

Poor graphics drivers can cause simple graphics calls to take exorbitant (by the cpu's standards at least) amounts of time. Try reinstalling them to the latest and re-test.

Virus scanners can cause apps with heavy file I/O traffic to slow down immensely. Try disabling any virus scanners and re-test.

Disk fragmentation can cause performance problems (but typically not trigger lots more CPU activity) when opening large files, such as huge CAD drawings. You could try to defragment your drives and re-test.

Heck, even something as zany as a poorly cooled (or uncooled) CPU can trigger the P4's thermal protection circuitry to kick in periodically, slowing things to a halt. As long as the CPU's fan is running, this shouldn't be an issue.

To really diagnose the problem, I'd download a tool called
Process Explorer from Sysinternals (freeware) and see where the CPU usage is going. It can even show you what THREAD inside of a process is using CPU time, and unlike Task Manager, device driver deferred procedure calls are reported accurately rather than listed as "idle cpu time". That helps spot broken device drivers.

Basically, run Process Explorer and then re-run the app that is so slow in the background. See if you notice anything unusual and report back.

Good luck.