Prime95: Why, when I pull up task manager, is Prime95 only taking 50% CPU time?

[Urinal Mint]

I thought it was a 100% stress test... and I can't seem to find an answer yet

Any insight? Thanks

 

[BlackMountainCow]

Is that a P4 with HT?

 

[Urinal Mint]

Your astuteness wows even the sharpest of the sharp

Originally posted by: BlackMountainCow
Is that a P4 with HT?

 

[slag]

since hyperthreaded processors count as 2 processors (1 is virtual), it spreads the load across both processors. 100% utilization / 2 processors =50%.

I have read that you can run 2 instances and get 100% utilization, but I dont know the trick to doing that. If I have one instance running and go to click on it in the start menu to start up another instance, it doesnt do it because it is already running.

 

[myocardia]

What a name! Anyway, Urinal, I believe you have to setup two separate Prime95's, but I can't say for sure, since I don't have a P4C. All you have to do, though, is ask that same question in the Distributed Computing forum (a.k.a. the SETI forum), and I'm sure you'll get 50 correct responses as soon as it's posted.

 

[THUGSROOK]

you need 2 instances of P95 ...or turn off HT

 

[BlackMountainCow]

Your astuteness wows even the sharpest of the sharp

Well, as I'm all into AMD and just wanted to help u ... nevermind then

 

[Urinal Mint]

OK, so can this be considered a valid P95 stress test? I guess that's my main question here.




BlackMountainCow: Dude, I was just giving you a hard time

 

[Duvie]

At 50%??? Which by the way is really more then 50% as max temps will attest...You run 2 instances you may see a rise in temps of about 3-4c more....

I would say this is not a true stability test since you really need 100% load....Either just run that 1 instance ofr pime and couple it with a 6 hours divx or mpeg2 encoding, or if you could game that long.....

the long way is to:

Start - run - CMD (type in the command line) - cd.. - cd.. - cd program files/Prime95 - prime95 -a

That will start a 2nd instance and just torture test both of them....


Other good stresses are FH (2) instances for 24-48 hours....multiple instances of SETI... these are all good to stress the cpu in terms of voltage and heat, but ultimately priem95 compares results against known. errors in seti and FH will be if the system is truly flaky, but a unstable computer can still finish them but send erroroneous data you would be unaware of.....


I would also run memtest86 for a period of time (overnight....all test, cache on, etc. )

 

[slag]

My temps go from 44c up to 58.5 when running prime95 for 12 hours. If it doesnt crash in 12 hours, its not going to crash.

This is a 2.6c @ 3.27 ghz (252 mhz fsb) on an abit ic7 board using a zaalman 7000a at full speed. @ 1.625 volts