Prime95 Vs. IntelBurnTest

ice445

Which of these apps is really better for stress testing? On an Intel system? On an AMD setup? I'm curious, because I hear differing opinions everywhere.

DSK · 11-25-2009, 03:39 | posts: 16,635 | Location: Scotland | User is Online

I usually think OCCT is better for AMD systems imo.

sajibjoarder · 11-25-2009, 04:54 | posts: 148 | Location: Dhaka,BD | User is Offline

intel burnin is for xtreem conditions. where as prime 95 doenn't give the processor that pressure. if ur system is stable with prime 95 at least 24 hours its perfectly stable system for general use gaming etc. but if u want more stable use intel burn in.

burebista · 11-25-2009, 05:06 | posts: 995 | Location: Romania | User is Offline

If you want maximum heat then Linpack clones. At least for Intel, I dunno about AMD.
For Nehalem/Lynnfield prime95 Large FFTs put ~same temps like Linpack.

Mega_Death · 11-25-2009, 05:21 | posts: 867 | Location: Michigan | User is Offline

Quote:

Originally Posted by kylzer

I usually think OCCT is better for AMD systems imo.

Have to agree. Nothing pushed the heat quicker on my 720 than OCCT.

IPlayNaked · 11-25-2009, 08:25 | posts: 1,273 | User is Offline

I use IntelBurnTest for quick checks of OC settings, and Prime95 for long-duration stability tests.

ice445 · 11-25-2009, 13:06 | posts: 1,157 | Location: Toledo, Ohio | User is Offline

I find that Intelburntest will call an OC stable that Prime95 fails in 7 minutes.

Omagana · 11-25-2009, 13:47 | posts: 849 | Location: Scotland | User is Offline

My OC is 6 hours prime stable, been running it for nearly a year now and I've never had an issue...

...tried Intelburntest last month and it crashed almost instantly, so me thinks its un-realistically harsh on the CPU.

ice445 · 11-25-2009, 14:56 | posts: 1,157 | Location: Toledo, Ohio | User is Offline

Hmm, maybe IntelBurnTest really only works on an Intel CPU.

Pyro the Dragun · 11-25-2009, 16:33 | posts: 233 | Location: Manhattan, Kansas | User is Offline

Quote:

Originally Posted by Omagana

My OC is 6 hours prime stable, been running it for nearly a year now and I've never had an issue...

...tried Intelburntest last month and it crashed almost instantly, so me thinks its un-realistically harsh on the CPU.

That's basically the point though. It's supposed to rape the CPU. If your CPU can handle a certain number of runs using IBT/Linpack apps then it should be almost rock-solid.

I've always thought of Linpack apps (IBT and LinX) as the "quick" way to tell if an overclock is stable. I'd rather not wait several hours for Prime95/OCCT to crash when Linpack can crash a system in mere minutes.

Quote:

Originally Posted by ice445

Hmm, maybe IntelBurnTest really only works on an Intel CPU.

Its my understanding that IBT takes advantage of some instruction sets on Intel CPUs, allowing the CPU to poll the RAM less and generate maximum workload/heat. I could be completely wrong on this, though.

Regardless I'm sure IBT will heat up an AMD proc plenty as well. You might try LinX, I prefer it over IBT for a few stupid reasons. (LinX seems to get my CPU slightly hotter, and seems to be a tad more tolerant of computation errors whereas IBT will just stop dead, which pisses me off when I walk away for several hours)

ice445 · 11-25-2009, 18:34 | posts: 1,157 | Location: Toledo, Ohio | User is Offline

Quote:

Originally Posted by Pyro the Dragun

That's basically the point though. It's supposed to rape the CPU. If your CPU can handle a certain number of runs using IBT/Linpack apps then it should be almost rock-solid.

I've always thought of Linpack apps (IBT and LinX) as the "quick" way to tell if an overclock is stable. I'd rather not wait several hours for Prime95/OCCT to crash when Linpack can crash a system in mere minutes.

Its my understanding that IBT takes advantage of some instruction sets on Intel CPUs, allowing the CPU to poll the RAM less and generate maximum workload/heat. I could be completely wrong on this, though.

Regardless I'm sure IBT will heat up an AMD proc plenty as well. You might try LinX, I prefer it over IBT for a few stupid reasons. (LinX seems to get my CPU slightly hotter, and seems to be a tad more tolerant of computation errors whereas IBT will just stop dead, which pisses me off when I walk away for several hours)

IBT does in fact heat up an AMD proc just fine, that's not the problem. The problem I have is the fact that an overclock that passes 10 runs of IBT flawlessly won't pass 7 minutes of Prime 95.

Pyro the Dragun · 11-25-2009, 18:48 | posts: 233 | Location: Manhattan, Kansas | User is Offline

Quote:

Originally Posted by ice445

IBT does in fact heat up an AMD proc just fine, that's not the problem. The problem I have is the fact that an overclock that passes 10 runs of IBT flawlessly won't pass 7 minutes of Prime 95.

Strange indeed. Tried OCCT with that overclock?

ice445 · 11-25-2009, 19:29 | posts: 1,157 | Location: Toledo, Ohio | User is Offline

Quote:

Originally Posted by Pyro the Dragun

Strange indeed. Tried OCCT with that overclock?

I think I did a while ago, I'll try it after I do a full Prime run. I was fiddling with it earlier and got Prime to run just fine for almost 2 hours, going to let it run overnight.

UnrealGaming · 11-25-2009, 19:48 | posts: 1,431 | Location: Serbia | User is Offline

Use Prime. As you said yourself, sometimes Intel burn passes , while prime fails. I only use intel burn to test speed ( Gflops ). Prime for stability .

MAD-OGRE · 11-25-2009, 20:14 | posts: 2,262 | Location: Alabama | User is Offline

I use 20 loops of IBT @ max memory, after that its stable as a rock.