Intel Burn Test...how much is 'enough'..? - Page 4 - Overclock.net

NrGx · 10-03-09

With IBT and LinX and OCCT, I consider 1 hour to be stable. I don't fold or anything though so my computer probably doesn't need to be as stable as other people's (or rather, stability isn't as critical).

LiquidForce · 10-03-09

For me if it passes IBT 10 times on max I'd say my OC is stable. At first when OCing I'll find the minimum voltage it will boot at and just keep increasing the vcore until it passes 10 runs. Normally once I get it to pass 10 runs it will be able to pass waaayyyy more until I just decide to turn it off (a day later if I'm not busy)

**CL3P20** · 10-03-09

So, for most here..IBT is working just fine for CPU stability with ~10passes...and unless full mem usage is selected..sounds like Large FTT's with Prime95 seems to still be necessary for NB stability testing.

..Anyone passing 10x loops with full mem useage on IBT and still failing Large FTT's? Just curious...

** I'll post a screen shot after voltage tweaking a bit...would love to see some others to compare FSB and GB/sec throughputs..

crashdummy35 · 10-03-09

IBT/Linx are good for quick tests to check the cpu stability. I usually do 10 regular passes, if I get through that without freezing/bsod'ing I run 10 Max passes.

After the IBT/Linx passing I run Prime95 Small FFT's for about 8-10 hours.

After passing the Small FFT's I run a Prime95 Blend for 8 hours.

I used to run only IBT until I could get 20 Max passes successfully then call it stable. After a few weeks my games started freezing and crashing; I got random errors in apps that were memory intensive.

Stabilizing the cpu oc is only one part of the oc process. That's the main thing I've learned here at ocn. Especially if you are running a high fsb.

#1 IBT/Linx for a quick "feel" for the oc's stability.
#2 Small FFT's for at least 10 hours. To ensure the cpu processes data correctly over a longer period of time without errors.
#3 Blend for at least 8 hours. To ensure the cpu and RAM are communicating correctly via the NB.

I can tell you first hand, os corruption happens under a not fully stable oc--because it was starting to happen to me. It may not happen over night but the errors in files will accumulate.

M1 Abrams · 10-03-09

""sounds like Large FTT's with Prime95 seems to still be necessary for NB stability testing.""
This isn't my full course stress menu, just a part of it.
For NB/Video stability I run 2 hours prime blend/small/large.
In the last hour of each(blend/small/large) I play unreal tournament while running prime.
All my cpu overclocks are with 285 sli enabled/overclocked, and ram timing/freq overclocked.
Again this isn't my full stress menu, but it seems to work for my setup.

Sheyster · 10-03-09

IMO if you get through 5 passes you're good to go, if it's a home/non-mission critical system. That test is seriously intense. I can't believe how high my temps got with it vs. Prime95.

drBlahMan · 10-04-09

30 passes of ITB
30 passes of LinX (Standard)
3hrs of OCCT Linpack (Max)
(2) 6hr Prime95 tests (small ftt & blend)

That's enough assurance for me regarding to stability

sdla4ever · 10-04-09

i run mine through 25 max test before calling it safe cuz ive had unstable errors as late as 19 before

lowkickqop · 10-17-09

I have the same question. I pass 100 intel burn tests. prime will run for about 6 hours. Then as soon as i reboot or start using my computer while not under stress it crashes shortly after.

Clockadile Dundee · 10-17-09

Quote:

Originally Posted by CL3P20

So, for most here..IBT is working just fine for CPU stability with ~10passes...and unless full mem usage is selected..sounds like Large FTT's with Prime95 seems to still be necessary for NB stability testing.

..Anyone passing 10x loops with full mem useage on IBT and still failing Large FTT's? Just curious...

** I'll post a screen shot after voltage tweaking a bit...would love to see some others to compare FSB and GB/sec throughputs..

In the thread in my sig, the "Striving For Ultimate Stability" club thread, we do varying amounts of loops, depending on installed RAM amount. For example, 300 loops for 2 GiB, 170 loops for 6 GiB, and 140 loops for 12 GiB; then we do 22hr custom 8k-4096k P95, a 4hr custom 4096k-4096k, and a SuperPi 32M run.

		Overclock.net - Overclocking.net > Intel > Intel CPUs
Intel Burn Test...how much is 'enough'..?

10-03-09	#35 (permalink)
M1 Abrams Overclocker Join Date: Apr 2009 Posts: 642 Rep: 59 Unique Rep: 51 Trader Rating: 0	""sounds like Large FTT's with Prime95 seems to still be necessary for NB stability testing."" This isn't my full course stress menu, just a part of it. For NB/Video stability I run 2 hours prime blend/small/large. In the last hour of each(blend/small/large) I play unreal tournament while running prime. All my cpu overclocks are with 285 sli enabled/overclocked, and ram timing/freq overclocked. Again this isn't my full stress menu, but it seems to work for my setup.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)