IBT confusion

willhub · 10 Jun 2010 at 14:06

Hi.

I've being using IBT to stress the system, and so far it's stable over high stress level, 20 runs.

Problem is, or maybe not a problem, sometimes, the same test runs around 44GFlops, but like now, I'm running it after that, each run is only doing around 25GFlops.

Why is this run going slower? When I look at temps it's clearly not stressing it as hard, despite being on the same test, this time, 63-69 Vs 70-81 degrees.

Any ideas?

Thanks
Will.

Liampope · 10 Jun 2010 at 16:51

That sort of thing happened to me once when I accidentally left prime running (by closing it with the 'x' button rather than stopping the test properly) then starting IBT. You done the same? Or just running a bunch of other stuff at the same time as IBT?

alex24 · 10 Jun 2010 at 16:53

I had this once too. I never found the reason why, but check you're running the same stress level. When I restarted the test it worked fine again, perhaps it was because I started it immediately after a restart whilst windows was still finishing loading.

willhub · 10 Jun 2010 at 23:37

I had to exit it and start it again, still odd tho.

Don't like IBT, runs 20 runs, then another 20, then another 20, all stable, run Prime95 blend and the PC crashed :\

Load of crap when they say 20runs, which is like 30 mins is a better test than small FTT's and Blend over the course of 40 hours in Prime95.

wolvers · 11 Jun 2010 at 08:12

I've had the same with IBT being stable and Prime not and I've come to the conclusion that IBT doesn't really stress the RAM. Prime blend I find is a better overall stress test for Ram/Board/CPU overclock.

HeX · 11 Jun 2010 at 08:23

I find IBT to be pretty buggy, try Linx instead, same kind of thing, but i've never had a problem with it like I have IBT.

Sayso · 11 Jun 2010 at 14:23

IBT stresses the CPU mainly thats why it gets so hot, where prime blend tests everything (CPU, RAM and chipset) if its 8 hrs prime stable you are 99% ok, IBT is a quick and dirty test to test CPU stability nothing else.

Sem · 11 Jun 2010 at 17:52

Sayso said:
IBT stresses the CPU mainly thats why it gets so hot, where prime blend tests everything (CPU, RAM and chipset) if its 8 hrs prime stable you are 99% ok, IBT is a quick and dirty test to test CPU stability nothing else.

thats what i do

if im testing an OC i run Linx for around 5-10 runs which takes 15-30mins

if all is ok will lower vcore till it fails

once you find out the rough stable vcore then you run prime blend overnight

Rroff · 11 Jun 2010 at 17:57

willhub said:
Load of crap when they say 20runs, which is like 30 mins is a better test than small FTT's and Blend over the course of 40 hours in Prime95.

TBH 40 hours of Prim95 doesn't really prove/disprove stability an ostensibly 100% stable setup can freak crash out under those kinda conditions. I usually start with 20 passes of IBT to give me an idea of where its stable and finish up with 2 hours of OCCT. I'm yet to find a setup that passes that, thats failed any other stability test.

Sayso · 11 Jun 2010 at 18:22

Rroff said:
TBH 40 hours of Prim95 doesn't really prove/disprove stability an ostensibly 100% stable setup can freak crash out under those kinda conditions. I usually start with 20 passes of IBT to give me an idea of where its stable and finish up with 2 hours of OCCT. I'm yet to find a setup that passes that, thats failed any other stability test.

I use OCCT sometimes depends on how much turnaround time i have on a system.

willhub · 11 Jun 2010 at 18:30

Rroff said:
TBH 40 hours of Prim95 doesn't really prove/disprove stability an ostensibly 100% stable setup can freak crash out under those kinda conditions. I usually start with 20 passes of IBT to give me an idea of where its stable and finish up with 2 hours of OCCT. I'm yet to find a setup that passes that, thats failed any other stability test.

2 hours of OCCT on what settings?