View Full Version : Orthos = EPIC FAIL
I tried running 260x11 for 2860MHz on a core/9 memory divider... DDR633 pretty much. Orthos blend ran ok for 13 hours on priority 9. Yet I start up F@H and browse TPU... within minutes it locks up with an EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE sound coming from my speakers.
:banghead:
What tests do you suggest?
imperialreign
03-03-2008, 09:19 PM
F@H is the system stability test of mythological proportions. You can throw all the big league stability tests and benchmarks at a OCed system for weeks and might not have problems - but 15min of F@H could make your CPU ball up into the fetal position and start howling like a kitten accidentally closed into the pantry.
The processes require extreme accuracy on part of the CPU, and the slightest discrepancy could cause a SYS crash.
You have two choices - back down the OC, or don't run F@H.
What about testing the memory then? Or NB(FSB too high)?
Solaris17
03-03-2008, 09:30 PM
i agree F@H is legend....but if you dont want to use that as a stability test i use prime 95 and then i customize it high iterations using the max possible ram i dont go with the presets i test everything at once at the highest possible setting..
imperialreign
03-03-2008, 09:32 PM
you might see an improvement if you loosen the MEM timings, or bump MEM or NSB voltage a bit . . . but I don't really think that'll make much difference. It might be worth a shot, but just be aware it might not abate the problem, either.
BTW, how far OCed is the CPU/SYS?
Hawk1
03-03-2008, 09:34 PM
I test the memory with Memtest86+ overnight.
I test the CPU with Orthos Priority 9 small FFT's 8 hours +
I test complete system with Orthos Blend 24 hours minimum (I have had it fail after 18 and 19 hours so I always do at least 24 to be somewhat reassured).
Memory was DDR 633 from DDR 533 same timings 4-4-4-12-16-1T at 2v
cpu is at 2.86ghz 1.4v I know it's stable -_-
imperialreign
03-03-2008, 09:44 PM
what kinda CPU?
hell, what happened to your sys specs under your avatar? :confused:
if messing around with MEM timings and chipset voltages don't show any improvement in F@H stability, you might want to lower the CPU multi a bit (if you can), or drop the BUS speed back down. If, i.e., 2.86GHz isn't stable, try it again at 2.75GHz.
DaMulta
03-03-2008, 09:55 PM
No fucking fire?
How is this Epic?
Can you pass memtest?
I'm just gonna run it at DDR700 5-5-5-15 where I know its stable.
vBulletin® v3.7.0, Copyright ©2000-2008, Jelsoft Enterprises Ltd.