Winchester 90nm and Prime95

[andyOCZ]

Hey guys,

I am seeing some odd things on my 3500+ 90nm Winchester and Gigabyte K8NS Ultra-939. It seems that at any decent overclock above 250mhz Prime95 will always fail. CPU Burn In and other stress tests seem fine.

Have any of you had luck with 90nm and Prime95?

Let me know what your results are. Thanks

Evidence of Prime95 failing at 200fsb for those of you contemplating RMA's on memory due to failure: http://www.xtremesystems.org/forums/...340#post624340

[Ol!ver]

Couldn't it just be instability? We all know what prime is like for finding problems.

Could be the RAM, Board, CPU or mem controller that isnt happy.

Oli

[DevilsRejection]

I'm thinking it is the memory controller itself on the chip, which is why I'm going to spend the little extra and get a 3500+ and not a 3200+ so I can clock it at 2-2-2-10 at 240x11 = 2640

i hope i can make it happen!

[bachus_anonym]

i can run D2OL Distributing Computing Project with no problem at 265x10=2650MHz @ DDR530 6-2-2 cas2 3.6v.... SuperPI 16M and 32M @ 270x10 but Prime95 fails instantly at 2700Mhz
at 2650Mhz it runs but i stopped it after 20mins....
i'll fire it up again and see what happens

[Goli]

i can run D2OL Distributing Computing Project with no problem at 265x10=2650MHz @ DDR530 6-2-2 cas2 3.6v.... SuperPI 16M and 32M @ 270x10 but Prime95 fails instantly at 2700Mhz
at 2650Mhz it runs but i stopped it after 20mins....
i'll fire it up again and see what happens

I have the same problem, any high 200Mhz and i have prime95 fail, but i can run 3Dmark 2001 48Hours without problem, then 3Dmark2001 and Econding at the same time in 24Hours and didn´t freeze or reset my computer, i don´t know what is the problem.

[esdee]

andyOCZ kinda same problem here too with the neo2
prime fails within a second even if i can supePI 32mb at that MHz
The solution for me is not to run prime at all,
it makes me go nuts!

[MaxxxRacer]

my prime seems fine as long as im at 2500mhz or below.. one mhz above 2500 and nothing is stable.. but whats odd is that i can get to 2500mhz with decent stability at stock voltages.. but it takes 1.6volts up from 1.4 to actually get it totaly stable at 2500mhz... ohwel... looks like i just got a bum chip.

[WiCKeD]

If it won't run one program it might have problems with others. Therefore YOUR SYSTEM is not stable.

[DevilsRejection]

If it won't run one program it might have problems with others. Therefore it's not stable.

why doesn't anyone just email the author of the program?

[joe2004]

Prime will fail when there is some calculation that has not been done correctly. That can be seen as a small pixel glitch in say 3dMark but Prime will not just say small glitch, it will fail. That shows you that your system isn't doing it right and no matter what you can run there it is not stable. Scientific calculations are not done correctly, as simple as that.

Prime95 is the ultimate measure of the stability of the system.

[DevilsRejection]

agreed with above post

[MaxxxRacer]

ironic though. my system is 9 hours prime stable, but it crashed playing day of defeat. hopefully its just a random windows crash and not a cpu crash.

[WiCKeD]

ironic though. my system is 9 hours prime stable, but it crashed playing day of defeat. hopefully its just a random windows crash and not a cpu crash.

I've had IE and other programs crashing all around me while Prime95 was running. It's not the be-all-end-all of stability tests, but it's probably the most taxing. Anyway, I'm sure if I had let it run long enough, it would have had problems.

[Ol!ver]

Prime will fail when there is some calculation that has not been done correctly. That can be seen as a small pixel glitch in say 3dMark but Prime will not just say small glitch, it will fail. That shows you that your system isn't doing it right and no matter what you can run there it is not stable. Scientific calculations are not done correctly, as simple as that.

Prime95 is the ultimate measure of the stability of the system.

I think what people are wondering is if something Prime is doing is causing the mem controllers to cause errors, when in fact nothing is wrong......

Oli

[Zanr Zij]

here U see...

[Goli]

I do some testing and i note that in the cell menu, when you are moving the FSB, the spread spectrum gone enable then you go to another FSB an spred spectrum goes disable, u have to go to FSB 200 then put disable and then pump the FSB, now i am not getting error in prime95.

[trans am]

Guys, i've had no problems running torture test/ blend at 2700 on a neo2/winnie 3200+ this is after 6hrs. But of course it will eventually fail Prime no matter what because it's a winnie.

[blinky]

If it won't run one program it might have problems with others. Therefore YOUR SYSTEM is not stable.

if it runs every program u use without crashing its stable



unles.... u have the need to calculate prime numbers all day long

[Boogotop]

can run prime at 9x320 with no prob here

[dnottis]

can run prime at 9x320 with no prob here

Specs, volts, screenshot of CPUZ?

[Mrki]

can run prime at 9x320 with no prob here

BIOS?

[OmegaSys]

I am having real odd problems with Prime95 as well. Sometimes I can only last 2 hours and then for some odd reason it will go 13, 18, or even 24 hours. There is just no rhyme or reason to it. I've been dealing with this for a month now and it surprises me every day.

-OmegaSys

[reject]

if it runs every program u use without crashing its stable



unles.... u have the need to calculate prime numbers all day long

exactly

[SkyGod]

yeah I hate prime 95 too, I've been pulling my hair out over it, I can do 250X10 on my winnie and play HL2 for hours on end but when I prime it it fails immediately!

[OmegaSys]

When people say that their computer is "Prime stable," what do they mean? I know I would think my computer was Prime stable if I only ran it for 6 or 12 hours but many times it fails at 13, 18, or 21 hours. It's making me think that not even 24 hours is good enough due to the random pattern of the failures.

-OmegaSys