Prime95 Errors at Stock

doktorstick · Jan 4, 2005

Howdy.

I think I know the answer to this, but before I start the RMA process, I wanted to get feedback from the community. I debated on whether or not to post this to the AMD CPU forum... but hopefully this is the correct location.

I recently purchased AMD64/939 3000+ Retail, MSI Neo2 Platinum, OCZ Powerstream 420W, OCZ PC3200 EL 1G (2x512), and XFX 6600GT AGP 8X as an upgrade to my computer. Before starting the overclocking, I wanted to check that the system was stable at stock installation.

- Windows XP SP2 with MSI drivers and BIOS updated according to their Live Update! tool;
- Prime95 has generated errors after 6m and 1h 20m when running the torture test;
- memtest86 has completed 14 successful loops over 4h 50m; and
- BIOS temperature reads 40C idle and MBM5 using the MSI Neo Platinum (not Neo2) profile reads 45C when under Prime95 load.

My guess is that the memory is rock solid, but the processor is flaky and I need to RMA it. I suppose it could be the motherboard or PSU, but I don't have much experience in debugging these types of problems (i.e., stock configuration problems).

Any advise? Thanks.

Cheers,
/ds

doktorstick · Jan 4, 2005

Well, now. I grabbed a copy of mprime on a Linux boot floppy and it has been running fine for some time now (~2 hours). It basically does what Prime95 does but without the heft of Windows XP running under the covers.

Perhaps there is something wrong with the Neo2 drivers? I did install that filesystem driver...

/ds

Charloz24 · Jan 4, 2005

Have you tried other stressing program like superPi 16M test...3dmark03 loop test....

I think I read that winchester and prime95 does not live well together.....

doktorstick · Jan 5, 2005

It's a thought, but it wasn't just Prime95 that died. World of Warcraft and lsass.exe both died as well.

I tweaked around some settings in the BIOS and Prime95 has run for about an hour without incident... but it had made it to 1.5h before... so I wait with baited breath.

I've submitted the RMA and have 14 days to get it shipped though I still have a fair amount of testing I need to do if Prime95 doesn't fall over.

/ds

rseven · Jan 6, 2005

I think I read that winchester and prime95 does not live well together.....

That's the first time I've heard this and I have been spending considerable time reading forums. Could you possibly remember where you read it? I'd like to check it out.
Thanks

Hoot · Jan 6, 2005

The current revision of Prime95 gets along fine with my Winchester As Long As I'm not overclocking beyond my particular hardware's capability, but then that's why I test with Prime95. Nothing makes my crazier than reading someones claim to their overclock that's "really stable" and then says something like"I still need to get it to work with Prime95 though...". With the exception of overclocking an individual graphiics card too far, in all the systems I've worked up, I've never had one that passed Prime95 Torture Test using the "In-place Large FFTs" option, through one complete pass, however long that took, that failed any other program I threw at it. I have had situations where I pushed a graphics card too far and though the system as a whole passed Prime95, the graphics card would flounder running 3D Mark. Sure, there are a plethora of system stability tests, but it seems the OC community most often uses Prime95 as a common reference point for establishing a system wide stability standard. I have had setups that passed Prime95 Torture Test "Blended" that failed some other programs, that's why I like the "In-Place Large FFTs" option.

Hoot

doktorstick · Jan 6, 2005

Problem solved?

Before I get into the "solution", let me say that mprime, the Linux Prime95, did eventually generate an error after 6h. So, it wasn't WinXP and any of the drivers.

I happened to stumble across a post on the AMD forums where someone had a problem with AMD 64 3000+/939, MSI Neo2 Platinum, and OCZ PC3200 1G (2x512) memory. He swapped out the OCZ memory and it was stable. I don't know if he has the same EL OCZ memory that I have, but regardless, I swapped out the memory with my Corsair XMS PC3200 1G (2x512) and then ran Prime95 on both computers (my other computer is a P4 2.8 system). Prime95 ran for 8h on both machines without incident.

The short of it is there is probably a compatibility issue between the MSI Neo2 Platinum and OCZ memory.

I have yet to re-overclock my P4 2.8 system with the new memory. I've lost about 50% performance swapping the memory because I reset it to stock. Hopefully I can get my P4 system running stable with this new memory. Then I can start work on overclocking my AMD system.

Cheers,
/ds

P.S. I agree with the previous poster that says "It is stable, except for Prime95..." If that's the case, then there is a problem.

cmitchellfly · Jan 6, 2005

doktorstick said:
Problem solved?

Before I get into the "solution", let me say that mprime, the Linux Prime95, did eventually generate an error after 6h. So, it wasn't WinXP and any of the drivers.

I happened to stumble across a post on the AMD forums where someone had a problem with AMD 64 3000+/939, MSI Neo2 Platinum, and OCZ PC3200 1G (2x512) memory. He swapped out the OCZ memory and it was stable. I don't know if he has the same EL OCZ memory that I have, but regardless, I swapped out the memory with my Corsair XMS PC3200 1G (2x512) and then ran Prime95 on both computers (my other computer is a P4 2.8 system). Prime95 ran for 8h on both machines without incident.

The short of it is there is probably a compatibility issue between the MSI Neo2 Platinum and OCZ memory.

I have yet to re-overclock my P4 2.8 system with the new memory. I've lost about 50% performance swapping the memory because I reset it to stock. Hopefully I can get my P4 system running stable with this new memory. Then I can start work on overclocking my AMD system.

Cheers,
/ds

P.S. I agree with the previous poster that says "It is stable, except for Prime95..." If that's the case, then there is a problem.

No Neo2/OCZ incompatability in my my rig. And, in my view, 8 hrs. of Prime is not necisarily stable. Ive had prime fail at 19 hrs. Obviously to each his own, but if I cant run Prime for AT LEAST 24 - 48 hrs it aint "stable"

Charloz24 · Jan 6, 2005

rseven said:
That's the first time I've heard this and I have been spending considerable time reading forums. Could you possibly remember where you read it? I'd like to check it out.
Thanks

Then take some time reading on this forum:

http://www.xtremesystems.org/forums/showthread.php?t=46885&highlight=0442

rseven · Jan 6, 2005

Charloz24 said:
Then take some time reading on this forum:

http://www.xtremesystems.org/forums/showthread.php?t=46885&highlight=0442

Thanks, I'll check it out.

rseven · Jan 6, 2005

doktorstick said:
Problem solved?

The short of it is there is probably a compatibility issue between the MSI Neo2 Platinum and OCZ memory.

I think you've nailed it. The compatibility issue may be something particular to your system. Even systems with the same components still have their differences. As for Prime, I think 8 hours does it. The makers of the program themselves say to run it between 6 and 24 hours. Your system is more stable than most personalities, my own included

Good luck.

Prime95 Errors at Stock

doktorstick

New Member

doktorstick

New Member

Charloz24

Member

doktorstick

New Member

rseven

Member

Hoot

Inactive Moderator

doktorstick

New Member

cmitchellfly

Member

Charloz24

Member

rseven

Member

rseven

Member

Similar threads