CPU Burn-in, in reverse?

[st0nedpenguin]

Righty, my 2500+ was running happily at 2.3GHz @ 1.8v for about three days, Prime95 and 3DMark stable.

Now for some reason, it is requiring 1.825v to run Prime95 for longer than 10-12 hours, and I haven't got a clue why.

I reseated my HSF last night, in the hope if sorting the problem out, thinking maybe I wasn't getting good contact or my AS3 application was just crap, and while I was doing this I dropped a few extra washers into the mounting screws to up the pressure on the CPU, this managed to drop my temps by about 2-3 degrees, even before the AS3 has had time to cure, but has done nothing for my voltage issue.

The only thing I have changed, except for reseating the HSF (which happened after the voltage problem arose anyway), is that I have dropped my RAM and chipset voltages back to 2.6/1.6 respectively, I don't think this could have caused the problem, but I could be wrong, it wouldn't be the first time. :D

Is this normal or have I buggered my chip in some way?

[TheWeaseL]

Good question. I'm guessing you're running the nf7-s 2.0, and in my experience with it, if you change one thing, and it shouldn't affect another thing in any way, shape, or form, it probably will. :)

Try putting your ram and chipset voltages back to where they were before, and give it another go.

[st0nedpenguin]

Yeah, it's worth a try I suppose, I love this board, but damn can it be awkward at times...

[Marci]

The best board always are... be glad it isn't as bad as the KX7-333 was... took stoopid amounts of vmods etc to get that board into full benching mode but once it was there it was unstoppable! With the NF7-S u can get away with fiddlin the bios... vmods are just for that extra-special somethin'

[Niven]

Same problem here...
CPU was 100% stable (F@H for 3 weeks) @ 200x11.
Now im getting BSOD's and program crashes all day long.
Temps seem to be higher as well, yet the AS3 has only been on for a couple of weeks.

[st0nedpenguin]

Well, I was going to try bumping the voltages back up again last night, but my rig was 12 hours into a 24 hour Prime95 run at 210x11, so I thought I may as well leave it to finish, and it did, it's running 3DMark now, so if it finishes that ok I'll try dropping the CPU voltage a notch and try my luck.

On a positive note, the CPU temp appears to have droped a few degrees from my HSF reseating, before the CPU temps were about 46c-48c and now they are down to 44c-46c after a day, and the ambient temps are actually warmer now than they were before.

[st0nedpenguin]

Hmmm, I had a shot at upping the RAM and chipset voltages, no luck, I really haven't got a clue what's going on with this chip.

:(

[DisposableHero]

computers are plain weird sometimes =\

[shrae]

This happened with my K7S5A rig right after I installed a new CPU and hs/fan. I switched from a Tbird1.4 with retail cooling to the XP2400+ with Vantec Thermoflow. My temps started low, around 32C load, but even as my ambient temps dropped as winter progressed my temps inexplicably increased about 4C. I tried reseating and reapplying AS3, but nothing seemed to help so I just gave up :/

Could it be a hardware monitor issue perhaps? It's hard to believe that it could be hardware related.

[st0nedpenguin]

The story gets odder.

After a few days of running at 1.825v, my machine crashed after 30 hours of 3DMark, despite being perfectly stable at 1.8v previously.

Obviously, I wasn't very impressed. :D

Having given the AS3 a bit of time to cure, I thought I would have another shot at dropping the CPU voltage back to 1.8, what the hell, nothing seems to be working at the moment.

Dropped CPU voltage to 1.8v, upped RAM voltage up to 2.9v, upped chipset voltage to 1.7v, fired up Prime95.

It has been running 13+ hours of Prime95, and was still running when I left for work this morning, have to see what it's doing when I get home.

I'm really beginning to get confused now.

[Marci]

my machine crashed after 30 hours of 3DMark

Why was it running 3DMark for 30 hrs? Could've crashed due to a momentary dip in 12v or 5v rail, GPU overheat... multiple reasons... not necessarily anything to do with CPU...

Think u may be taking stability testing a bit too seriously d00d.... do u ever game fer 30 continuous hrs solid?? (If u answer yes then u REALLY need to get laid d00d!)

My rig fails Prime every time at current speeds, but it 3DMark's & Aquamark's, Plays BF1942, does all me graphic / web design (Coreldraw / Photoshop / Dreamweaver / Actinic), does all me audio recording (CubaseVST etc), Folds... all perfectly. Which means, as far as is relevant to me, it's stable.

On your temp issues, you never mention what the difference in Ambient temp has been when the CPU has gone up or down... it has been fluctuating temp-wise over here recently...

[st0nedpenguin]

I was a little unclear earlier, the 3DMark crash was whilst burning in the AS3, leaving it to cure for a few days, I thought I had cranked the voltage up to 1.9v, to get the temps up, but I actually had it at 1.825v, I should really pay more attention in future.

Crashing after 30+ hours, yeah I can cope with, I'm not likely to manage a 30 hour gaming session, but hell, getting laid would be nice. :D

The temp drop was with the case temps at 28/29 degrees, it has been for a few days now, not counting the odd fluctuation, though the temps here have been all over the place as you mentioned.

I'll settle for 24 hours Prime95/3DMark, I just don't want a crash in the middle of a BF1942 match, can be rather frustrating. :D

[st0nedpenguin]

Well, a wee update for anyone whose interested, yes, you, the lone guy at the back.

It would appear that stability at 1.8v, low NB/RAM voltages has returned, 18 hours as I left for work, no clue what the problem was, but it seems to have gone now.

On a plus note, my temps dropped to 39c yesterday, on air, I'm loving this change in weather. :D