OC stability issue with E2140 @ 3.2GHz?

[ExcaliburMM]

Hey guys,
I'm running an E2140 @ 3.2GHz (400x8) w/A7FP w/MX-2 compound and getting idle temps of around 30C and load temps in between 50-55C. CPU-Z reads the voltage at 1.328v while under load. RAM is G.Skill 2x1GB DDR2 800 running 801MHz @ 6-6-6-18, 2.0v. Case is a CM RC690 with quite a few fans: 2x stock 120mms, 3x Yate Loon HS 120mms, and a YL 140mm HS on the bottom, so everything seems to be staying cool enough. I think my load temps are ok all things considered, I'll be replacing it in the next 6-8 months anyway. I've had it at this clock for around six months.

Sometimes when playing WC3 I've been getting restarts, and there's no error log or blue screen, the screen blacks out, and the PC restarts. I'm wondering if this is due to the OC or not since ORTHOS has it stable with a blend CPU and RAM test of about 10 minutes, I haven't really had the time to watch it for too long.

Also wondering if it might be the soundcards (Audigy 2 ZS) instead since the sound goes all funky when it crashes but I figured that could be due to the crash instead of being the cause of it. I will note I haven't had a crash since switching out my 8600GT for an HD4870 1GB DK, but that's probably because I've had my clock down to 3.0GHz when I made the switch, was trying to make sure my PSU could handle the new card and was putting less volts into the CPU just in case. However it seems everything's fine. PSU is a SeaSonic 500w.

Any ideas or additional info appreciated.


Wow, just as I was about to hit post, ORTHOS error'd. FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 9 minutes 11 seconds - 1 errors, 0 warnings.
Execution halted.

Edit: Found stress.txt. Whoops.

Edit2:
Well I guess the problem is I want 3.2GHz but I'm not sure what to do about the errors. The stock RAM timings were 5-5-5-15 and I've loosed them up and given it 2.0v. I would up the CPU voltage but I don't want to if I don't absolutely have to. Shortens the life and I cannot replace this thing for a few months.

Ideas?

 

[Bl0cks]

Have you tried memtest? From reading your problem I was thinking it might be your video card. My 7600gt had problems where I would be playing BF2 and the computer would eventually freeze with the sound looping, and the computer would restart itself. I did not get a BSOD and it wasn't driver related.

 

[VirtualLarry]

My E2140s are perfectly stable at 3.2Ghz (8x400), on GA-P35-DS3R v1.0, BIOS F4, at 1.36v (CPU-Z load), 1.425v (BIOS) (actually, one step down from that).

I think you just need a touch more vcore to stabilize.

Your temps are good though, mine get up to 80C under load, with a CoolerMaster HyperTX2.

 

[demiurge3141]

More vcore will do. It will not die in a few months, these 65nm guys are tough and can take a beating.

 

[ExcaliburMM]

Thanks guys, I'll be giving it a shot tonight and will post back with results. I'm also thinking of trying a little more NB voltage if needed because I have a 120mm blowing right on it from the door.

Edit:
Upped vcore and NB voltage. Results:

Had it stable for 20 minutes of testing @ 1.38v. I can't say I like that I had to up the vcore that much, nor can I say I like the temps it produced, and I can't even say I feel I tested long enough but it seems I never have time on my hands these days. Load topped out at 59-63C after 20 minutes, but was hovering around 55C before then for quite awhile. I have yet to see anything stress both cores to 100% for longer than a few minutes other than a stress test so I guess I can live with it. Idle temps are around 30C which I guess isn't so bad.

Will be doing more tests when I have the time.

 

[MustangSVT]

just run it a tad slower, 3 or 3.1 which it should be completely stable with.

you wont notice any difference, it will consume less power, and output less heat.

 

[soccerballtux]

Originally posted by: ExcaliburMM
Thanks guys, I'll be giving it a shot tonight and will post back with results. I'm also thinking of trying a little more NB voltage if needed because I have a 120mm blowing right on it from the door.

Edit:
Upped vcore and NB voltage. Results:

Had it stable for 20 minutes of testing @ 1.38v. I can't say I like that I had to up the vcore that much, nor can I say I like the temps it produced, and I can't even say I feel I tested long enough but it seems I never have time on my hands these days. Load topped out at 59-63C after 20 minutes, but was hovering around 55C before then for quite awhile. I have yet to see anything stress both cores to 100% for longer than a few minutes other than a stress test so I guess I can live with it. Idle temps are around 30C which I guess isn't so bad.

Will be doing more tests when I have the time.

Dude 1.4v is fine relax. I run mine at 1.5v in CPU-z for my 3.4Ghz, have had it for over a year now. You DO however need to worry about the ram at 2.0v and you probably don't need to increase your NB voltage. Turn the ram back to 1.8v or whatever the RAM says it should run at.

 

[Kraeoss]

hrmm running at 1.55 for 3.0 is too much lol... vdroop on my board is like .05v lol 1.55 bios and cpu-z/occt reads it as 1.5 is that too much ? for so little increase in speed ? currently @ 2.7 @ 1.35v cpu-z 1.425v bios

 

[ExcaliburMM]

Bumping this up. 3.2GHz stable but if I want to use more than 400 FSB I can't get a proper boot. Comp will start but no display / mouse light stays off. Any ideas? FSB wall?
Mobo: DFI LANPARTY DK P35-T2RS

Tried more vcore, tried more nb volt, tried more RAM volt, no dice.

However my DDR2 800 is running at DDR2 801 already because there's no lower divider than 333/667. The stock timings are 5-5-5-15 but Ive got it down to 6-6-6-18 @ 2.0v. Tried higher on the volts for it too, no dice. Perhaps the memory is limited here?

 

[VirtualLarry]

More than 400Mhz FSB is stretching it, especially on an E2140. I got mine to 410 FSB during testing, but the CPU wasn't stable there.

 

[soccerballtux]

Originally posted by: Kraeoss
hrmm running at 1.55 for 3.0 is too much lol... vdroop on my board is like .05v lol 1.55 bios and cpu-z/occt reads it as 1.5 is that too much ? for so little increase in speed ? currently @ 2.7 @ 1.35v cpu-z 1.425v bios

Edit: Yeah that's too high for 3.0 likely the stability issue is somewhere else.