Is this i7 Degrading?

[jguy]

Built a rig for someone in March,

Asus P6T SE
920 @ 1.2875V 3.8Ghz, Ultra 120
12GB Gskill RipJaw
Radeon 4870
Audigy 2 ZS
OCZ Vertex LE i think 50GB
2X WD 1TB hdds

BIOS settings were:
Qpi - 1.3375v
Cpu 1.2875v
IOH 1.3v
ICH 1.3
CPU PLL 1.84v
Memory 1.62v 1600MHz a 7,8,7,18
LLC off from day 1

I tested this with IBT 500 passes (i know, i know.. i tend to do 100 now max) and 48hrs p95 blend, both checked fine and there were no complaints. Games fine etc..

However, i had the system back for some routine maintainence and before i give her back i just ran some stability tests to ensure everythings ok, it now fails IBT within an hour, bluescreen and reboot, prime passes fine.

Finally managed to so far get it back to some stability, all voltages are the same apart from QPI, which i have now set to 1.35v. Surely these voltages could not have degraded the chip? Any reason this might have happened? Its running over an 2hrs so far with IBT so hopefully this is it.. Going from experience, the original Qpi voltage might have been a bit on the low side especially with 12GB RAM at 1600mhz? Strange it went fine before though.

[sin0822]

are the temps skyrocketing? or different from when you built it? It should be just fine. maybe try re-install?You are well within intels limits.

[jguy]

Temps are 72C on the hottest core with IBT, others are max 68C, i built it in March.. so it was a wee bit cooler but i also tested it for a few hours in dead summer where it was 28C indoors.. right now its 20 lol. It seems to be plugging away at IBT for the last 6hrs now at 1.35v qpi, the old voltage of 1.3375 fails within minutes with a blue screen, only set to run 50 passes, 1 pass takes about 660 seconds with 12GB hehe.

[fordf250]

If I set my 1600 ram to xmp it defaults to 1.35V for qpi at 1600 even for 133bclk,you could have been to low when you started.

[sin0822]

A lot of things can degrade, the most likely culprits are on the motherboard in this case because you haven't gone past any of intels max's. Capacitors can wear and so can other electrical components, requiring a bit more juice. Also take a look at the PSU, i wouldn't factor that out.

[jguy]

Well.. heres the odd thing.

BIOS was Feb 10.. upgraded to latest.

So far its been over 2 hours and 11 runs of IBT at

Qpi - 1.3v
Cpu 1.2875v
IOH 1.24v
ICH 1.3
CPU PLL 1.84v
Memory 1.62v 1600MHz a 7,8,7,18
LLC off

Forgot to mention psu is a Corsair HX620, i tested it with a TX650 as well with same results previously..

It'd reboot almost instantly with these settings previously.
Gonna see how this pans out lmao. Even if it dosen't pass all 100 tests its a start...

[sin0822]

yup, still doesn't look bad at all.

[mk-ultra]

I tested this with IBT 500 passes (i know, i know.. i tend to do 100 now max) and 48hrs p95 blend, both checked fine.

WHY would you ever need to run so much fake stressing?

[Loser777]

@OP, I've had similar issues when my Q6600 was flaking at 3.6GHz... and in that case it was my PSU that was slowly dying. You're not going to see it in everyday use, but PSUs can cause instability during stressful tests. I'd say power supplies are the most likely to degrade, followed by motherboard and then CPU.