E8500 become unstable over time???

I had this E8500 E0 CPU for about 7 months now and I was able to get it to 4208mhz @ 1.328v overnight orthos small FFT stable. Recently, I decided to run orthos small FFT again just to test out the CPU and then it crashed in 4 minutes . I thought it was degradation since my CPU only needed 1.296v to run 4.2ghz when I purchase it, then 1.312v 2 weeks later, finally 1.328v another 2 weeks later and it was stable ever since. So I bumped up my voltage to 1.344v and it crash 4 minutes in small FFT again WTF? I use Asus P5k-E motherboard BIOS 1202 and here are my settings below.

Multiplier = 9.5x
FSB = 443mhz
STRAP = 333mhz
DRAM = 1:1 887mhz 4-4-4-12
PCI-E = 100mhz

CPU Vcore = 1.35v in BIOS (1.328v in CPU-Z)
CPU PLL = 1.5v
CPU Termination VTT = 1.1v
DRAM = 2.2v
NB = 1.4v
SB = 1.05v
CPU over voltage = Auto
CPU GTL = auto
NB GTL = auto
Load line calibration = Enable
CPU Spectrum = disable
PCI-E Spectrum = disable

4gb ram 4x1gb OCZ Platinum
400 watt corsair VX400 power supply
evga 9800gt with thermaltake DuOrb
OS = Windows XP x64 SP2

I use thermaltake cl-p0024 full copper tower with a vantec stealth 92nm fan with a very thin layer of AS5. Idle at 31/38 and load at 59/60 orthos small fft. I also ran memtest86 for 2 hours and it's stable, and I also have a thermaltake spirit on the NB. Could this be cpu degradation??

This really seems to be a CPU degradation.

You can try with lower multi and see if its stable to test the MOBO/RAM.

It must be something else, I dont think its CPU degredation, mainly because I routinely feed my cpu between 1.4-1.53v. Thats what I need to run 4300-4700mhz. ive had this cpu for 6 months, and its on 24/7 except for maintenance

Is your ram rated for 4-4-4-x? Maybe try 5-5-5 and see what happens

Are you sure you initially had it stable? Once you get something stable, you generally dont test it unless something is going on w/ your computer (reboots, crashes, ect). Are you having those issues?

1.328v @ 4.2ghz is pretty good and sounds spot on for an E0. Maybe, just maybe, your single test wasnt good enough for stability and maybe you need to test a bit further?

Also, Id also like to know what happens when you raise the ram timings. Also, try giving it some VTT, it may help without raising the CPU voltage.

I tried 5-5-5-15 and still no luck

When I set my vcore to 1.328v couple months ago, I can run 7-8 hours over night orthos small FFT, large FFT and blend without any trouble. I just leave it on when I go to bed and tomorrow morning wake up still running. I did not experience any blue screen or crash, just I want to re-test my system for stability every month or so. I never raise the vcore more than 1.3625v and only tried 1.2v VTT for couple hours. Later I found out the max VTT for my CPU is 1.15v, so I immediately lower that to 1.1v and re-test my system for stability, passed 7-8 hours small/large/blend orthos without any issues. I can run large FFT for couple hours without any error, just can't run small FFT right now ... I can run games, application all day and won't give me any crash. So should I be worrying about orthos small FFT???

Hes speaking of SMALL FFT ... over night 7 months ago, and now 4 minutes crash ... Small is mostly CPU related.

Who said 1.15v is the max for VTT? VTT max from intel is 1.45v. Alot of people bring it up to 1.2-1.3v to get stability. I know for me, 1.26-1.28v worked great.

My VTT is like yours.

But i agree he can raise it ... its damn low !

Can bring the PLL up, too. Intel says 1.54 max.

I can set PLL from 1.5 to 1.8v and VTT 1.1 to 1.5v with 0.1v increment. I can't set it to 1.54v for PLL or 1.15v for VTT, so I have to set PLL at 1.5 and VTT at 1.1.

Today I pull out my motherboard and took off the heatsink. I clean off the old arctic silver 5 paste and re-apply fresh pea sized in the center, and reinstall my heatsink. Then I reset the CMOS and then use the old settings I had in BIOS with 1.35v cpu vcore. I fired up OCCT and pass 1 hour of CPU test, then tried orthos small FFT another 3 hours+ stable. I will stress it overnight today and see if the system is stable. The thing I forgot to mention is I switched from Windows xp 32 bit to xp x64 recently, maybe 64 bit OS stress the CPU more and lead to orthos small fft crash??

Update: I went to fry's and grab a new antec three hundred case and I was shocked with the result. I carefully hide all the extra cables at the back of the case to ensure expectional air flow and then now I can run my E8500 at 1.304v still at 4.2ghz. I'm going to try orthos small FFT overnight today. Thanks for all your help.

That's pretty random to get a new case when fixing a CPU problem.

Glad it worked out for you!

Final update: 5 hours+ stable on orthos small FFT now at 1.312v 62/60C when I got back from home. 6C difference with antec three hundred compraed to my old thermaltake mambo case, pretty shocking..

A 64bit OS does indeed stress a CPU differently, often a 32bit OC that is right on the knife edge of stability will be distinctly unstable in a 64bit OS.

Yeah, I do notice orthos small fft run hotter in windows xp x64 compared to 32 bit edition. That's why I grab a new case from frys to compensate the temperature difference.

http://icrontic.com/articles/pc_airflow_heat_cooling_guide