E8400/8500 degradation myth possibly busted?

[zlojack]

Yeah, but running hotter than it used to could be attributed to the change in seasons, no?

[railmeat]

How about a prime result? Super PI doesn't proove any stability sorry to say.

agree
ya 10 seconds is not stable anything...fire up a game and it will fall on its face fast thats why im not into benchmarking etc.get a solid clock(prime 8hrs etc)and have some kind of piece of mind vs electrifying for it for 10 sec runs,means nothing except it survived another run and u didnt cook it that time around.guess some ppl could care less about the whole rma process..?

[railmeat]

Same thing I was thinking. Real games dont use 4 cores (except UT).

cod4 does fully

[javadshahvand]

cod4 does fully

[003]

UPDATE/CONCLUSION

Well guys, I can confirm now that my CPU's did not actually degrade.

It is apparent now that the 45nm CPU's have a burn-in period of about
1 - 2 weeks, after this time the CPU's will need a small bump in vcore
(.024mv - .050mv) to regain initial prime stability.

My CPU's are rock solid stable now, and are not showing any other signs
of further degradation, I even had the vcore on one up close to 1.5v for
some suicide runs, and still did not hurt it.

Any more updates here, or is this pretty much the new agreed upon consensus?

[adamsleath]

i thought degradation was supposed to take a long time?... ie electromigration or whatever it is..?

how many have actually killed their chips with overvolts?

Yep, It does seem very apparent. The second e8400 I bought I kept it @
1.32v and It eventually needed more volts, so it's obvious voltage was not
causing this. The good news here is once it completes burn-in it slips
no further.

this sounds familiar to me
i've experienced something like this 'burn-in' (possibly) with the e4300 'L2' 65nm...i ran it for 3 days at 3.4GHz before it started to bsod @ very reasonable volts also..1.4vish
back down to 3.1-3.2 now for the past year or so.

[CrazyNutz]

Any more updates here, or is this pretty much the new agreed upon consensus?

I would say so. It's still holding up for me.

[CrazyNutz]

i thought degradation was supposed to take a long time?... ie electromigration or whatever it is..?

how many have actually killed their chips with overvolts?



this sounds familiar to me
i've experienced something like this 'burn-in' (possibly) with the e4300 'L2' 65nm...i ran it for 3 days at 3.4GHz before it started to bsod @ very reasonable volts also..1.4vish
back down to 3.1-3.2 now for the past year or so.

I've seen peeps volting these things past 2.00v on extreme cooling for
benching, and I have asked them if it they have noticed any damaged
to the CPU, and the general responce is No.

I would say that heat is more a factor in damaging CPU's than voltage.
Keep it cool, so you dont damage your Jewel...

On a side note, I have a 4300, and now that you brought that up, I do
remember it too needed either a bump in vcore, or a drop in MHZ after a while.
I have been running that one for well over a year, still going
strong!

[ScottFern]

Just wanted to let you guys know my e8400 OC has degraded in the last 2 days. Random reboots and tonight I finally ran an Orthos Smal FFTs test and as I suspected the machine rebooted. So I took my OC down from 3.9 to 3.5 and took my voltage down from 1.35v to 1.16v. Sucks!

Any other CPUs that OC better than these damn e8400?!

[icecpu]

Just wanted to let you guys know my e8400 OC has degraded in the last 2 days. Random reboots and tonight I finally ran an Orthos Smal FFTs test and as I suspected the machine rebooted. So I took my OC down from 3.9 to 3.5 and took my voltage down from 1.35v to 1.16v. Sucks!

Any other CPUs that OC better than these damn e8400?!

Whenever it degraded, increase the voltage.
That's what I've done so far, when I got it first day, stable at 4.09Ghz for 1.35v. Now 1.406v for the same speed 4.09Ghz.

Until it blow the heck out of it

[mrcape]

FYI I had that happen in the first month degrading slightly down .05v but then it settled in. I test every week or two and it's still just fine at 4.25ghz for hours with only 1.32v.

Today -

[CrazyNutz]

Just wanted to let you guys know my e8400 OC has degraded in the last 2 days. Random reboots and tonight I finally ran an Orthos Smal FFTs test and as I suspected the machine rebooted. So I took my OC down from 3.9 to 3.5 and took my voltage down from 1.35v to 1.16v. Sucks!

Any other CPUs that OC better than these damn e8400?!

It most likely just completed the burn-in period, and now it needs a small bump, +.025 to +.050 mv.

We were all freaking out about this, and now we know it's just need a little
more vcore after a inital period of time. After that it should stay rock solid.

[Fuji]

When you increase the voltage, you increase the current as well (ohm's law).

So you'd be putting even more current through the processor interconnects.

So given that the metal is already really thin don't you risk blowing out some of the interconnects?

Can't you also risk blowing your gate oxide? A good SiO2 gate will breakdown at 5-10 MV/cm or 5-10 V across a 10-nm oxide. (Intro do microelectronic fabrication). They say that the equivalent thickness of the gate dielectric to SiO2 is 1 nm. That seems to be pushing the limit. How does that work then - you'd be having .5-1 V across a 1-nm oxide. Or is the voltage supplied to the gate no where near the vcore range.

[003]

Can't you also risk blowing your gate oxide? A good SiO2 gate will breakdown at 5-10 MV/cm or 5-10 V across a 10-nm oxide. (Intro do microelectronic fabrication). They say that the equivalent thickness of the gate dielectric to SiO2 is 1 nm. That seems to be pushing the limit. How does that work then - you'd be having .5-1 V across a 1-nm oxide. Or is the voltage supplied to the gate no where near the vcore range.

Not sure how it works but my Q9450 at 1.475v (llc disabled) has shown absolutely no degradation -- and I have tested it over the last week or so with linpack 64-bit, 30,000 problem size. Passes with flying colors. What ever "burn in phase" exists, mine has passed it already.

[icecpu]

My 8400 keep degrading , never stop
now at 4.09Ghz 1.406v no longer stable, I'm getting rid of this bull$hit cpu.
What should I get without degrading ?? any recomendation ?

[CrazyNutz]

My 8400 keep degrading , never stop
now at 4.09Ghz 1.406v no longer stable, I'm getting rid of this bull$hit cpu.
What should I get without degrading ?? any recomendation ?

What was your initial stable votage @ 4.09ghz?

Also I'm curious, what multi & FSB did you use to get 4.09ghz?

[icecpu]

What was your initial stable votage @ 4.09ghz?

Also I'm curious, what multi & FSB did you use to get 4.09ghz?

First week 6hrs prime stable @ 4.09Ghz 1.35v
Now it needs 1.406v and maybe more.
multi 9, 455FSB

[CrazyNutz]

First week 6hrs prime stable @ 4.09Ghz 1.35v
Now it needs 1.406v and maybe more.
multi 9, 455FSB

0.056mv = 1.406 - 1.350, that's about on par. However for 24/7 use it would be best if you set your FSB @ 445 for 4005Mhz, and leave your vcore @ 1.35v. It will be better on your CPU, and your not gonna miss that extra 90mhz
in real apps or games.

[Darkatom]

My Q748A chip originally needed 1.344 vcore (load, cpuz) to be prime stable at 4.0 (1778 fsb). Now it requires 1.352 vcore at load to be stable at the same speed. I'll accept that but with my 790i board I idle at 1.387 vcore to acheive the 1.352. Is that safe for 24/7 use ?

[003]

My Q748A chip originally needed 1.344 vcore (load, cpuz) to be prime stable at 4.0 (1778 fsb). Now it requires 1.352 vcore at load to be stable at the same speed. I'll accept that but with my 790i board I idle at 1.387 vcore to acheive the 1.352. Is that safe for 24/7 use ?

My idle voltage with CPU-z is 1.440v with my Q9450 and I have had no ill effects. I would not worry about it.

[CrazyNutz]

My Q748A chip originally needed 1.344 vcore (load, cpuz) to be prime stable at 4.0 (1778 fsb). Now it requires 1.352 vcore at load to be stable at the same speed. I'll accept that but with my 790i board I idle at 1.387 vcore to acheive the 1.352. Is that safe for 24/7 use ?

I see, It's probably ok, but I understand It concerns you. Maybe you should
consider a vdroop mod, If there is one for the 790i. I just did the vdroop
pencil mod on my p5k-e and it keeps my idle, and load @ 1.352

[tmatrix]

sigh! Finally, I can post. I caught wind of this thread after I followed the Official Gigabyte X38-DQ6/X38T-DQ6 Info/BIOS thread almost religiously for the past couple of weeks since I built my first Core 2 duo machine - been over two months now. There was something wrong with the image verification tool at registration - I might still have my E8400 but oh well. Live and learn.

I suppose I have satisfied the criteria in my mind for serious degredation, so much so I ended up buying an E8500 when my E8400 had issues maintaining a 3.6GHz overclock. I almost thought it was the motherboard, but it was more practical to get a processor on short notice in my calculated opinion. Even more damning than the high Vcores in excess of 1.4875V or thereabout I pushed through it, was the high Vtt I used to run a 4.25GHz OC which lasted a little over a week before my OS started hanging and rebooting on me - 1.55V if I got the default of 1.20 right. And as if that wasn't bad enough, I ran into some "stupid" luck when my water cooler's coolant gave out on me running one of the high OCs - I can only imagine how hot it ran before it kept its cycle of restarting every so often when it got hot enough. If I recall correctly, it was a Q740A batch. After getting Realtemp which reported temps about 10 lower than Coretemp's readings, I picked up on how the chips sensors were kinda whacked with more than 7 degrees difference between the cores' readings. The readings weren't that bad before; I'm supposing it got worse especially after that failed coolant incident.

Forgive the story-like nature of my post. I'm at work at the moment and I'm relying on memory for enough of the information I'm reporting here. Before I got to work about an hour late, I managed to put the E8400 back into my system. I couldn't boot into WinXP at a 4GHz OC until I was pumping about 1.6V through it - 1.552 idle in CPU-Z. For the life of me, I couldn't maintain 4GHz with this chip as well as I did today; I'm of the suspicion that my stable settings I used for the other chip, the E8500 are holding true for the E8400 despite it's beaten-up state - I have VTT aka FSB termination voltage at normal now all the time. Sadly though, my Norton Antivirus scanner subsystem encountered an error, but everything else kept running. Since I want to watch this thing like a hawk to see how it behaves, I decided to shut it down before I stormed out to work. I did manage to do an XS bench with Realtemp and it gave a score of 1323 in about 11s.

Let's recap. There was too much VTT, too much heat, and too much Vcore. I'm probably missing something. Oh well! I'm not one to throw blame at all, but those DTS readings were very misleading. Also, even though I understand Gigabyte's need for widespread support, I'm quite disappointed in them for not doing enough of their homework for the new Penryns before releasing this motherboard to the market with a normal VTT of 1.20V - E8x00 only needs 1.1V. I'll keep you updated on what I find.

[003]

VTT will damage 45nm chips in excess. That is the most likely culprit of your problems.

[prime81]

I've had mine at 4ghz 500 x 8 for 2 months now, no stability issues. Constant vcore of 1.288 load/idle in windows, actual in BIOS is 1.30625

[tmatrix]

VTT will damage 45nm chips in excess. That is the most likely culprit of your problems.

Of all the things I mentioned, I firmly believe VTT did me in with that chip in particular.

Prime81, did you do a Vdroop mod or something to have your Vcore constant when both idle and under load in windows? It's very odd to see such especially with all the Vdrop values being reported - average of about 0.03V from supply voltage. You may have a gem of a motherboard there.