**Official e8400/e8500 Retail OC Thread

[cadaveca]

@ARIcIU and Cadaveca, what Vpll were you guys pushing?

1.557v cpuPLL. lowest setting available.




Not the problem at least for me. vFSB(VTT) @ 1.123v.

Readings by DMM, not bios, software, etc...

[Capt Proton]

When you folks give a voltage, are you talking about what you are setting in the BIOS, or what is being reported? I understand that some are saying something like - 1.25V in CPU=Z, or 1.175 on my Multi Meter. I've got those ones.

[maxwolfie]

I'd bet that most people are quoting CPU-z, as there is vdrop to account for in most situations.

Whether that's load or idle is a different question though. I personally will not go over 1.4v idle CPU-z

[Vinas]

Here's my e8400, seems week q748 is still OK despite is taking a few more volts.

Week: Q748
Batch: A219
Spec: SLAPL "C0"
Packaged: 1/29/08
CPU = 8x500 (4GHz)
vcore bios: 1.42
vcore CPUz: 1.392
vcore orthos: 1.38

http://valid.x86-secret.com/show_oc.php?id=316728

[laragirl83]

Now I am having again the same problem as before with P5B Deluxe BIOS 1226...
I set 8*500MHz for example, with proper voltages (cpu 1.352V, fsb termination 1.30V, NB 1.45V) and sometimes, like 3 out of 5, when I get to XP login screen, i type my password, hit enter... and it all slows down... after like 30 second only the background appears, no icons, no tray no nothing... tough mouse is still operating , and the keyboard as well, but still, the only thing i can do is reset....
Sometimes, it enters desktop all right, and I can run anything, prime95 small ffts, games, dvd, it doesnt matter, and it work like a charm, no errors, no freeze, nothing...
Now why is that...?
Like I said it happend to before with BIOS 1226, but guys here on XS adviced me to change back to 1219, cuz' its more stable, and better for OC...
Now same thing happen with this BIOS rev as well...
Any ideas?
Settings are the same, the ones worked well before, now doing the same old

[ixtapalapaquetl]

... and it all slows down... after like 30 second only the background appears, no icons, no tray no nothing... tough mouse is still operating , and the keyboard as well, but still, the only thing i can do is reset....
Sometimes, it enters desktop all right, and I can run anything, prime95 small ffts, games, dvd, it doesnt matter, and it work like a charm, no errors, no freeze, nothing...(

Sounds like Windows is corrupted to me. I would reinstall.

Incidentally, you should seriously consider getting a program like Acronis True Image. Then it would be a matter of restoring rather than reinstalling. I might also suggest using two OS partitions, one for stability testing/benching/screwing around, and another to be used only with known stable configurations for 24/7 use.

[Mumid]

Now I am having again the same problem as before with P5B Deluxe BIOS 1226...
I set 8*500MHz for example, with proper voltages (cpu 1.352V, fsb termination 1.30V, NB 1.45V) and sometimes, like 3 out of 5, when I get to XP login screen, i type my password, hit enter... and it all slows down... after like 30 second only the background appears, no icons, no tray no nothing... tough mouse is still operating , and the keyboard as well, but still, the only thing i can do is reset....
Sometimes, it enters desktop all right, and I can run anything, prime95 small ffts, games, dvd, it doesnt matter, and it work like a charm, no errors, no freeze, nothing...
Now why is that...?
Like I said it happend to before with BIOS 1226, but guys here on XS adviced me to change back to 1219, cuz' its more stable, and better for OC...
Now same thing happen with this BIOS rev as well...
Any ideas?
Settings are the same, the ones worked well before, now doing the same old

I had similar issues when my memory was overheating, i added active cooling to it and everything then worked ok.

[Vinas]

I was going to also mention try to lax up on your memory settings and/or vdimm. Maybe this is a good place to start if you haven't already...

[Mumid]

I had similar issues when my memory was overheating, i added active cooling to it and everything then worked ok.

Also, other things to try would be to increase northbridge voltage and\or increase your memory performance level (i.e loosen it, for example change it from 7 to 8)

[emoners]

it's funny these wolves...they may be power efficient on the proc but it makes you push your NB & RAM voltages

[rge]

mumid, what is your NB volts and FSB volts to run 500FSB at perf 7? Do you ever get random shutdowns running that tight?

[Mumid]

mumid, what is your NB volts and FSB volts to run 500FSB at perf 7? Do you ever get random shutdowns running that tight?

My volts are

Northbridge (MCH) +0.325 (Prime stable) or +0.35 (OCCT stable). Can run it at +0.3 for general usage but i leave it on +0.325.

Fsb (VTT) is set to +0.15.

Im not sure what the default voltages are for these on the Gigabyte boards and I havnt got a MM so im not sure what the actual voltages are.

Anyway, it runs rock solid stable, not one crash or reboot in Windows ever even after playing Crysis for hours on end. The only issues I had were I were getting fails in Prime and OCCT after a while but after adding active cooling to my Ballistix everything is fine again.

One thing to note is not to use the "performance enhance" option in the Gigabyte mobo bios but instead to set these timings manually instead.

Hope that helps

[HaTaX]

I only checked my VCore, VTT and VDimm with my DMM, but all of them were within a few hundredths of what the bios was reporting so I thought that was close enough. I do have a small difference in what I set vs what I see reported, normally an small increase from voltages I set are reported.

Also my vcore looks like it consistently is jumping from 1.27 to 1.28 in the hardware monitor, but with my DMM it reports 1.274 - 1.276. It's just twitching between that .005 mark where I bet the bios rounds up. Not really a big deal but it is irritating to see it jump around in logs when it's really pretty steady.

We should make note of what people are seeing reported by a program like hardware monitor by CPUID for their voltages, it would be more helpful I'd think to someone trying to mimic another's success.

[yokomo]

temps are quite high i think but considering i just got in within an hour so the AS5 hasnt set inyet im doing ok....3.789 on 1.2v bios as ive just got .
coretemp says 35/40 idle ! and 40 and 45 load whihc seems quite high tome but its lower than my e6400 so im guessing is just due to my bios being crap (965 gigabyte)
hmm loading here wiht orthos im getting 52 and 55 c !!!vid is aprently 1.1v exactly !

[rge]

My volts are

Northbridge (MCH) +0.325 (Prime stable) or +0.35 (OCCT stable). Can run it at +0.3 for general usage but i leave it on +0.325.

Fsb (VTT) is set to +0.15.

Im not sure what the default voltages are for these on the Gigabyte boards and I havnt got a MM so im not sure what the actual voltages are.

Anyway, it runs rock solid stable, not one crash or reboot in Windows ever even after playing Crysis for hours on end. The only issues I had were I were getting fails in Prime and OCCT after a while but after adding active cooling to my Ballistix everything is fine again.

One thing to note is not to use the "performance enhance" option in the Gigabyte mobo bios but instead to set these timings manually instead.

Hope that helps

Thanks

[mrcape]

Hey Hey! Chip #2 has slightly degraded, needs a little more vcore to be stable at old clocks and idles 2 or 3C hotter. This time with max vcore = 1.415v, no prime or anything but PI at that high v. It almost seems like running high FSB period is doing the damage. The temps never broke 56C and like I said vcore 1.415v. Other max v - VTT - 1.15v, PLL - 1.55v, NB - 1.45v, VDIMM - 2.3v.

This was all the result of a little PI and 3dmark benching on Saturday. The next day, I noticed temps a bit higher so I checked to see if it had degraded by going back to a previously prime stable setting in cmos reloaded. It needed a vcore bump from 1.22v to 1.25v to pass.

Oh well, it still passes prime at 1.25vcore set in bios, but I want my baby back! lol I have a record of before and after, and what I did along the way, but before I plug it all in as a reply here, is there a database specific to degradation anywhere? I'm hoping it doesn't continue it's downhill slide as long as I stay under 1.3vcore from here on out. Has anyone else had chip seem to degrade, backed off, and then it stayed put? Most of the experiences I read about say that it kept going down the tubes.

[Mav451]

Nope. We can make a separate thread, and someone (or me lol) can put up another database.

[PhilDoc]

Hey Hey! Chip #2 has slightly degraded, needs a little more vcore to be stable at old clocks and idles 2 or 3C hotter. This time with max vcore = 1.415v, no prime or anything but PI at that high v. It almost seems like running high FSB period is doing the damage. The temps never broke 56C and like I said vcore 1.415v. Other max v - VTT - 1.15v, PLL - 1.55v, NB - 1.45v, VDIMM - 2.3v.

This was all the result of a little PI and 3dmark benching on Saturday. The next day, I noticed temps a bit higher so I checked to see if it had degraded by going back to a previously prime stable setting in cmos reloaded. It needed a vcore bump from 1.22v to 1.25v to pass.

Oh well, it still passes prime at 1.25vcore set in bios, but I want my baby back! lol I have a record of before and after, and what I did along the way, but before I plug it all in as a reply here, is there a database specific to degradation anywhere? I'm hoping it doesn't continue it's downhill slide as long as I stay under 1.3vcore from here on out. Has anyone else had chip seem to degrade, backed off, and then it stayed put? Most of the experiences I read about say that it kept going down the tubes.

Maybe there's something I dont understand, but why would degradation make your temps go up? TRUEs are know for their mounting problems. Is it possible that you're not making as good of contact as before?

[mrcape]

Maybe there's something I dont understand, but why would degradation make your temps go up? TRUEs are know for their mounting problems. Is it possible that you're not making as good of contact as before?

No change, the sink is solid and the load temps are the same as before degradation. It's only idle temps tat went up. I had the same symptom on my first chip that degraded.

Q745 - Pre high FSB (545) it was idling around 35C @ 4.2ghz after that it idled at about 38c.


Q748 - Pre high volts (1.46v) it was idling around 35C @ 4.0ghz after that it idled at about 42c.

It almost seems like the sensor itself got a bit sizzled.

[PhilDoc]

Shoot, sorry to hear that. I was hoping that wasn't the case. Hopefully, it wont get any worse. That's a great chip you have there.

[mrcape]

I hope it stays put. But if not, it's okay, I can chalk it up to early adoption. I really do wish I kept it under 1.3v max though. lol It's just too tempting.

I think I'll be getting a yorkfield or two and maybe the c1 of wolfdale if it comes out soon. When I do, I'll put the Q745 on the workbench and jack it up.

Hopefully this info saves someone from damaging theirs. I hope we can identify a trend with the way they degrade soon. I'm really curious to see the FSB of other degraded chips and what was run, how long, voltage etc..

Wasn't there some recall regarding FSB scaling problems with the first batch of intel 45nm? Any details on that problem?

[PhilDoc]

I don't know of any recall, but irrc, they delayed the yorkfields because of instablility at high fsb on lower grade mb. That's all I remember about it. There should be a thread in the news section about the delayed yorkies and why. Don't know how old it was though.

[mrcape]

I thought there was some article stating that Lenovo bailed on a bunch of intel 45nm mobile chips as well.

Well, it's definitely hard to dig up any info in forums. A lot of it's buried in posts by people defending intel and saying it's not true yada yada. I can't tell you how much of that crap I have to skip over, enough already. It hilarious to read that stuff from armchair experts who don't even have the chips.

[.OCX]

There's no viable data. Whenever you track trends you gotta have good data. Good data goes in, Good data comes out. Without a uniform standard to test against, this is more baseless proof of degradation. Factor in the Air cooling folks with 1.4v for 4.0ghz...now your "Control Data" is worthless.

"Official E8X00 Overclocking Database"

Where art thou?

[PhilDoc]

Oh, I agree, way to many experts. Here's a couple of articles I found. Don't know if they'll help any.

http://www.xbitlabs.com/news/mainboa...rocessors.html

ftp://download.intel.com/design/proc...t/31872703.pdf

Here's a quote from the last article I dug up.

Quote:
Problem: In a synthetic testing environment, Intel has observed that some processor, chipset, and motherboard configurations may experience reduced Front Side Bus (FSB) voltage margin during some certain die-to-die data transfers. This combination of configurations and data transfers is rare. This lower voltage margin could lead to FSB data bit errors, which can lead to unpredictable system behavior.

Implication: When this erratum occurs, it leads to FSB marginality in the system during processor die-to-die transactions, which can lead to unpredictable system behavior. Intel has not observed this erratum with any commercially available software.

Workaround: None identified.

Status: For the steppings affected, see the Summary Tables of Changes.