The Wolfdale Vtt ( FSB Termination ) game.Place your bet!

[TTimmy]

No updates yet?

[Marvin_The_Martian]

BZ I'm sorry for calling you an CPU destroyer, busting this Myth you farted intro Intels smile and I clap for that.
But there is one thing that should concern al Silicium based things on this planet. You will eventualy get them and OC the crap out of them ...
And I don't want an BZ Microwave ! *Runs on 1020V consuming 24A but turning diamonds intro coal and back.

Well I think this "test" is over and somehow "busted", feel free to grab a couple of Quads and OC and Overvolt the hell out of them. I'll subscribe that !

Only, and only if the stress test consists of crunching

jk, waiting on updates as well BenchZowner

[BenchZowner]

Some of the rigs are still running.
I had to stop some of them because I needed the hardware.

[dinos22]

No updates yet?

wow first post at XS and in this thread
nice going
welcome to XS

[BababooeyHTJ]

Some of the rigs are still running.
I had to stop some of them because I needed the hardware.

I have to know which ones. You are the man.

[zsamz_]

happy to see em still goin strong
n i thought 1.5v on a 45nm was pushin it

[BenchZowner]

Just finished a review with one of the E8600s... still the same max stable frequency at the same settings.

I'll also use this CPU on LN2 this Monday/Tuesday to see if the absolute max has changed (degraded)

[Hyperhorn]

To be honest, when I saw the threads at the beginning I didn't thought there would be zero problems after such a long time.

[TTimmy]

wow first post at XS and in this thread
nice going
welcome to XS

Hehe, been stalking on XS for a several months now and I just had to know how Benchzowner's 45nms were doing.

[BababooeyHTJ]

Just finished a review with one of the E8600s... still the same max stable frequency at the same settings.

I'll also use this CPU on LN2 this Monday/Tuesday to see if the absolute max has changed (degraded)

Are you trying to make me beat the hell out of my e8400? I cant wait for some results!

[sup3rfly]

did increase voltage for my crappy E8500 @ vtt to 1.9 and pll to 2.0 and orthos at pri10 for 1 day, it still alive but degraded haha
previously it need 1.425v to run spi 32m now it need 1.55v to be stable xD

[BenchZowner]

did increase voltage for my crappy E8500 @ vtt to 1.9 and pll to 2.0 and orthos at pri10 for 1 day, it still alive but degraded haha
previously it need 1.425v to run spi 32m now it need 1.55v to be stable xD

Stepping & Batch info ?

[NBF]

Just finished a review with one of the E8600s... still the same max stable frequency at the same settings.

I'll also use this CPU on LN2 this Monday/Tuesday to see if the absolute max has changed (degraded)

Nice to hear that BenchZowner, great job you did. We know all CPUs are fine, but is there any dead motherboard?

Now you could start all over, we wanna know how high we can set QPI and Vmem on i7!

[BenchZowner]

Nice to hear that BenchZowner, great job you did. We know all CPUs are fine, but is there any dead motherboard?

Now you could start all over, we wanna know how high we can set QPI and Vmem on i7!

Haha!
I'm going to move on to Nehalem soon.
Just let me get some CPUs first and then some extreme testing is guaranteed

Unfortunately, all my motherboards are still alive
Some of those CPUs & MBs will be used in various rigs that I'm planning to use for various purposes in my house network.

No wonder, if I decide to sell some of them... it's going to be really hard to find a buyer

[TTimmy]

bump for the result.

[xlastshotx]

Bump again for results

[mikeyakame]

did increase voltage for my crappy E8500 @ vtt to 1.9 and pll to 2.0 and orthos at pri10 for 1 day, it still alive but degraded haha
previously it need 1.425v to run spi 32m now it need 1.55v to be stable xD

thats probably a result of the extremely large gap between vcc and vtt values. too much vtt with respect to vcc and you'll overshoot the transient vcc peak far beyond anything its designed to handle. too much vcc with respect to vtt and you'll end up damaging the chip over time from switching the instaneous voltage during on/off cycles, and having no vtt buffer to reduce the max voltage at peak and cut supply if the swing period exceeds 10ns if i recall.

maintaining a little overhead with vtt on vcc i've found works the best in my testing, and is probably why i've been able to push some reasonably high voltages 24/7 on my 65nm chip with no degradation as far as i can prove during the 1.5yrs ive owned it.

[nullface]

So what is the conclusion so far, if you dont plan on running your wolf for 10 years, 1.6-7V VTT is ok, yes?

[mikeyakame]

just keep vtt under 0.3v above vcc and you wont have a problem. once u increase potential across cpu landing pin input voltages u gradually begin to risk stressing the silicon as it tries to cope with constant differences between vcc and vtt which is used for input buffer threshold on low/high swings.

[BenchZowner]

The CPUs are okay.
Some of the motherboards ( the Striker II Extremes ) aren't
I RMA'd and sold them, I only have one S2E now, and I'm stuck between selling it and keeping it.
The CPUs used in the testing process will undergo an stability evaluation and will be circulated ( some of them will end up in some HTPCs ).

Now I gotta move on, Core i7 testing is 'bout to start

[Sneil]

sorry for my ignorance but what is the vcc?

edit: im guessing its the vcore...

[mikeyakame]

sorry, yes vcc = vcore!

[PanzerIV]

thats probably a result of the extremely large gap between vcc and vtt values. too much vtt with respect to vcc and you'll overshoot the transient vcc peak far beyond anything its designed to handle. too much vcc with respect to vtt and you'll end up damaging the chip over time from switching the instaneous voltage during on/off cycles, and having no vtt buffer to reduce the max voltage at peak and cut supply if the swing period exceeds 10ns if i recall.

maintaining a little overhead with vtt on vcc i've found works the best in my testing, and is probably why i've been able to push some reasonably high voltages 24/7 on my 65nm chip with no degradation as far as i can prove during the 1.5yrs ive owned it.

Hi Mike! You seem to know a lot about overclocking, and I'm pretty sure you could have the answer to my question which is quite about the quote I've done. First of all, HERE's my computer specifications.

1- On Intel's website about my E8400 "E0", it says:
Thermal Design Power: 74.1°C
Thermal Specification: 0.85V – 1.3625V

• Does it means that if the sensor of the cpu, not the one for core0/1, goes over 74C that it will start degrading the cpu?

• Why does it shows the maximum as being 1.3625v if anybody can go beyond that? Does it just means that its the highest safe voltage if you are under the stock heatsink?

2- HERE's a picture showing my current temperature & voltage. So now come the question that is about the quote. A lot of people were saying that VTT voltage was pretty much useless and that it wouldn't help overclocking if you would raise it. At stock it's at 1.10v so I've raised it to 1.20v just for the hell of it, in case if it would help. I noticed that I also had to raise a lot my Nortbridge's voltage as I was getting close to the 500FSB range. It's set to 1.57v but the monitoring application "SmartGuard" reports 1.60v. I hate the way it shows the Vcore in the bios, on my DFI board at the current setting I've chose it shows +162mv and since I guess the stock voltage is 1.11v I should be at 1.273v if I'm not wrong but the average on load seems to be 1.34-1.36, with a peak 1.39v

• At 4.30Ghz, if I do a blend test on Prime95, 1 thread on 2 failed after 10 houres. However with all the same voltage if I would attempt 4.4xGhz, every single time the threads on blend would crash exactly at 17min.

I tryed raising the Vcore of 1 step above (+172mv?) and it would still fail. So if the NB and Vcore voltage can't save me anymore, then would it have something to do with the diffrence you were talking between Vcore and VTT voltage???

[George_o/c]

I have been using 1.6V+ for all my benches, for like a month now
No problems yet

[PanzerIV]

I have been using 1.6V+ for all my benches, for like a month now
No problems yet

Hi again, I've tryed something else now that I've read more information about secondary voltage settings. So I've bumped the VTT voltage from 1.20 (stock=1.10) to 1.30v AND also raised the "Clockgen Voltage Control" from 3.45v to 3.60v since the next step was like 3.9v... then I wanted to see if at 4.41Ghz I could go beyond that 17min wall with a blend Prime95 because without this voltage bump, 4.30Ghz worked 10 houres. Unfortunately it still failed at 17min and 19min for the other core.

It was also recommanded to raise a little the PLL voltage to gain stability at high fsb, stock being 1.55v (Intel maximum says 1.575v) but in the DFI bios, the next closest step is 1.75v which is way too high (-_-). So now what should I try. Raise even more the VTT voltage, raise my NB from 1.57v to 1.61v, cry and go back to 4.30Ghz? It was pretty unclear about how to set the VTT, do I have to set it as close as possible to my Vcore but never 0.3v over my Vcore? If yes, then what happens when the system goes in "Thermal Management/EIST" and lower the frequencies + the voltage of the cpu to 1.31v, my VTT will get much higher so it will cause damage?!? x_x

http://www.stooorage.com/show/30/34536_untitled0.jpg