Wolfdale Vcore 1.45v: Max, or Myth?

Printable View

04-04-2008, 06:15 PM
slugzkea

Wolfdale Vcore 1.45v: Max, or Myth?

Okay I can't take it anymore. I've been reading a lot on these wolfie cpu's and a lot of people are stating that 1.45v is the rated, safe max vcore temp for the wolfdales.

However, lots of hardcore enthusiasts have breached that by large margins to achieve extreme overclocking results. There's been people applying 1.6v on air, and above on water reporting no negative complications or regrets of doing so with proper cooling.

So lets narrow it down a little bit:

Intel States:

- 1.45v MAX vcore
- Rated Thermal Threshold: 72.4c
- Max Tj. 95c

Intel gives you for stock speeds:
- really small heatsink
- crappy paste.

Now I don't know how well the wolfdales do at stock, with stock hsf, in a normal case with normal temps. I'm assuming the cpu hits 60'esh on load, since the default hsf is very very small.

Now here is what has caught my attention:

- Why rate the thermal threshold at 72.4c when only rating max vcore of 1.45? I doubt the cpu would ever reach 72.4 on 1.45 unless you have some serious crappy cooling, or you throw your rig in the microwave/oven.

- Why a max Tj. of 95c when the cpu will never come close to even 72.4c on stock speeds, stock volts, stock hsf? that's a lot of headroom.

So I decided to do some extreme, risky tests to find out the truth. I honestly thing the rated max specs are bull-poo, so lets find out.

- No one can say "oh yeah i used over 1.45v on my wolfdale and it died in a year or two" because the cpu isn't that old yet.

- No one can state that the cpu will not last 5-10 years using over 1.45v, since once again, unless some one comes back from the future, it's all in the air as theory.

Okay now on with the tests.

WARNING! I DO NOT EXPECT ANYBODY TO ATTEMPT WHAT I HAVE. ALL WARNINGS AND RISKS HAVE BEEN TAKEN INTO CONSIDERATION. I DO NOT RECOMMEND ANYBODY ATTEMPTING OVER 1.53V ON YOUR WOLFDALE UNLESS YOU HAVE EXTREMELY GOOD COOLING!

If my cpu burns, I won't make a huge deal. I accept the risks.

-----------------------------------------------------------------------
Test one: Thermal-Threshold

Okay so if the wolfie breaches 72c it's suppose to declock itself and its volts to prevent damage, right? We've seen this behavior in previous intel cpu's such as the prescotts or "preshotts". I enabled TM2 in my bios and applied 1.6v to my e8400 to attempt to sustain 4.43ghz on orthos.

Temps went as high as 77c, with no sign or report of thermal-throttling. CPU-Z didn't show any declock, or any voltage drops.

well, guess that isnt' the actual thermal threshold then.

Test Two: CPU High Temp Auto Shutdown

Later Intel cpu's shut themselves off once they reach a certain temperature. With a max Tj. of 95c and a rated thermal threshold of 72c, i would assume the cpu would shut down around 75c'esh but it did not, with TM2 enabled or disabled.
With TM2 disabled, the max temp I saw my cpu jump to was 86c and that was too close for comfort for me. So I backed off.
My best record for orthos is around 20 min's but decreasing my northbridge voltage has shown greater stability with orthos. My goal is to reach 2 hours with 1.5-1.6v 4.43ghz. Temps at load are bouncing between 74c and 77c.

So What's the deal? i mean sure, at those temps a wolfdale would probably not last in life for years to come, but who is going to have a wolfdale 3-5 years from now anyways? no one, unless you live in thailand or some other primitive country (technology wise).

PLEASE DO NOT COMPARE THESE RESULTS WITH AMD CPUS. AMD CPUS HAVE ZERO TOLERANCE TO HEAT DUE TO THE ONBOARD MEMORY CONTROLLER AND CURRENTLY, AMD IS VERY FAR BEHIND WITH TECHNOLOGY.

I'm looking for feedback from people who have really pushed the wolfdales. Please keep all comparisons with Core2duo e8xxxx WOLFDALE CPUS ONLY!

Thanks :) :welcome:
04-04-2008, 07:01 PM
RunawayPrisoner

The 72.4C is the maximum recommended thermal limit. Meaning... if you operate your processor for long terms above that limit, your processor might burn out or die out faster. It's no big deal to those who are going to update again next year anyway, but to those who don't plan on updating next year or so, or to those who fold 24/7, the processor's longevity is of great importance.

Next, even at stock volts and stock clocks, the processor will still heat up past 75C if ambient temp is high enough. Imagine what ambient temp in Vietnam is like? 38C! It's around 18-22C here in California, and that's no big deal. The processor loads its max at 55C. Fine. But a jump of 25C in ambient would result in the almost same jump in CPU temp, or even more. 55C would then become something like 75C or even above. That's why to run stock in such countries, extreme air cooling is required. And to run OC clocks, they need to either put the computer in an air conditioned room, or have it water-cooled.

Last but not least, only when you reach 95C will you see thermal throttling. It's easy to get there: just don't install a HSF. Your processor will throttle within a minute or two. Maybe even faster in hot places. 86C shows that you're too scared to go higher... but then, why do you even question this in the first place if you are too scared to do it yourself? Experiment and see what fits you best, ya?

Oh, and Intel's specsheet probably meant that the processor may survive with up to 1.45v (but not really sure if the processor will operate normally and stably for long-term use), but beyond that, they are not sure if the processor can even stay for a minute. The maximum "Safe" voltage specified is 1.3625v for all 45nm processors according to the specsheet. Those who are running above specs are suicide runners who only do such things for like a day or two. Most of them then go down to 1.36v or even lower, anyway. The true goal of real overclockers is to reach high clocks at as low a voltage as possible, not to reach high clocks only. I mean... if you want 6GHz stable, you just need to crank in more volts than it's required, then put the processor under some phase change or dry ice, ya? No big deal! :p:
04-04-2008, 07:08 PM
slugzkea

a lil scared, yes

It's not a matter of being afraid of reaching those temps, since to me, the behavior and result of my e8400 in the 90c's is unknown.

I love my wolfdale, and my goal isn't to burn it. My goal is to find out what it can truly handle without degrading performance, and without hitting Max Tj.

Yes, every OC'ers dream is high OC with low volts but realistically, Higher frequencies require more power, it's ohm's law is it not?

Anyways, I don't want to argue. I'm not trying to prove a point about all go for high volts or anything like that.

I just want the truth! ^.^
04-04-2008, 07:17 PM
RunawayPrisoner

Quote:

Originally Posted by slugzkea

Yes, every OC'ers dream is high OC with low volts but realistically, Higher frequencies require more power, it's ohm's law is it not?

Mmm... people once thought the earth was flat, too! Just experiment, and see for yourself. :) Sometimes... not believing is believing!

Quote:

Anyways, I don't want to argue. I'm not trying to prove a point about all go for high volts or anything like that.

I just want the truth! ^.^

And I offer you one truth: no truth is the truth. Like... I still don't believe I'm among the few people who can ever get a X3350/Q9450 to 3.80GHz, and I'm working to see if it's chip-related or motherboard-related here. Life is a big experiment to me. :)
04-04-2008, 07:22 PM
slugzkea

It's all good my friend, i totally agree.

But lets wait for other people to share their experiences so we can all combine knowledge to achieve practical science that works :)

Theory don't mean shiat!
04-04-2008, 07:28 PM
dragonhunter

I have been running my E8400 @ 4.4ghz using 1.525V with Vapo LS seen the CPU first came and still running strong here. By the way, my PC is always running at full load whenever it powers on.
04-04-2008, 07:29 PM
RunawayPrisoner

Ya, and I've only owned two chips this way. I'm getting another E3110 next week, and then I'll see what happens. :p
04-04-2008, 07:39 PM
slugzkea

stability

Quote:

Originally Posted by dragonhunter

I have been running my E8400 @ 4.4ghz using 1.525V with Vapo LS seen the CPU first came and still running strong here. By the way, my PC is always running at full load whenever it powers on.

Were you able to run prime95 or orthos for 2hrs? what temps?
04-04-2008, 09:16 PM
mrcape

I think we're all on new ground with these suckers so my only vote would be time will tell. Of course, I hope that degradation is a myth, but I've seen things that make me think otherwise.

Bottom line I say go for it! I will continue to push my chips.
04-05-2008, 03:28 AM
_hiVe

Quote:

Originally Posted by slugzkea

Temps went as high as 77c, with no sign or report of thermal-throttling. CPU-Z didn't show any declock, or any voltage drops.

well, guess that isn't the actual thermal threshold then.

You didn't stated the means of temperature monitoring... Which leads me to believe that the temp's you've witnessed were not correct.
So how did you measure the temperature lvl's?

I'm pretty sure the specified 72,4c is indeed the thermal threshold where the CPU begins to throttle itself down.

Additionally to verify the truth behind the "tj. max" please first reach 95+ degrees.~
I think the writer of RealTemp ...umm unclewebb has already been through this, it'll be much easier to just look here.

As for max voltage, personally I didn't seen/heard of anyone with a wolfdale that died at 1.45v "yet" but it doesn't mean it won't happen in about 8 months of 24/7 usage....I think we all know the line from Intel's datasheet for 45nm CPU's that "neither functionality nor long-term reliability can be expected" :rolleyes:

Anyway throwing things asside, dragonhunter's feedback would be warmly welcome as he claims to be running his 24/7 at voltages that lie absolutely outside manufacturer's specified max limits.
04-05-2008, 04:08 AM
slugzkea

Quote:

Originally Posted by _hiVe

You didn't stated the means of temperature monitoring... Which leads me to believe that the temp's you've witnessed were not correct.
So how did you measure the temperature lvl's?

I'm pretty sure the specified 72,4c is indeed the thermal threshold where the CPU begins to throttle itself down.

Additionally to verify the truth behind the "tj. max" please first reach 95+ degrees.~
I think the writer of RealTemp ...umm unclewebb has already been through this, it'll be much easier to just look here.

As for max voltage, personally I didn't seen/heard of anyone with a wolfdale that died at 1.45v "yet" but it doesn't mean it won't happen in about 8 months of 24/7 usage....I think we all know the line from Intel's datasheet for 45nm CPU's that "neither functionality nor long-term reliability can be expected" :rolleyes:

Anyway throwing things asside, dragonhunter's feedback would be warmly welcome as he claims to be running his 24/7 at voltages that lie absolutely outside manufacturer's specified max limits.

real temp is false. it showed 10 c cooler than lavalyst and cooltemp.

When my next paycheque comes in, perhaps i'll reach max Tj. but until then, maybe some one else would like to test it before me? :P
04-05-2008, 04:36 AM
Grnfinger

I run at 4.4HGz @ 1.46v 24/7 and have never had a hicup.
Temps are around 28 idle and 55 load both cores, if the fireplace is on the temps get a tad hotter on load but nothing I would be concerned about.
04-05-2008, 04:39 AM
IanB

Quote:

Originally Posted by slugzkea

real temp is false. it showed 10 c cooler than lavalyst and cooltemp.

Or alternatively, both lavalyst and cooltemp are false and are making the same mistake. Having read unclewebb's thread and his very careful, methodical analysis, I'd be inclined to trust him over the unconfirmed guesstimates of other programs, even if they look more professional. :rolleyes:
04-05-2008, 07:05 AM
Zucker2k

Myth!
04-05-2008, 08:43 AM
slugzkea

Quote:

Originally Posted by IanB

Or alternatively, both lavalyst and cooltemp are false and are making the same mistake. Having read unclewebb's thread and his very careful, methodical analysis, I'd be inclined to trust him over the unconfirmed guesstimates of other programs, even if they look more professional. :rolleyes:

true, but what if he's wrong as well?

that means that the temps are showing 10c cooler than they actually are, which puts us at even more risk.

Either way, in the near future I will apply probably 1.7v which will allow the cpu to reach 90c'esh in orthos and then we'll see if it'll thermal throttle or not.
04-05-2008, 08:53 AM
ownage

Quote:

Originally Posted by Zucker2k

Myth!

Is it?
http://www.xtremesystems.org/forums/...ight=degration
04-05-2008, 11:13 AM
dragonhunter

Quote:

Originally Posted by slugzkea

Were you able to run prime95 or orthos for 2hrs? what temps?

Prime95 +24hrs stable and folding whenever the PC is on. For temp, I don't know cause CoreTemp stuck at like 7C (buggy) but my Vapo LS LED temp is like around -47C max.....
04-05-2008, 11:29 AM
slugzkea

Quote:

Originally Posted by dragonhunter

Prime95 +24hrs stable and folding whenever the PC is on. For temp, I don't know cause CoreTemp stuck at like 7C (buggy) but my Vapo LS LED temp is like around -47C max.....

You must have one of those l33t chips, mine could never do 4.4 at 1.43v
I'm at 4.43 right now, using 1.57v and things are good for now....can't do prime or orthos though.
04-05-2008, 01:55 PM
fordf250

Quote:

Originally Posted by ownage

Is it?
http://www.xtremesystems.org/forums/...ight=degration

How many posts have you seen of people with ram problems and new motherboards? I think alot of the people had ram problems and thought it was the cpu.
04-05-2008, 02:05 PM
RunawayPrisoner

Quote:

Originally Posted by slugzkea

true, but what if he's wrong as well?

that means that the temps are showing 10c cooler than they actually are, which puts us at even more risk.

Which is why you should measure TJ yourself, and see which program would fit you best. To me, even "--" (which takes off an additional 5C) still does not reflect the actual temperature on the CPU die. It's actually 2C cooler than that... so for my CPUs, RealTemp with "--" was off by 2C, and CoreTemp or Everest or any other programs were off by 17C. It's about 68F here around evening, and the computer is running 38-40C (open case) then even when nothing is running? Yeah, right... :p

Quote:

Either way, in the near future I will apply probably 1.7v which will allow the cpu to reach 90c'esh in orthos and then we'll see if it'll thermal throttle or not.

Or just run without a HSF. :)
04-05-2008, 03:29 PM
twiggy

Quote:

Originally Posted by slugzkea

You must have one of those l33t chips, mine could never do 4.4 at 1.43v
I'm at 4.43 right now, using 1.57v and things are good for now....can't do prime or orthos though.

I wouldn't say its because of a "l33t" chip, I'd say its because of his "l33t" cooling.;)
04-05-2008, 05:42 PM
slugzkea

cooling

Quote:

Originally Posted by twiggy

I wouldn't say its because of a "l33t" chip, I'd say its because of his "l33t" cooling.;)

Is he on water? I'm on Truck Air :)
04-05-2008, 05:51 PM
003

I have not read the thread and I voted only time will tell. However I have some interesting information.

I damaged my E6750 (a 65nm CPU) with 1.71v (as set in bios; actual voltage was lower due to vdroop) on water, and now anything over stock speed on it causes the system to crash and reboot after a minute or two of OCCT.

All I know is that when I get my Q9450 on the 8th, I'm not going over 1.3625v :)
04-05-2008, 06:56 PM
RunawayPrisoner

Quote:

Originally Posted by 003

I have not read the thread and I voted only time will tell. However I have some interesting information.

I damaged my E6750 (a 65nm CPU) with 1.71v (as set in bios; actual voltage was lower due to vdroop) on water, and now anything over stock speed on it causes the system to crash and reboot after a minute or two of OCCT.

All I know is that when I get my Q9450 on the 8th, I'm not going over 1.3625v :)

I guess you've damaged your board, not your CPU. If you damaged your CPU, the poor thing wouldn't even boot... at all. :p

And by the way, I thought 4400MHz was normal? Here are my results with an mATX board:

http://i235.photobucket.com/albums/e...s/SuperPI3.jpg

4500MHz would be stable at 1.55v, but I didn't bother using that much. :) Just thought I'd point this out.
04-05-2008, 07:07 PM
003

Quote:

Originally Posted by RunawayPrisoner

I guess you've damaged your board, not your CPU. If you damaged your CPU, the poor thing wouldn't even boot... at all. :p

Well not damaged, but badly degraded the CPU? How would it have damaged the board? I really hope that's not the problem...

Has anybody ever actually damaged your mobo with to much vcore, before damaging (or "degrading") the CPU? That dosen't really sound plausible, but then again... now I'm worried.
04-05-2008, 07:09 PM
slugzkea

Quote:

Originally Posted by RunawayPrisoner

I guess you've damaged your board, not your CPU. If you damaged your CPU, the poor thing wouldn't even boot... at all. :p

I agree.
04-05-2008, 07:14 PM
RunawayPrisoner

Quote:

Originally Posted by 003

Well not damaged, but badly degraded the CPU? How would it have damaged the board? I really hope that's not the problem...

Has anybody ever actually damaged your mobo with to much vcore, before damaging (or "degrading") the CPU? That dosen't really sound plausible, but then again... now I'm worried.

The voltage still runs through the board before it goes to the CPU... you know. :) If your CPU is degraded, it'll require more voltages for higher overclocks, that's all. But if your board is damaged, it won't be able to take high volts anymore. Actually, stock volts (which are quite low) may still have troubles later on.
04-24-2008, 09:58 PM
slugzkea

I switched mobo's to an asus rampage which is able to lock the vdroop. I have way more stability at higher clocks.

I believe there is no degration, and that 1.45v is not max. I've pushed my wolfdale (everest and coretemp readings) close to 80C and the cpu is still operating fine, passing stress testing at 4.3ghz 1.57v

So if your cpu is demanding more volts, and crashing..and demanding more volts, i believe it's the lack of load calibration/ too much droop causing instability.
04-24-2008, 11:28 PM
Leeghoofd

Well do the test then push it up to 1.6volts and let her sweat, see if the CPU can continue to operate at the same voltage level 24/7... All I have seen is that some CPU's give in and require more volts than before, some stay stable...

It's hard to gather bulletproof information from anyone, as too many factors are involved... but if the CPU does the same on 3 different boards with 3 different ram sets and PSU's I am pretty sure it's either my inability or that the CPU that has degraded...
I sold my "degraded" Qx and the guy is pretty happy with it, but if I told the dude that it did 4ghz at 1.31 before and now almost requires 1.39volts he will be pretty pissed... and that was under high end watercooling...

I have no idea what caused it too degrade ( max was 4.6ghz for short benching sessions) was it the Vcore, CPU VTT, CPU PLL (never needed above 1.6 here), GTL voltages that made it degrade ? No idea really...:shrug:

You can do all the testing you want,as long as there is no 100% reliable info from Intel, I will still keep on advising my friends to limit max volts to 1.45 and FSB Term voltage to below 1.35 for 24/7 use on high end air or water... just to be on the safe side...;)

To claim degradation doesn't exist seems to be a harsh statement... there are many peeps out there that are not aware what they are playing with, but many many here are...

And they person that uses 1.55volts daily is using a pretty high end cooling setup providing sub zero temps 24/7... there's no way you can compare data between that setup and everyone else that use air or water cooling...
04-24-2008, 11:38 PM
RunawayPrisoner

I believe degradation exists, but at this point, I will have to say that degradation is not caused by vCore. I have had my processor at up to 1.55v to prime, bench, kill, and basically for anything imaginable and it's still fine. I believe, however, that the degradations are caused by VTT being set too high. Remember back a few days please? What did Intel just confirm for Anandtech's dead QX9650?
04-25-2008, 12:01 AM
Leeghoofd

In their specific case it was caused by CPU VTT or FSB termination voltage, though mine have never been above 1.4 at all... But in reality I really would like to know how Intel determined the cause of death ? I'm pretty convinced in my case that the 780i tests damaged the CPU... board I had had horrible Vdroop/drop and due to lack of time I didn't test the voltages at all with the DMM... I won't make that same mistake again...
04-25-2008, 12:16 AM
RunawayPrisoner

How high was VTT when you put the CPU in the 780i board? I mean... if you can remember that much.
04-25-2008, 12:27 PM
slugzkea

I believe degration exists too, but it doesn't with my e8400 and neither for my friend.

He uses 1.55 VTT and I use 1.5VTT and he uses 1.62v to sustain 4.2ghz while I use 1.57v to sustain 4.36ghz.

I haven't crashed yet. Prime stable, everest stress test stable.

Ever since my rampage has been able to lock the vcore unlike any other boards i've owned (0.01 droop, or no droop at all) my OC has been far more stable.
08-15-2008, 07:04 AM
oli_ramsay

I'm wondering if the 1.45v max is tru too. I've just got a new e8400 and am having to run 1.525volts to reach 4.2GHz which was my target for 24/7 usage. But I'm wondering if that voltage is too high and I should settle for maybe 4Ghz at a lower voltage. I plan on keeping this setup for a couple of years at least and don't wanna kill my chip.

What do you guys reckon :confused:

thanks :toast:
08-15-2008, 09:30 AM
ReverendMaynard

Quote:

Originally Posted by oli_ramsay

I'm wondering if the 1.45v max is tru too. I've just got a new e8400 and am having to run 1.525volts to reach 4.2GHz which was my target for 24/7 usage. But I'm wondering if that voltage is too high and I should settle for maybe 4Ghz at a lower voltage. I plan on keeping this setup for a couple of years at least and don't wanna kill my chip.

What do you guys reckon :confused:

thanks :toast:

I really believe that if you have a high vcore, but keep your temps nice and cool without much full load stress, your cpu should last indefinitely. It's some of the other cpu related adjustments that pose a short term threat. PLL at 1.6v is my cut off. 1.3v FSB term (VTT) is also my cutoff. Some may think that's high, but that what I'm personally comfortable at. People who go nuts on those adjustments are the ones that mutilate their cpus.

1.525v seems a bit aggressive for 4.2ghz. I can't really compare my scenario to yours due to different configs, but 1.53v gets me to 4.5ghz pi stable.
08-15-2008, 09:47 AM
oli_ramsay

Thanks, I've decided to settle for 4Ghz at 1.45v (1.416 load). Don't really wanna damage it as I don't plan on upgrading for a while now.
08-15-2008, 08:10 PM
CryptiK

There has not been enough time to determine whether or not here is long term damage caused by 1.45v +. However there are many reports of degrading cpu clocks, some of these reports are from meticulous people whom I am inclined to believe.

Remember the old Northwoods? They were run at up to and beyond 1.75v (intel's stated absolute max for them) and they took it for months, then the reports of degradation began. Some degraded so badly they would not even run at stock speeds anymore regardless of voltage.

Intel give their safe max limits (vcore = 1.3625v, vTT = 1.155v, vPLL = 1.575v), and they base that on extensive testing, and the need to not have people returning their cpu's in X months due to degradation. As every electrical device degrades due to gradual electromigration (which is affected by temperature) over time, they have designed the cpu's to last a reasonable amount of time (years) at the said max voltages and operating temperatures.

To go beyond these stated safe max limits, is sure to degrade the cpu faster, perhaps in less than 12 months. Hence, only time will tell, but those limits are real limits, intel is not "lying" to us.
08-15-2008, 08:15 PM
Loser777

I would never feed a wolfie more than 1.38v for 24/7 clocks... Even the 1.55v max for 65nm seems a bit too liberal for me.
08-16-2008, 05:07 AM
natedog420

im given my e8400 1.465 at idle and 1.44 at load for just 4ghz which sucks big time. i dont know if its my settings or not but it needs high vcore to be stable and always has since i bought it. pll is at 1.5 and vtt is at 1.27 so them volts are low to me but dont seem to help bring the vcore down and raise those to be stable still. i just think i have a dud e8400 on my hands.:down: