Question for the expierienced overclockers

nlel1975 · 12 Jul 2008 at 20:33

I have a Q6600 1.325 VID, it can only manage 3ghz @ 1.3875v BIOS on my P35-DS3L that is Prime95 stable for at least 17hours. When I up the speed to 3.2ghz it fails rather quickly, upping the voltage makes it last a bit longer but I find I have to give it something like 1.45v REAL to make it run for over 2 hours and on stock cooling its way too hot and then it fails after approx 2hr 30mins.

How long should you run Prime95 before you know that the voltage is right and do you think mine is failing at 2hr 30mins becuase of heat or it a lack of voltage. At 3ghz it gets up to 78C but it never fails, do you think if I could keep the temps at 3.2ghz down to 78C it might last a bit longer or is the reason its failing purely down to a lack of voltage in which case a new cooler won't make much difference.

In theory the idea that it lasts for 2hr 30mins then fails becuase of heat is a good one, but then its been running at a very high temperature for most of that time anyway becuase I run small ffts. Can a CPU suddenly fail Prime after 2 hours becuase of heat or is it usually down to not enough voltage.

If I felt a new cooler would increase my overclocking I would invest in one but I have my doubts it would make any difference to Prime95 stability.

Cob · 12 Jul 2008 at 20:57

Using the stock cooler? It shouldn't fail Prime95 because of heat. The chip should keep going til it throttles itself.

8 hours of Small FFT's is considered to be stable. 2hrs 30mins won't be anywhere near long enough to run a full cycle.

Tbh it sounds like you have a crap chip (sorry) if it can't do atleast 3ghz on stock volts. A new cooler may help a bit, but, as they say, you can't polish a turd. Personally I'd offload the chip at auction and buy a new or used one.

gareth170 · 12 Jul 2008 at 21:09

its just pot luck on what cpu's can clock too. if it's stable at 3ghz i'd be happy with that, extra 200mhz u won't much ddifferents tbh

hundred-and-one · 12 Jul 2008 at 22:35

I've the same 1.3250v VID. I got to 3.2 GHz by using a x8 multiplier and a 400Mhz bus speed. Voltage is at 1.4v and DDR2 RAM at 800Mhz (ie, 1:1 ratio). Been running like this for a few weeks 24x7. Prime runs for at least 48hours without problems.

mmorpheus1 · 12 Jul 2008 at 22:59

I have the same VID on my q6600 and can get it to 3.6Ghz (8x450) on air prime stable on an asus rampage formula x48 chipset , same chip on the Asus P5n32e-sli nf 680i chipset would not go over 3.0Ghz. same ram and cooler , personaly i dont think the vid has much to do with it more the motherboard..

nlel1975 · 13 Jul 2008 at 12:51

mmorpheus1 said:
I have the same VID on my q6600 and can get it to 3.6Ghz (8x450) on air prime stable on an asus rampage formula x48 chipset , same chip on the Asus P5n32e-sli nf 680i chipset would not go over 3.0Ghz. same ram and cooler , personaly i dont think the vid has much to do with it more the motherboard..

I thought the P35-DS3L was a good board for overclocking in fact many people will testify to this, I will see if I can test my chip in my brothers ABIT IP35 board to see if it makes any difference.

Cob · 13 Jul 2008 at 14:15

The P35 DS3L should be grand for clocking a quad-core.

Corasik · 13 Jul 2008 at 14:42

Hows the power regulation on the DS3L, there's another post on here from someone with a DS3R who seems to have a bad case of vdroop.

PhoenixUK · 13 Jul 2008 at 15:26

0.008 drop on my p35ds3l.

It can also manage anything u to 3.85ghz ( at least that was my highest kamikadze shot on e2200 on this board ).

mrthingyx · 13 Jul 2008 at 16:01

PhoenixUK said:
0.008 drop on my p35ds3l.

It can also manage anything u to 3.85ghz ( at least that was my highest kamikadze shot on e2200 on this board ).

Which is sucking significantly less wattage than a Q6600 at such speeds.

Overclocking is, has never been, and never will predictable or a remotely precise science. It may well be one of those instances which everybody recommending Q6600s overlooks when stating "they're guaranteed to reach 3.2GHz" although it could be any number of other things.

Bung a fan over the PWM area, mem-test your overclock and see what happens. It could be that you have a rubbish chip, but it may well be something else.

DrSteve · 13 Jul 2008 at 18:19

Have had 2 chips with the same vid as yours. The 1st was a crap chip and needed silly volts to do anything above 3.2. Despite what Cob said earlier I did indeed try to polish a turd and lapped the chip.

Didn't help.
Ordered a second one which will do 3.6 but needs 1.5+ load to do that. As suggested best to flog the one you have and grab a new one which will hopefully do better.

marc2003 · 13 Jul 2008 at 19:15

PhoenixUK said:
0.008 drop on my p35ds3l.

It can also manage anything u to 3.85ghz ( at least that was my highest kamikadze shot on e2200 on this board ).

mine drooped a lot with my 2180 (can't remember how much exactly). and it droops 0.1v with my q6600.

mine isn't the best cpu as it takes 1.4625v set in the bios for 3.42ghz (380x9) stable. cpu-z shows 1.36v when priming in windows. but it's quick enough for me.

nlel1975 · 13 Jul 2008 at 21:09

I`ve just bought a used OEM one from a popular auction site which has been ran at 3.4ghz "flawlessly" for the last few months, so I should be able to get the same results with decent cooling and a 3.4ghz overclock is good for a Q6600, I know some are getting 3.6ghz but I`d be happy with 3.4ghz from the 3.0ghz I have now.

It cost about the same as a new one ex VAT but at least I have a pretty good chance of getting 3.4ghz with this one and the same is not always true for a new one bought randomly. I am trying to sell my system with the 3ghz chip in, so I will most likely hold onto the popular auction site one and use it in a future build. If I get fed up of trying to sell my system I can always flog the 3ghz chip at anytime.

With ocUK charging £ 120 ex VAT + postage for a random chip guaranteed to 3.3ghz (which isn't pre-tested) I`d figured I`d take my chances and believe what the seller says in that it runs at 3.4ghz flawlessly, there are plenty of chips out there that do this so I have no reason to suspect this is a lie, of course there is a risk that when I get it - it won't do 3.4ghz but I feel it was worth the risk.

gareth170 · 13 Jul 2008 at 21:18

hope that works well for u..

btw u should remove the price and site name.

Reality|Bites · 14 Jul 2008 at 14:52

It can be because of heat. The MAX temp quoted is only for stock speeds, If you overclock and increase voltages your max stable temp drops as well.

That's why you can get a better overclock using water/phase cooling. Not due to the fact you can increase the voltage more using better cooling.

nlel1975 · 14 Jul 2008 at 15:03

Reality|Bites said:
It can be because of heat. The MAX temp quoted is only for stock speeds, If you overclock and increase voltages your max stable temp drops as well.

That's why you can get a better overclock using water/phase cooling. Not due to the fact you can increase the voltage more using better cooling.

I thought the reason for water cooling was so that you can increase voltages thereby allowing you to push it a bit further to get a stable overclock at more reasonable temps then is achievable on air, I didn't think better cooling could make an unstable overclock stable. Surely as has been pointed out the chip should just keep going untill thermal throttling kicks in, these chips can run at up to 100C before that happens.

Greebo · 14 Jul 2008 at 15:33

nlel1975 said:
I thought the reason for water cooling was so that you can increase voltages thereby allowing you to push it a bit further to get a stable overclock at more reasonable temps then is achievable on air, I didn't think better cooling could make an unstable overclock stable. Surely as has been pointed out the chip should just keep going untill thermal throttling kicks in, these chips can run at up to 100C before that happens.

It's both really. I have one core on my q6600 which fails prime95 once it hits 73C.

I have been told technically that my cpu is faulty and I should send it back. Problem is, under water I can get it stable at 3.8Ghz and it's VID is 1.25v.

If I send it back, I may get a none faulty one with a VID of 1.325 which might not clock stable to 3.8Ghz so sticking with my current cpu.

If I wasn't on water, I certainly wouldn't be able to keep the core temp below 73C and hit 3.8Ghz stable.

nlel1975 · 14 Jul 2008 at 15:39

Greebo said:
It's both really. I have one core on my q6600 which fails prime95 once it hits 73C.

I have been told technically that my cpu is faulty and I should send it back. Problem is, under water I can get it stable at 3.8Ghz and it's VID is 1.25v.

If I send it back, I may get a none faulty one with a VID of 1.325 which might not clock stable to 3.8Ghz so sticking with my current cpu.

If I wasn't on water, I certainly wouldn't be able to keep the core temp below 73C and hit 3.8Ghz stable.

There is def something wrong with your chip then, my Q6600 @ 3ghz on stock cooling reaches 78C and its passed prime 17 hours before I stopped it so a q6600 shouldn't fail prime at 73C in fact they should reach 100C and keep going, at 1.45v 3.2ghz my chip did reach 97C but it didn't fail immediately ran for 2hr 30mins like that and no-one can prove it failed becuase of persistent heat. I just have a crap chip. Its never been worth me investing in a better cooler.

Greebo · 14 Jul 2008 at 16:03

nlel1975 said:
There is def something wrong with your chip then, my Q6600 @ 3ghz on stock cooling reaches 78C and its passed prime 17 hours before I stopped it so a q6600 shouldn't fail prime at 73C in fact they should reach 100C and keep going, at 1.45v 3.2ghz my chip did reach 97C but it didn't fail immediately ran for 2hr 30mins like that and no-one can prove it failed becuase of persistent heat. I just have a crap chip. Its never been worth me investing in a better cooler.

That's my dilemma though. I have a faulty chip capable of 3.8Ghz stable or do I risk a new, none faulty chip which might top out at 3.2Ghz like some peoples do on here.

If I had been on air, I would have returned it. I still might before it's warranty runs out but it will do for now

If thay have stopped making q6600 when I return it, what would I get as a replacement do you think since q6600 and end of line this quarter?

nlel1975 · 14 Jul 2008 at 16:19

Greebo said:
That's my dilemma though. I have a faulty chip capable of 3.8Ghz stable or do I risk a new, none faulty chip which might top out at 3.2Ghz like some peoples do on here.

If I had been on air, I would have returned it. I still might before it's warranty runs out but it will do for now

If thay have stopped making q6600 when I return it, what would I get as a replacement do you think since q6600 and end of line this quarter?

If you return it you have a chance of getting one which will do 3.0 - 3.2 a chance of getting one which will do 3.2 - 3.4 a chance of getting one which will do 3.4 - 3.6 and a chance of getting one which will do 3.6 - 3.8.

If it was me and it was 100% prime stable I wouldn't return it, and if you return it after they have stopped being made they will most likely send you one from their stock of replacements.