Tom's Hardware > Forum > Overclocking > CPUs > Overclocking without voltage increase

Overclocking without voltage increase

Forum Overclocking : CPUs - Overclocking without voltage increase

Tom's Hardware: Over 1.4 million members in 6 different countries available to answer all your high-tech questions. Sign up now! Its free!
Word :    Username :           
 

Hey Guys,



Planning on getting a core i7 920 (D0 stepping). If I was to do a mild cpu overclock without voltages, would this still increase the chance of CPU failure - even if it is the tiniest bit?

Thanks heaps

Sponsored Links
Register or log in to remove.

Short answer: Yes.
Long answer: No.
Longer answer: You won't be able to tell the difference, because the CPU will move on into the abyss of obsolescence before you reach the end of the ever-so-slightly-shorter lifespan.

Reply to randomizer

haha thankyou and nicely said :)

Do you think overclocking a core i7 D0 to around 3-3.33ghz with as minimal voltage as possible would affect lifespan - again even the tiniest bit?

thanks once again

Reply to fattymole123

Yes it will, but the difference is so negligible it's not even worth noting. Voltage does the most "damage" to a processor. If you want to try something interesting you could overclock to 3.33GHz and see if you can undervolt by some amount. This would effectively increase the lifespan even though you are overclocked. However, there's no guarantee that you'll be able to both overclock and undervolt by any significant degree, you just have to try. At the very least 3.33GHz should be doable at stock voltage, especially for a D0.

Reply to randomizer

I believe that heat is the killer of processors, there was an article on anandtech that discusses that a voltage increase does decrease life but heat does more damage. Besides after 3 or 4 years I think you'd be looking at something faster already.

Reply to amnotanoobie

Bah, heat does nothing to a modern CPU! Actually you'd be surprised what they can take, although it's not like you'd TRY to run it hot for fun (Ok, I would). On the other hand, 45nm HKMG CPUs are very voltage intolerant.

Reply to randomizer

randomizer wrote :

Bah, heat does nothing to a modern CPU! Actually you'd be surprised what they can take, although it's not like you'd TRY to run it hot for fun (Ok, I would). On the other hand, 45nm HKMG CPUs are very voltage intolerant.



hmm running it hot for fun... (uncfan imagines a heatsink shaped like a heat plate with eggs scrambling on top...yum)

Reply to uncfan_2563

randomizer wrote :

Yes it will, but the difference is so negligible it's not even worth noting. Voltage does the most "damage" to a processor. If you want to try something interesting you could overclock to 3.33GHz and see if you can undervolt by some amount. This would effectively increase the lifespan even though you are overclocked. However, there's no guarantee that you'll be able to both overclock and undervolt by any significant degree, you just have to try. At the very least 3.33GHz should be doable at stock voltage, especially for a D0.




Well done on Intel part for the overclockability (is that word..) of the i7 processor. You have been very helpful, i must thank you again! Do you believe it is worth overcloking the i7 920 to 3.0GHz or 3.33GHz for gaming purposes? or wud the difference be negligable?

Reply to fattymole123

amnotanoobie wrote :

I believe that heat is the killer of processors, there was an article on anandtech that discusses that a voltage increase does decrease life but heat does more damage. Besides after 3 or 4 years I think you'd be looking at something faster already.




Thanks for your response, i will be putting a good aftermarket cooler on it maybs the thermalright 120 extreme as it is around ther price of all good aftermarket coolers but i have heard is the best. I agree probably around 4-5 years maybe less, i will be looking for more speed..

Reply to fattymole123

I have an i7 with a DX58SO motherboard.

I have the option of using the default voltage or setting it manually.

The default voltage usually runs at .9v idle, 1.05v medium load, and 1.20v full load

The manual voltage is about 1.16-1.18v, to about 1.15v vdroop full load.


Which would be better for the cpu lifespan? (no matter how insignificant)

Reply to Bluescreendeath

fattymole123 wrote :

Well done on Intel part for the overclockability (is that word..) of the i7 processor. You have been very helpful, i must thank you again! Do you believe it is worth overcloking the i7 920 to 3.0GHz or 3.33GHz for gaming purposes? or wud the difference be negligable?


For gaming with how many video cards? If only one, you could probably leave the i7 at stock and not see much difference unless you play RTS games or Flight Simulator X. 330MHz is not a massive difference in any case, but if you can do both speeds at the same voltage then you might as well do it :D

fattymole123 wrote :

Thanks for your response, i will be putting a good aftermarket cooler on it maybs the thermalright 120 extreme as it is around ther price of all good aftermarket coolers but i have heard is the best. I agree probably around 4-5 years maybe less, i will be looking for more speed..


I think the 120 Extreme has a convex (concave?) base and needs to be lapped for optimal performance. The Noctua NH-U12P SE1366 performs well and is about the quietest cooling you can use for an i7. Plus you don't need to buy the fans separately :)

Bluescreendeath wrote :

Which would be better for the cpu lifespan? (no matter how insignificant)



Lower voltage is always better.

Reply to randomizer

randomizer wrote :

For gaming with how many video cards? If only one, you could probably leave the i7 at stock and not see much difference unless you play RTS games or Flight Simulator X. 330MHz is not a massive difference in any case, but if you can do both speeds at the same voltage then you might as well do it :D


I think the 120 Extreme has a convex (concave?) base and needs to be lapped for optimal performance. The Noctua NH-U12P SE1366 performs well and is about the quietest cooling you can use for an i7. Plus you don't need to buy the fans separately :)



Lower voltage is always better.




I am not suprised you are a moderator u have very good responses! Ok, so what does lapped mean? What would be the best air cooler that does not require to be lapped? Is lapping hard to do? I have heard good responses from the xigmatek s1283 and not sure about s1284 however I would presume it would be the same if not better, what do u think about these?

Reply to fattymole123

randomizer wrote :

For gaming with how many video cards? If only one, you could probably leave the i7 at stock and not see much difference unless you play RTS games or Flight Simulator X. 330MHz is not a massive difference in any case, but if you can do both speeds at the same voltage then you might as well do it :D


I think the 120 Extreme has a convex (concave?) base and needs to be lapped for optimal performance. The Noctua NH-U12P SE1366 performs well and is about the quietest cooling you can use for an i7. Plus you don't need to buy the fans separately :)



Lower voltage is always better.



I will be gaming on 1 GTX 260 core 216, however later on the track I might get two for sli. I mostly play FPS games like Crysis etc.

Reply to fattymole123

Lapping is using sandpaper to flatten and polish the contact surface of the heatsink and/or CPU (usually starting course to get it flat and moving to a fine one to polish it off). There's some guides around the Internet that are probably better and more up-to-date than the one that is here.

Reply to randomizer

randomizer wrote :


Lower voltage is always better.



ya, but which lower voltage? Lower voltage ever, or average lower voltage?

The default voltage usually runs at .9v idle, 1.05v medium load, and 1.20v full load ---> lower voltage on average

The manual voltage is about 1.16-1.18v, to about 1.15v vdroop full load ---> lower voltage ever

Reply to Bluescreendeath

Can't really say I know the answer to that. There's two ways you can look at it:

1) The maximum voltage does the most "damage" so reducing the maximum is best.
2) Reducing the voltage at which the CPU runs for the majority of the time is best.

I find it odd that you can't reduce both idle and load voltages though. I'm idling at 0.768V on my i7 920. All I did was drop the vcore by 0.16V from stock.

Reply to randomizer

randomizer wrote :

Lapping is using sandpaper to flatten and polish the contact surface of the heatsink and/or CPU (usually starting course to get it flat and moving to a fine one to polish it off). There's some guides around the Internet that are probably better and more up-to-date than the one that is here.




Thats odd, why would they make the base concave/convex for it to be lapped, I would have thought there machinery would do a better job.. What would you reccomend as a cpu cooler that does not require lapping?

Reply to fattymole123

I haven't got a clue. I've always wondered why they did that too. Never really looked into it much. As I said before, the Noctua NH-U12P SE1366 is good.

Reply to randomizer

randomizer wrote :

I haven't got a clue. I've always wondered why they did that too. Never really looked into it much. As I said before, the Noctua NH-U12P SE1366 is good.




Ok thank you VERY much for ALL your help it has not gone unnoticed! :)

Reply to fattymole123

randomizer wrote :

Can't really say I know the answer to that. There's two ways you can look at it:

1) The maximum voltage does the most "damage" so reducing the maximum is best.
2) Reducing the voltage at which the CPU runs for the majority of the time is best.

I find it odd that you can't reduce both idle and load voltages though. I'm idling at 0.768V on my i7 920. All I did was drop the vcore by 0.16V from stock.



aye...unfortunately my DX58SO doesn't have that option.

I have it on default, so it uses a lower voltage (down to .9) when idle.

If I set it manually, then the voltage hovers around that setting 24/7 and doesn't really go down during idle.


What voltage do you have your mobo/i7 set to? Did you change any other voltages/settings?

Reply to Bluescreendeath

I have done nothing but lower the vcore.

Reply to randomizer

I don't know. The BIOS I use only allows me to select offsets from the VID, so I'm at -0.16V now.

Reply to randomizer

fattymole123 wrote :

haha thankyou and nicely said :)

Do you think overclocking a core i7 D0 to around 3-3.33ghz with as minimal voltage as possible would affect lifespan - again even the tiniest bit?

thanks once again


Realize that the i7-920 and the 975 are exactly the same chip. Intel will produce more "best" binned 975 chips than they can sell for $999. They take the remainder and reduce and lock the multiplier to produce 950 and 920 cpu's. Each sample is a tiny bit different, but my take is that if you limit your overclock to 975 speeds, you are not doing anything that the cpu is not capable of handling.
If you raise the BCLK of a 920 from 133 to 160, you will get 3.33 with turbo, the speed of the 975. That does not require any voltage adjustments.

There is probably more difference in longevity from chip to chip samples than there is from a overclock with no voltage change.

Reply to geofelt

randomizer wrote :

I don't know. The BIOS I use only allows me to select offsets from the VID, so I'm at -0.16V now.



oh i see, what motherboard?

Reply to Bluescreendeath

Hmmm, what does cpu-z say? A range or a specific number?

I think Imma just copy your settings or something and test it out.

Reply to Bluescreendeath

A specific number but it bounces around with every little change in load. Plus reading voltages via software is not that accurate anyway.

 

EDIT: Checked my BIOS and it does actually say the current vcore which I didn't notice earlier. I'm at 0.992V.


Message edited by randomizer on 08-30-2009 at 09:31:00 AM
Reply to randomizer

geofelt wrote :

Realize that the i7-920 and the 975 are exactly the same chip. Intel will produce more "best" binned 975 chips than they can sell for $999. They take the remainder and reduce and lock the multiplier to produce 950 and 920 cpu's. Each sample is a tiny bit different, but my take is that if you limit your overclock to 975 speeds, you are not doing anything that the cpu is not capable of handling.
If you raise the BCLK of a 920 from 133 to 160, you will get 3.33 with turbo, the speed of the 975. That does not require any voltage adjustments.

There is probably more difference in longevity from chip to chip samples than there is from a overclock with no voltage change.




Thanks heaps geofelt, I believe you have commented nearly every one of my threads, thankyou very much u have been a great help! :) Ok im having a little bit of trouble understanding what you said last. So, the manufacturing process is not 100% accurate therefore every chip is different even of the same model, and the difference in longevity between every chip even of the same model would be greater than that from overclocking with no voltage? Is the performance difference in gaming from 2.66ghz to 3.33ghz on the i7 worth the void of warranty from overclocking?

Thanks heaps bud :DDD

Reply to fattymole123

It only voids the warranty if you tell them.
But I've never had a CPU fail on me my entire life, from 4Mhz 8088's to my last P4 which spent way too much of it's life at 70-80C.

You'll throw out or sell the i7 way before it fails I'd expect. I'd definately overclock it, just keeping voltages safe and temp under 50.

Reply to SpidersWeb

SpidersWeb wrote :

It only voids the warranty if you tell them.
But I've never had a CPU fail on me my entire life, from 4Mhz 8088's to my last P4 which spent way too much of it's life at 70-80C.

You'll throw out or sell the i7 way before it fails I'd expect. I'd definately overclock it, just keeping voltages safe and temp under 50.




So unless you admit to overclocking they have no way of knowing?

Reply to fattymole123

well if it's damaged they'd be able to tell but i doubt there is a way they can tell

Reply to uncfan_2563

If it's damaged they'd be able to tell that it's damaged. They won't be able to tell why.

Reply to randomizer

thats true... not all damage is caused by us overclockers :D

Reply to uncfan_2563

geofelt wrote :

Realize that the i7-920 and the 975 are exactly the same chip. Intel will produce more "best" binned 975 chips than they can sell for $999. They take the remainder and reduce and lock the multiplier to produce 950 and 920 cpu's. Each sample is a tiny bit different, but my take is that if you limit your overclock to 975 speeds, you are not doing anything that the cpu is not capable of handling.
If you raise the BCLK of a 920 from 133 to 160, you will get 3.33 with turbo, the speed of the 975. That does not require any voltage adjustments.

There is probably more difference in longevity from chip to chip samples than there is from a overclock with no voltage change.




Im a Total Noob, but I wanted to ask something... to achieve the 3,33 ghz at stock voltage, theres something else I have to raise besides BCLK? like ram speed or something? I want to build my first gaming rig,
thanks
sorry for noob question, but if its that easy then I will buy the components very soon :D



Albertomac...

Reply to albertomac

fattymole123 wrote :

Thats odd, why would they make the base concave/convex for it to be lapped, I would have thought there machinery would do a better job.. What would you reccomend as a cpu cooler that does not require lapping?



http://www.prolimatech.com/

"Prolimatech does not condone any type of lapping done to the CPU or to heatsink base. Every Prolimatech's heatsink base is designed on a pin-point scale of how the base is to be flat and/or curved where it's needed to be. We have programed our machines to machine the surface in a very calculated way. Any after-manufacture lapping or modding done to the base will alter the design, hence negating its performance factor as well as its warranty."

Reply to JackNaylorPE

fattymole123 wrote :

Planning on getting a core i7 920 (D0 stepping). If I was to do a mild cpu overclock without voltages, would this still increase the chance of CPU failure - even if it is the tiniest bit?Thanks heaps



http://www.tomshardware.com/review [...] 268-3.html

3.33 GHz and no voltage increase

Reply to JackNaylorPE
Tom's Hardware > Forum > Overclocking > CPUs > Overclocking without voltage increase
Go to:

There are 616 identified and unidentified users. To see the list of identified users, Click here.

Sponsored links
  • Ask the community now
  • Publish
Ad
They won a badge
Join us in greeting them