Will 1.68v kill my965 i7 CPU?

B

BadgerMC

n00b
Joined
Mar 23, 2010
Messages
4
I started crashing with my Corsair Dominator 1866 OCed to 1997. It run this way for 6 months then started BSOD. I can get it running stable with my DRAM voltage at 1.68. Isn't this dangerous for the Core i7? If I set the ASUS P6T m-board DRAM to Auto it drives the voltage to 1.9v. The XMP profile will set it at 1.66 which is RED. One module will drop out at 1.64v. Why do they rave about the 1.65v and my last two m-boards only allow 1.64, 1.66. etc? I have found a few posts about some folks getting max memory performance with 1.68 but I can't fry this CPU!

Thanks for any ideas,
 
What Intel (yes INTEL said this, publicly, which is pretty unusual and so it got everyones attention, bit like God speaking directly to us) said, is that running with memory voltage above 1.65 volts will reduce the lifetime of the onboard memory controller and lead to eventual CPU failure. No ifs ands or buts about it and recommended a max of 1.65 V for long life etc.

Now 1.68 is not much above 1.65 and with the accuracy of on-board voltage monitoring you could actually be anywhere around 1.65 more or less as these settings are seldom dead accurate.

And it is not like a setting above 1.65 will kill your CPU tomorrow. The higher above 1.65 you go the more the chance of reducing the life.

Its your stuff and your call, investigate more, but if 1.68 gets you to where you need to be it is my opinion that you are fine and the CPU will likely last longer than you will keep it. Now if you said you needed 1.73V etc. I would have responded you are killing your CPU stop it and if you had said 1.8V I would have said your CPU is likely to die in the next few months. See ? its all about how much over 1.65 you go, and you are barely over.

So thats the story, Intel recommended/warned that Vdiim over 1.65V on these CPUs will reduce the life. Several people destroyed there CPUs before the warning came out as they where used to the older board with a separate memory controller that would easily tolerate high(er) Vdimm voltages.

And the better your heatsink and cooling is I would think the better off you are at a higher Vdimm. However there are internal to the cpu "spot" temps that a heatsink of any size cannot instantly draw heat away from and, well it gets technical but bottom line a good heatsink and cooling would be a very good idea since you are running the internal memory controller about as hard as one would want to based on Intel's recommendations.
 
Thanks for the help. I woke up this morning with that sick scared feeling and dropped back to 1.66v. My system is so snappy with 1.68v and stable. Maybe it's in my head? I just can't seem to find much info on 1.65v + so I'll keep searching.

Thanks again!
 
hmm check back in a few, I know on anandtech they mentioned the intel announcement, I will try to find.

http://www.anandtech.com/show/2658/7 last two paragraphs on page but still have not found what I was looking for.

eh tired of looking for offical Intel statement. all so far is that intel let its chipset customers / board builder know about the advice, apparently Intel provided a "flyer" with Intel motherboards warning about the issue but I cannot find one online, and several board manufacturers included stickers or where going to on board sold to enthusiasts. A google of "Intel memory voltage warning" will find reams of stuff you have likely already found. I first hears about it in an article on Anandtech but cannot find it, where the author in testing something mentioned he had been told by an Intel engineer that over 1.65 was not a good thing to do. That was good enough for me even though I am big on Intel datasheets and exact numbers. Not having an iX cpu has not made it an issue for me.


Really tho, 1.68 is not pushing it. if you can really feel the difference and it was my board I would not worry about it I just would not go any higher.
 
Last edited:
I also thought it was actually a recommendation that the memory voltage should stay within a certain range of one of the other voltages (like within 0.5v of QPI or something). Considering that voltage is just a "pressure differential", this could make sense. In the same way that you can get 7v for your fans by using the 12v and 5v lines, keeping the memory voltage relatively close to another voltage could keep it from "blowing out" the memory controller. If you have other voltages raised a bit, the higher memory voltage might not be as bad as if you were using really low voltages for the other components. I also can't seem to find what I was thinking of either.

That's why I just got 1.5v RAM and played it safe. =) I can raise the voltage 10% and still be at the 1.65v limit.
 
I really appreciate the info. I do have my other voltages to less than .5 from my memory. I found some info stating the .5v difference was the danger. I just thought after more than a year I could see some more feedback on memory voltage and the i7. If I leave every setting on te P6T to Auto the memory goes to 1333. I didn't pay through the nose for this 1866 Dominator so I run the BCLOK up to get my 1866 and the board will set the DRAM to 1.9v. It indeed had a huge sticker on it about 1.65v.

I just wonder why it run for 6 months and then started BSOD's? I did a fresh install to eliminate program issues. I still crash. I read memory will start deteriorating after time.

Thanks so much for the help.
 
ok, this is something thats concerned me too. So I checked the manual for the 930 i7, and found nothing other than the keep ambient tempretures below 38℃. But on an asus motherboard manual I got last week it says:

"According to INTEL CPU spec, DIMM voltage below 1.65v is recommended to protect the CPU."

It goes on and on about "memory for overclocking not running at vendors stated rate" due to "SPD" and to set it manually. With everything, stated limitations are usually just ass covering, and i'm sure theres some kind of margin of error worked into everything.
 
BadgerMC said:
I really appreciate the info. I do have my other voltages to less than .5 from my memory. I found some info stating the .5v difference was the danger. I just thought after more than a year I could see some more feedback on memory voltage and the i7. If I leave every setting on te P6T to Auto the memory goes to 1333. I didn't pay through the nose for this 1866 Dominator so I run the BCLOK up to get my 1866 and the board will set the DRAM to 1.9v. It indeed had a huge sticker on it about 1.65v.

I just wonder why it run for 6 months and then started BSOD's? I did a fresh install to eliminate program issues. I still crash. I read memory will start deteriorating after time.

Thanks so much for the help.
Click to expand...

Are you getting the crashes from having too high a voltage or for running it at 1.65 v or lower? Memory does not deteriorate in 6 months. It might be defective if the memory is rated at 1.65 and needs higher than normal voltage to run. Contact Corsair and see if they can send you replacement sticks.
 
I get the crash with 1997mhz setting. The voltage bump seems to make it run as it did since 9/09 at 1997. I can drop to 1985mjz and be rock solid. This kit was one of the bad batch last summer and it has been replaced. It's just you have been there it's hard to drop back : |
 
You must log in or register to reply here.
Back
Top