- Joined
- 2 Feb 2010
- Posts
- 345
- Location
- London
looks like I guessed spot on then...
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Sandy Bridge 2500k Just Died
- Thread starter lollyhayes
- Start date
looks like I guessed spot on then...
Hi mate
This is spot on, core voltage is killing them, fact!
Memory voltage is recommended at 1.50v, but higher should not harm them, but 1.65v is maximum voltage, any higher will harm Sandybridge.
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
I set my i7 920 do @ 4.2ghz at the safer vcore 1.30v and memory ram @ 1.50v and still fine and stable for 24 hours. No need to increase any further in vcore nor memory voltage.
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
Just try 1.30v vcore and memory ram @ 1.50v for all sandybridge until it stable, if not stable, try push up notch on the vcore but do not go over 1.380v (at the bios)
Your CPU is 45nm whereas SB is 32nm, so a totally different kettle of fish when it comes to vcore.
Got my 2500k running at 1.195vcore @ 4.3Ghz and memory running at 1.5V, 1600Mhz effective and stock timings. Interestingly the memory is G.Skill Ripjaw rated at 1.6-1.65v but it runs very happily at 1.5V. As stated here though the CPU VCORE is the real killer.
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
Yes that's right. But, I got a cherry chip as it can cope with 1.30v as I had done 4.4Ghz at 1.34v but I leave as 4.2Ghz to kept my longlife on the cpu, memory, and the board on the safer side as I know 45nm is stronger chip than 32nm.
The next FEARSOME.....is ivybridge 22nm!!!!! Very scary time!
Yea, people said I kill my i7 920 at 1.45vcore but run it for almost 2 years without issue.
If you cool it wisely, and dont go bonkers with vcore, you can run it pretty high without running into a lot of issues.
Trying to benchmark it also will put pressure on the chip, instead of finding a good oc.
use common sense
Agreed, I may be taking the plunge friday (mis the no vat code but not pay day till then
). But the information in here is very useful.Also impressed by the response of Gibbo et al, giving us some info and doing a few tests.
Can I second that as this forum is great for help and OC are
Certainly showing how much they care
Thanks to Gibbo
Hm. This is interesting, and raises the question - what voltages (RAM and CPU) are the overclocked bundles and systems being sent out at? Having ordered mine last week, before the word from Intel and the tests, where were they set and will they actually be stable?
They will die eventually from electromigration (whether you overclock them or not), but I haven't personally heard of any CPUs that have died 2 days into being used. I guess the lower the die size the more sensitive they are to voltage and dying which doesn't bode well for future generations.
That being said these CPUs are designed to run at a certain speed so people who overclock them, in some cases well over 1Ghz faster than they are supposed to run, don't really have anyone else to blame but themselves when this stuff happens. That's a bit brutal but there is obviously no blame at the chip manufacturers door.
edit: I should point out that obviously running them in spec would mean their lifespan is well beyond its obsolescence.
Last edited:
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
If I buy sandybridge, I probably overclock it and leave as 4.5ghz below 1.30vcore and memory ram at 1.50v! (But I will not buy one cos I own my i7 920 anyway)
TBH I'm gonna stick with the Corsair, as that's the brand I always use, and just run it at 1333Mhz and 1.5v. The real-world performance difference will be minimal.
No killer 5GHz then at current showing, 4.5-4.6GHz looks to be typical 24/7 OC or about 35% OC. Not a bad result however it does mean that it is fairly easy to kill the CPU if you are not 100% sure of the settings to adopt.
My concern is that the mobos do not seem to fail safe. Example the PC fails to boot and you can recover on the next boot. This does not seem to be the case in some of the instances made. The CPU fails first.
Nice efforts OCUK in defining the limits, a pity that many users may not see this due to Intel not generally advertising limits prior to release. This may be because they do not actively encourage or wish to encourage over volting the CPU and the reason why bclk is fixed.
My concern is that the mobos do not seem to fail safe. Example the PC fails to boot and you can recover on the next boot. This does not seem to be the case in some of the instances made. The CPU fails first.
Nice efforts OCUK in defining the limits, a pity that many users may not see this due to Intel not generally advertising limits prior to release. This may be because they do not actively encourage or wish to encourage over volting the CPU and the reason why bclk is fixed.
Last edited:
Just for the sake of it I've swapped my ram out on my order from OCZ gold 1.65v 1333C9 for some Mushkin 1.5-1.7v 1600C9. Only cost an extra couple of quid and at least its sold as being 1.5. I'm sure the OCZ would have worked but what the hell.
Soldato
- Joined
- 13 Sep 2010
- Posts
- 2,564
so the current opinion is just to keep Vdimm@ 1.5v right?
my stuff just turned up and I'm itching to build but dont wanna fry my new kit
my stuff just turned up and I'm itching to build but dont wanna fry my new kit
Mushkins better anyway. Im quiet pleased how fast OC have gotten on the case and got back some usefull information.