Printable View
i think i have heard lots of people said 1.35v on a 45nm = 1.45v on a 65nm one. is this true?
now i am running a e8400 at 4320mhz @ 1.368v for 24/7 (primed for 1hr stable, still doing it). do you guys think this is too high for the 45nm chip?
oh, and the chip is watercooled.
24/7 config:
http://img204.imageshack.us/img204/8076/247ob4.th.jpg
max clock so far:
http://img204.imageshack.us/img204/6868/4500ma6.th.jpg
super pi 1m at 4.3g.
http://img101.imageshack.us/img101/9806/piyu7.th.jpg
What abt your NB voltage?
quoted from mrcape:
http://www.movement-radio.com/intelspec.jpg
as u can see where as Intel rated for 65nm as 1.55vcore as the absolute maximum, and for 45nm 1.45vcore. I would personally use a notch lower for 24/7 , so 1.400vcore, I bet tons of people here @ XS would disagree with me, but this is just what I would use for these chips, actually no1 can confirm for sure, we need to wait till we see some death cpu after some time exceeding 1.45vcore what Intel specced as absolute maximum
hahahahaahaa now i see why you thought my name was weird!
I used to be running 4300 @ 1.48v (1.5v bios) but dropped it to 4000 @ 1.32v (1.35v bios) for peace of mind. I'm not worried about it dyeing or anything but don't want to lose MAX overclocking potential in time... I'll surely bench it again at ~4800-4900 + do some phase and stuff with it... My E6750 lost it's top end overclock (4820) by keeping it at 1.61v for 24/7 (in only one week) so better safe then sorry with this chip.
aw m8 sorry LOL typo:X edit :DQuote:
Originally Posted by mrcape
Till now , i now 8 CPUS (E8400,E8500) that degraded here in Athens,Greece . All are from close friends of mine . Six of them were watercooled(other 2 with aircooling) with voltages up to 1.55V .
One of the watercooled cpus degraded with 1.41V in 4-5 days.
The one with 1.52V degraded in just 2 days .
If you're worried , don't give more than 1.32-1.34V .
My CPU degraded with 1.49V watercooled in 4 days . Lost about 70-80Mhz in orthos & same voltages.
Quote:
Originally Posted by Stelios
Ummmm not good.... Other chime in and corroborate this please.
Does this cpu oc degration occurs too if we use sub zero temps on the cpu?
Can anyone give feedback on this point ?
Thks :)
I am trying 1.4875V right now.. I'll keep it at that for a few days. I'll let you know if it becomes any less stable overtime.
interestingQuote:
Originally Posted by Stelios
this is probably my first post here despite i registered few time ago.
i was discussing the maximum safe voltage to this 45nm processors in a portuguese forum (from where i was linked to here) and the information i found is probably usefull here too. so.. here it goes :)
those tables (the ones mc2k posted above) were found in the e6000 and e8000 processor series datasheet. in the e8000 series datasheet (pag 17) above the image we can read the following:
i'm just not sure about what are the "functional operation limits". i believe that we can assume that the functional operation limits are the vid range limits. at least it makes sense for me because if intel is manufacturing processors with a vid of 1.3625v it's because it's a functional operation voltage for the manufacturing process (45nm).Quote:
Table 3 specifies absolute maximum and minimum ratings only and lie outside the functional limits of the processor. Within functional operation limits, functionality and long-term reliability can be expected.
At conditions outside functional operation condition limits, but within absolute maximum and minimum ratings, neither functionality nor long-term reliability can be expected. If a device is returned to conditions within functional operation limits after having been subjected to conditions outside these limits, but within the absolute maximum and minimum ratings, the device may be functional, but with its lifetime degraded depending on exposure to conditions exceeding the functional operation
condition limits
with the info posted above and this info in the intel datasheet i'm a bit afraid that this processor series can't safelly go beyond the maximum vid limit (1.3625v) or, at least, far beyond..
EDIT: sorry if there are any grammatical/ortographical erros in my post but, as for many users here, english it's not my native language.
Quote:
Originally Posted by Bad_Cop
that table was already posted is xs
Can you elaborate on this? Degradation in a few days with a relatively low voltage increase sounds unreasonable. Did you run only orthos or other tests too? Ambient same, temps same? All other setting same?Quote:
Originally Posted by Stelios
the point wasn't to post the table (it's quoted anyway) but post the reflection and info found on the datasheet that, in my point of view, may be usefull to some users. specially those who have a rock stable overclock 24/7 and wanna keep the processor as healthy as possible.
despite this is a extreme overclocking oriented forum a lot of subjects discussed here are related to 24/7 overclock's and a lot of users who post here have only that kind of rig's.
but i'll edit my previous post. i didn't saw that the table is already posted in this thread.
i dont think id feel comfy about 1.45v...more like 1.35V(or less?); im just attempting to read between the lines of course...
how 'delicate' is the 45nm chip?:confused:
Right now (5th day now) i'm using my single stage cooler with voltages up to 1.75V . I'm gonna remove it in 1-2 days to check it again with watercooling .Quote:
Originally Posted by VoRtAn_MaDgE
It was a lot easier to check with orthos as my friends don't bench so they want to have 100% stable systems . 12+ hours orthos testing is a very good indication (of course lots of memtest & 3d's) . The temperatures where the same . In fact weather is a bit colder here in Greece since last week .Quote:
Originally Posted by kiwi
I'll have to agree that it sounds unreasonable to have such results in only 2-3 days ... but it's what 8 people reported so far . I'm a bit dissapointed , but i like to bench ... so i'll bench and not care for the CPU .
"relatively low voltage" ??? From 1.1V to 1.5V is 27% increase . It's not like conroes from 1.35V to 1.5V . We deal here with 45nm and it'll take some time to find the limits.
Right now 2 friends of mine (after they seen the degradation) , lowered the vcore to 1.34V . I'm waiting for their findings after 2-3 days .
The VID range does not necessarily correlate with intel's functional voltage range alluded to in the spec sheet. If intel had some comp from AMD you might see a higher stock speed and thus higher VID range.
My E6850 does 4 stable, so when i get my E8400 next week, I want 4.2 minimum, and will be happy with 1.42volts for 24/7 from what I have seen in others. If I achieve that, I will keep the benching at 1.45V and less.
If I get a bad OC'er, than like some others, I will test out the voltage limits for 24/7, as I wont care if I degrade/kill it, I will just put my E6850 back if the E8400 gets below 4 stable.
Besides if this chip degrades in even a month at voltages under the listed max of 1.45, I would a) be shocked, b) not believe it until confirmed by multiple sources c) if true, then I am happier at 4 with my tougher 65nm E6850.
well....
i think my cpu just degraded in 5 mins 1.5v-1.6v vcore...
it used to be able to do at least 5 mins prime stable at 4320mhz at 1.368v vcore (as the pic shows on the top). now for the same clock, 4320mhz, prime stops even at 1.392v...
another possibility might be i take out my watercooling and went back to air cooling after the 1.5v-1.6v abuse. temp went really high (about 75c) when full load.
so now i am not sure, whether the high temp or the high voltage degraded my cpu...
guess its time to get a new one..
Try remounting the block in case you got a bad mount . Also , 5 mins prime doesn't give any stability indication . Try it at 20-30 mins at least if possible.Quote:
Originally Posted by haoster
This might be true because i was able to do OCCT 30 mins stable with 1.4v @ 4.32ghz 2 days ago, and last night when i tired to do OCCT 30 mins, it failed. hmmm... interesting. btw, my idle temp is at 42 (stock or 1.4v) and load temp is ~70C with watercooling.
If this is true, why haven't we heard this from any of the Yorkfield users, those are 45nm as well.
Also, the guys using ES wolfdale cpu's did not complain about this.
I have had mine at 1.64V for getting sub 10 seconds 1M but do not notice any difference after that, mine wasn't very good from the start but did not turn worse.
Tested my CPU again (single stage cooled) .... 4.5Ghz with 1.49Vcore and -36C load ... orthos errors after 10 secs !
I remember leaving orthos at 4.7Ghz and 1.49Vcore 5 days ago with no problems.
I'm curious to test it with water again , as it seems i can't bench at my previous clocks & voltages.
@Zeus , a friend of mine has a QX9650 that was benching with air and 1.7Vcore . A few days ago he gave it to Hipro5 to check if it's a good CPU . When he get's it back , we'll check it if it can bench at known clocks.
EDIT : I put it again on watercooling . Loaded my previous saved profile in bios , and what a surpise ... orthos error . Now i'm 50Mhz lower at 3880 and 1.28V .
When i first tried the CPU , it was running at 4040Mhz and 1.28V . 4 days ago i could run at 3940 approx .
Hmm, I'm running my Q6600 at 1.52v, VID is 1.3v so the "overvoltage" is only 17% (and this under ss).
While I wait for my E8400, I'm wondering... if the VID is say 1.1v and I put 1.5v into it, the voltage increase would be 4/11 or 36%. That's quite a hefty whack for any component. 1.7v would be 6/11 or 54%. Perhaps as the VIDs of these CPUs gets lower and lower, we need to consider overvoltage as a percentage and maybe not get too surprised if they degrade with +40% or +50% volts?
Those seeing degradation after pushing 1.6-1.7 V is one thing, and there were some Q9650 owners that killed theirs I thought from high voltage, there is a RIP thread in here from one Q9650 owner. Given intel has lowered the max voltage on these chips to 1.45, degradation from 1.6 to 1.7 would not surprise me, regardless of cooling.
But degradation in a short time from 1.45 volts and less, is a little hard to believe, unless a lot of people report it, and do so using reliable methods such as long prime/orthos runs. Short stability tests can yield variable results that are not repeatable, so I would have a hard time making conclusions. 10-12 hour orthos/prime runs though that shows a significant change in required voltage would be different.