Ok, i recently upgraded the PC from a C2D to this i7. I decided I wasn't going to OC until i was finished moving for a while. So now that I am, i went to OC it and my temps are waaaay too high, at least according to the C2D/Q temp thread. I'm thinking that maybe possibly during the move my water block became a tad dislodged from the CPU so that it isn't mounted correctly.
Basically, is remounting the heat sink the first thing i should try? i don't have any artic silver on me so i have to order it, and wait a week for it to come in, etc. so if there's any suggestions i could try before that it would be awesome.
I'll try to get these settings in the post, if i'm missing something let me know.
thermochill pa 120.3 radiator, with 6 fans in push-pull
dtek fuzion v2 water block
swiftech MCP655 pump
Now when i had my 4ghz OC (200x20) my idle temps where around 38/38/41/41. when i went into Prime95 and did a stress test, i saw the temps jump to 80+ and i panicked and tried to shut the test down in time, but the BSOD happened first.
I tried enabling the "Dummy OC" that my x58 board offers, which put my cpu around 3.5 Ghz, and the idle temps are the same, but the load temps are shooting into the 70's as soon as i start the stress test, i can stop the test before BSOD though.
Now completely stock the idle temps seems about the same still (it's weird, that normal?) and the load temps are about 63/63/65/65
Now it's to my understanding that idle temps are probably too high, and that load temps are definitely too high. It could be that I'm not fully grasping all these voltages that seem to be new to the i7 (QPI, etc etc) and I'm setting something too high without realizing it, but I've read a few OC guides and they seem to be within what other people choose.
So i have the artic silver on order so i will eventually try to do a remount, but if anyone has any other suggestions I'd really like to hear em. thanks for your time.
i7s run a lot hotter than the intel quad cores. He is overclocking it heavily, so he is never going to see load temps in the 50s. He is not that far off, but if he makes some adjustments to his cooling rig he can get it down a few degrees. And remember, it is unlikely real world temps will ever reach as high as they will when doing an extended stress test....that is the point of the test, to give your cpu the worst case scenario.
voltage depends on your board and vdroop. Also you can get a bum chip that just needs a lot to get there. The D stepping is better in this regard, but he has a 940 which to my knowledge never changed to the D stepping. He would actually be better off with your 920...but too late for build advise.
I've tried lower voltages but I get BSOD's or it won't boot up. And my Vcore is only .0875 higher than yours (Cooperstown) but I'm at 4Ghz
DRAM volt has to stay where it is, it's the rams recommended voltage. And i've tried lowering the CPU VTT voltage to lower than 250, but i get BSOD's. Any other voltage i can lower to reduce temps would be appreciated.
I'm also using arctic silver 5, and i just heard that OCZ put out that new thermal compound that's superior, so i'm going to buy one of those and try it out when it get's to my house.
Also i haven't passes the AS5 burn in period so temps may drop another 2c over the next week or so.
Also i didn't use much paste. But i did apply it in a line according to another thread i saw on Tom's that showed a few people getting much lower temps.
Edit: I also looked at my cooling system. Water flow seems to be normal, i mean quite a bit is moving around in the reservoir which is the final spot in the loop, so i would assume the radiator and CPU block are getting good flow. The only thing that i could do really is clean off the radiator, which i have been meaning to do for a while, *but* when i first built the PC, i couldn't find ANY screws that would go through the fan and into the radiator hole correctly. either the screw was too wide, or too long and it wouldn't hold the fan against it, etc etc. So out of desperation i ended up super gluing all 6 fans to the radiator lol. I know probably not the best thing to do, but it's been a total champ. I've had that reservoir hooked up to the PC for a good year or 2 now and it hasn't budged. I'm using that Swiftech mounting system so the whole radiator is being held to the PC by one super-glued 120mm fan lol.
I would for sure check into cleaning the rad, and if you have any way to increase fan speed and/or air flow I would do that also. Most of my water builds using the 920s on water @ 4.0 run in the mid 60s during stress test. But once again, that is NOT a bad temp at all for the i7, and you'll see similar temps at stock speed with the stock cooler.
When you re-seated your water block, did it seem to be making full contact with the proc? ie. from observing the thermal compound on the block and the proc.
Other than that, your vcore still seems a bit high. Most people seem to be able to get 4.0 somewhere in 1.21-1.23, and you will notice temp differences at those lower voltages. Are you using a quality power supply?
Edit: As far as vcore and temperature, IIRC I was running at about 3.5GHz with 1.35 vcore at one point, and dropped my vcore to 1.2 and saw about 8C lower temps.
Your temps look great for such a high OC. Even at stock (no OC and stock fan), the normal idle temps for me were in the 40s. It seems the i7s simply run hotter than what we're accustomed to with previous processors. I"m air cooled, and prime stressed it goes into the 80s easily. You reported 66/67/70/71 stressed? You're fine.
your temps are great lol, I'm running a D0 and its sitting at 4Ghz (200*20) and has temps of ~75 C across 4 cores under Linx reported by realtemp. I'm at 1.33V as well (I got a more leakage chip it seems, it's giving out a hell lot of heat for the voltage..)
BTW I have a TRUE with push/pull, one stock fan and the other is a delta/panaflow 120mm*38mm 100 CFM fan that kicks in under load (and boy is it loud!!)