Originally Posted by Cotton
Hey there, thank you very much for a reply! I've been getting some help on this issue over on the p45 mobo owners club as well. The question about the plastic cover made me laugh. A lot. YES I took the big, heavy plastic cover off the cooling block prior to installation. It's a fair question; you don't know me, and these things DO happen. But not this time. Hahaha...
No animosity meant
I have seen it happen here before.
Large in place FFT using LESS THAN "auto" voltages and 433 x 7.5 is STABLE and not too hot, hovering around 65 and maxing at 69 for six hours. Increasing the mult, however, requires BIG increases to vCore, resulting in BIG jumps in max temps, and this process has yet to yield a stable OC above 4GHz.
So as far as I can tell, the issue is NOT with the FSB or Ram, both of which test stable at 433 x 7.5 with 1.18 vCore, 1.28 vNB and 1.14 vFSB. (These are down from the "auto" voltages of 1.19, 1.39, and 1.20, respectively.) So leaving the FSB, NB, and Ram completely alone now, the whole effort shifts to getting the CPU to play at 4.35 like it did before, by increasing the mult above 7.5. Simply upping the PLL to 1.52 and VCore to 1.30 is enough to at least get me into windows (which will BSOD trying to boot under 1.2125v vcore).
You may need to establish term voltages. How confident are you with your RAM clock? Have you tried a post 4GHZ clock with 5-5-5-15? Once you start getting past 4-4.3GHZ locking in overclocks take quite a bit more time investment. In the past I ran my NB at 1.34 and my FSB at 1.28 or 1.30 (1.28 was ran at the higher clocks since my board actually enjoyed a little under-volting). I would also try lowering the FSB clock and upping the CPU multiplier, especially at 4.3GHZ. But first I would try the voltage changes I mentioned. It may require a small vcore bump, infact I know it will if its bsoding under Large FFTs, and your cooler may just be suffering from what I bring up later. My E8500 was taking 1.42vcore at 4.5GHZ and that was a very good chip. I am not going to lie, getting a 24/7 4.3GHZ will take some very fine tuning.
Another suggestion would be to track down someones proven MIT template from the motherboard thread that you have been reading. This may give you a good benchmark to start locking in all of your settings. Every chip is different.
Also, when it comes right down to it, if you're trying to keep 4.3GHZ you just may be hitting the cooler's limits of stabilization. If its steadily gaining in temperature while Large FFT testing that means the cooler just is not doing the job you need done. Sometimes it happens and that is when you move to a custom loop or find a more conservative overclock, such as 4GHZ. Sh** really does start to get real for these chips past 4GHZ.
You got any personal recommendations for the best available TIM?
The TIMs that ocman suggested are perfect. If you happen to shop locally, depending on your stores, Arctic Silver 5 will be a good "3rd place" if the MX series or IC7 are not available. It will hold you over if you decide to purchase one online. The differences will be subjective, but the pastes listed by ocman are better pastes in my opinion.
Thanks for the reply. What confuses me about all this is just that I had a 4.35GHz stable OC with this chip, mobo and ram, but I was using an Asus v60 air cooler, which should be far less capable. However, the v60 created a totally stable 4.35GHz OC for THREE YEARS without a hitch. It was only two weeks ago when I changed out the cooler that I have suddenly had all these problems, including a total inability to reach the previously stable OC of 4.35GHz which, for the record, didn't take much of anything when I did it back then.
Additionally, I'm not getting a BSOD when I use large FFT and a low mult. If you re-read again, you'll notice that I am able to go as far as 433MHz with a 6x mult on UNDER-volts, six hours stable. SIX HOURS STABLE! So I'm *very* confident that I'm not running into a problem with the Ram or FSB. The only thing left is the CPU, right? Please tell me if I'm missing something here, but if I can run a Large FFT stress test with 433x6 and have no problems (with either stability OR heat) then that doesn't need any more tweaking right?
I'll drop the settings down on it just to see, but I can assure you that for 3 years that ram has run at 867 with 126.96.36.199 timings. I'll order a good TIM and reseat when it arrives. Right now I'm getting more frustrated with this than I should, mostly because like I say, I've had a stable 4.35OC for 3 years and then I "upgraded" the cooler. That's a laugh.
I have used templates on several occasions to try to short circuit some of the process for OC'ing my rig, however, I have found that the best way to approach an OC is with due diligence. If you read back through some of the stuff I've posted here and in the Mobo thrad, you'll see I've done a lot of step by stepping my way into this and the very real facts are that this cooler isn't performing anywhere near as well as my Asus v60 did, whether because it's a piece of junk or because I've messed something up I'm not yet sure, but I'd be more inclined to believe it's the latter.
I put all my BIOS settings in a post in the mobo thread, too, so if you'd like to peek at them and see what I've got going on, that'd be cool. I am once again running out of ideas and the technical skill required to solve this problem, basically at the same point I had when I first posted this thread.
EDIT: Just ordered some IC Diamond 7 from Ebay and I'll do a reseat when it arrives. Any recommendations for the meanwhile would be welcomed.Edited by RedStapler - 2/19/12 at 11:10am