You made a mistake. *Warning graphic*
You made a mistake. *Warning graphic*
Feeling nervous about giving it another shot with my working X5660 after completing the butcher job, especially since I just found out it's a soldered die.
I received it at no cost. Purchased it online at a discount but didn’t get the correct fit, so I sent it back. The box arrived a month later after multiple failed delivery attempts, and I located the replacement parts on Cooler Master in the Netherlands.
I recall my initial attempt with the de-lid. Thinking it was soldered, I used a soldering torch on the chip. It turned out to be a Phenom I Agena core 9850BE. I was really shaken like a tree in bad weather. After a few more tries, that fear faded away. The thread linked shows DinerCore removing the IHS without heat, almost identical to the de-lid tool but done in a vice. Personally, I prefer using heat. Both methods are effective. If you ever need help or advice, let me know.
The optimal method for running a bare chip is to skip the IHS plate and go straight to the end. This involves removing the hold down bracket on Intel boards. It may not be worth the effort for everyday use. For casual practice, it's enjoyable and provides a good challenge.
The cooling system isn't working well if you can't manage these chips. The Gammax 400 I talked about earlier with two NF-F12s tied on? It kept the i7-950 (45nm chip) at 1.43-1.45v around 86-87°C. You're applying much less voltage to a 32nm chip, which should perform better. The issue with clocks not stabilizing comes from only providing about 1.3v, which isn't enough. These chips can handle up to 1.4v or even 1.45v if you manage the heat better. I've been out of X58 for a while (will return after finishing an X79 build), but your voltage levels seem quite low. However, I usually pushed things hard for daily use, so I kept voltages at their maximum safe settings most of the time.
But increasing the voltage only made things worse, and lowering it from 1.29 also caused instability. Of course, the readings I see aren’t exactly 1.29375—they’re significantly higher. On my 2550k, the voltage set in the BIOS isn’t reflected in CPU-Z or the hardware monitor info. The same happens with this x5660; at 1.29375 in BIOS it appears on CPU-Z and HWMonitor during stress tests as 1.344v.
I found this thread on OCN about AMD CPU overclocking. It’s a few years old but the content is still relevant. Check it out: https://www.overclock.net/forum/10-amd-c...hread.html
Oof there are some other settings that might be off somewhere. X58 is a tricky balancing act—adjusting one thing usually means tweaking several others. In your situation, higher voltage probably leads to instability since your cooling system isn’t performing well enough, which will push temperatures up. Cutting the voltage further would hurt an already overclocked chip. The best approach seems to be checking out the HEDT forum (link in my profile); there are many X58 enthusiasts there, and some top overclockers also use P6Ts. I’ve tried a Rampage III Formula about two years back, used an MSI board, and added EVGAs afterward. I don’t have firsthand experience with this specific board, so I can’t say for sure what BIOS settings are available.