- Joined
- Jul 24, 2002
- Location
- Seattle, WA
I've been working on overclocking the rig in my sig for the last week or so, and I'm running into some very disappointing results. I was wondering if anyone could help me out.
First, I wasn't able to include my chipset waterblock this time around, so now I'm using the Swiftech MCX159CU. I know it's making good contact, so I'm not worried about that. The other watercooling components are two Chevette heatercores in parallel with two Sanyo Denkis each in push/pull config @ 5v, an MCP655, MCW6002, and DD Maze4 GPU. Maybe that's OT, but I'm pretty sure my coooling is more than adequate. Even at 1.5v I haven't seen CPU temps over 36C or so, so I know that's not the problem.
The CPU just seems like it's requiring WAY more voltage than it should. At 3.9 I have to bump it up to 1.45/1.47v to get some stability, and even then it crashed one of last night's Prime95 instances. The same happens at 4GHz, but I have to pump it up to 1.51v for similar stability. I have done the vdroop mod, and the reported voltage stays very consistent. I've never seen more than a .02 variation from it. My rails simply don't budge on my Seasonic 600w.
I know it's not the memory...this is good TCCD, and it Primes away quite happily at 260x14 when I lower the multiplier...so it's got to be the CPU, right? My PWM temps no longer exceed 43C now that I added a fan over the heatsinks that were already in place.
Any ideas? Can I run more voltage through this? How much would be safe? If it's just a crappy chip, I feel pretty shafted by the classifieds user who sold it to me...
First, I wasn't able to include my chipset waterblock this time around, so now I'm using the Swiftech MCX159CU. I know it's making good contact, so I'm not worried about that. The other watercooling components are two Chevette heatercores in parallel with two Sanyo Denkis each in push/pull config @ 5v, an MCP655, MCW6002, and DD Maze4 GPU. Maybe that's OT, but I'm pretty sure my coooling is more than adequate. Even at 1.5v I haven't seen CPU temps over 36C or so, so I know that's not the problem.
The CPU just seems like it's requiring WAY more voltage than it should. At 3.9 I have to bump it up to 1.45/1.47v to get some stability, and even then it crashed one of last night's Prime95 instances. The same happens at 4GHz, but I have to pump it up to 1.51v for similar stability. I have done the vdroop mod, and the reported voltage stays very consistent. I've never seen more than a .02 variation from it. My rails simply don't budge on my Seasonic 600w.
I know it's not the memory...this is good TCCD, and it Primes away quite happily at 260x14 when I lower the multiplier...so it's got to be the CPU, right? My PWM temps no longer exceed 43C now that I added a fan over the heatsinks that were already in place.
Any ideas? Can I run more voltage through this? How much would be safe? If it's just a crappy chip, I feel pretty shafted by the classifieds user who sold it to me...