lopri
Elite Member
Hopped on the quad-core bandwagon thanks to recent price drop on Q6600. I should admit that it was an impulse buy since I was semi-high when I pushed the 'place order now' button, but it's already delivered to me and there is no going back. 😀 My original plan was to wait till Barcelona's debut or Intel's August price cut (Q6600 for $280?) - Although I purchased a quad-core at this time, my desktop usage just doesn't warrant the need for a quad-core. (except Virtual PC 2007, and I think the OS and HDD/VHD are more often the bottleneck than the CPU power in virtualization on my system) So yeah I don't really need a quad-core right now but I just had to buy one. I guess it's a part of the endless road to my 'Dream Machine' that we all have.
Enough of digressing, on to the topic.
Day 1:
Received the package and checked the batch # but didn't have time to install. (L640F179)
Day 2:
Installed and attemted to boot @3.60GHz (9x400) at first try. 😀 Failed miserably so tried 9x350. Again, failed. 9x333, no-go either. Gave up and reset the BIOS to default. Boot to Windows fine (2.40GHz, stock) The highest I could go was something like 2.45GHz. lol. And even that was flaky in that the system became jerky although stress tests ran fine.
It wasn't until too late that I found my board just isn't meant for FSB overclocking with a quad-core. Memory controller on the chipset seems fine in that it behaves similarly to a dual-core. (1T, 4GB, etc.) It's a known issue with the board (680i original reference board) but still disappointing. I was kinda hoping to lucked out. Called EVGA and requested a replacement for a new revision.
Until the new board arrives, it seemed like only thing I could do was undervolting. (as far as overclocking is concerned) And that's where the bright side slid in. So far, I could go as low as 1.06~1.08V (Idle: ~35C, Load: ~45C) for the stock speed. (2.40GHz)
I changed my HSF also. Kinda tired of worrying about hot CPU socket area on the motherboard, for the first time in 2 years I switched to push-down HSF from the popular L-shaped HSF (Ninja, Tuniq, Ultra-120, etc.) It's this:
I think this HSF will be quite popular in near future. I found this one by accident and decided to give it a shot. I replaced the 1200RPM stock fan with an 1800RPM Thermaltake fan, and the results are very satisfying. It might let your CPU run 2~3C higher than the No.1 HSF, but it makes me feel more comfortable because I know there is some sort of air-flow on the hottest spot of the board. Definitely No.1 HSF among this type of HSFs. (better than Big Typhoon IMO)
Day 3: Encouraged by the low voltage/temps to run Q6600 at its stock frequency, I took the fan off the HSF and ran quad-Prime95! Only moving fans are: NB fan / SB fan / Video card fan. Load temps stabilized just under 60C after 15 mins of LargeFFT run.
But at this point I should let everyone know that my setup isn't inside a case yet. :laugh: Still, the ambient air is at normal room temp. Here is the setup for the screenshots above.
Back with the 1800RPM fan, I ran some benches with Q6600 @2.40GHz (stock speed)
Day 5: Received the new board. I love EVGA. This company's customer service is the best (by far) that I've experienced. I thought BFG was decent (they are) but EVGA is even better.
So anyway after some cleaning up and moving parts (NB/SB HSF) I started testing. The first boot was 8 x 400 which I'm aiming with this chip. Everything was set to auto except disabling Spread Spectrum, etc. No voltage changes, no memory tweaking.
I will let the pics do the talk.
Needless to say I am thrilled right now.
Enough of digressing, on to the topic.
Day 1:
Received the package and checked the batch # but didn't have time to install. (L640F179)
Day 2:
Installed and attemted to boot @3.60GHz (9x400) at first try. 😀 Failed miserably so tried 9x350. Again, failed. 9x333, no-go either. Gave up and reset the BIOS to default. Boot to Windows fine (2.40GHz, stock) The highest I could go was something like 2.45GHz. lol. And even that was flaky in that the system became jerky although stress tests ran fine.
It wasn't until too late that I found my board just isn't meant for FSB overclocking with a quad-core. Memory controller on the chipset seems fine in that it behaves similarly to a dual-core. (1T, 4GB, etc.) It's a known issue with the board (680i original reference board) but still disappointing. I was kinda hoping to lucked out. Called EVGA and requested a replacement for a new revision.
Until the new board arrives, it seemed like only thing I could do was undervolting. (as far as overclocking is concerned) And that's where the bright side slid in. So far, I could go as low as 1.06~1.08V (Idle: ~35C, Load: ~45C) for the stock speed. (2.40GHz)
I changed my HSF also. Kinda tired of worrying about hot CPU socket area on the motherboard, for the first time in 2 years I switched to push-down HSF from the popular L-shaped HSF (Ninja, Tuniq, Ultra-120, etc.) It's this:
I think this HSF will be quite popular in near future. I found this one by accident and decided to give it a shot. I replaced the 1200RPM stock fan with an 1800RPM Thermaltake fan, and the results are very satisfying. It might let your CPU run 2~3C higher than the No.1 HSF, but it makes me feel more comfortable because I know there is some sort of air-flow on the hottest spot of the board. Definitely No.1 HSF among this type of HSFs. (better than Big Typhoon IMO)
Day 3: Encouraged by the low voltage/temps to run Q6600 at its stock frequency, I took the fan off the HSF and ran quad-Prime95! Only moving fans are: NB fan / SB fan / Video card fan. Load temps stabilized just under 60C after 15 mins of LargeFFT run.
But at this point I should let everyone know that my setup isn't inside a case yet. :laugh: Still, the ambient air is at normal room temp. Here is the setup for the screenshots above.
Back with the 1800RPM fan, I ran some benches with Q6600 @2.40GHz (stock speed)
Day 5: Received the new board. I love EVGA. This company's customer service is the best (by far) that I've experienced. I thought BFG was decent (they are) but EVGA is even better.
So anyway after some cleaning up and moving parts (NB/SB HSF) I started testing. The first boot was 8 x 400 which I'm aiming with this chip. Everything was set to auto except disabling Spread Spectrum, etc. No voltage changes, no memory tweaking.
I will let the pics do the talk.
Needless to say I am thrilled right now.