You bet I'm trying to overclock it! The good news is it fired up without any hiccups.
Windows 7 boot time is 8.5 seconds, from main screen turn on (for POST) to desktop. I have a OCZ Vertex 3 120GB SSD in here. The drive started out a bit flaky, it would freeze up every couple of weeks and required a cold boot to come back online, but it has stopped doing that entirely the last couple months. I couldn't believe it the first time I saw the loading animation end right when the colored balls were passing each other the FIRST time.
There's a nice thick thread over on another forum about this board and the consensus there is that overclocking is more or less limited by mosfet heat output: The chips are not sinked at all. This issue only tends to rear its head when running synthetic torture programs.
I'm noticing a few extra options in the BIOS menu than some of the earlier users did: Yes I still do not have an option to modify the vcore setting, but there is a setting that adjusts a voltage offset under Turbo. I do intend to keep power-saving features enabled if possible.
The southbridge heatsink remains at the high-sixties Celsius even when prime95 is running and never gets hotter than that, so I guess I'm not going to worry about it now. I might not even bother to install the heatsink I got for it at this point...
At 40x multiplier, the voltage goes up to between 1.208 and 1.248V. These voltage values are higher than they were when I left it at stock settings. Under Prime95, CPU-Z shows the machine throttling down to 33x intermittently after about a minute, as I expected it might. I'm going to see if I can't come up with a way to cool some of these mosfets...
If there is some way I can get it to provide about 1.3V when it is ramped up to 44x multiplier, stays stable, and does not throttle down under typical loads, I would be very very happy considering the enclosure is a Silverstone SG06. I was able to devise a method to mount a Corsair H80 cooler in here, while still allowing the slim optical drive AND a 10 inch videocard to fit.
Update: I am unable to tell if changing the "Turbo Boost Voltage" setting in the BIOS is actually doing anything. I set it to +0.180V but I have no idea if that is supposed to be the offset from the stock voltage, or what. I was able to get a reading of 94 degrees C with my infrared thermometer of one of the mosfet chips (!) as I ran prime95 at 4.2Ghz... CPU-Z reported the voltage hovering between 1.280 to 1.304V before it throttled down to 3.3. I'm now covering up the row of mosfets near the socket with mini heatsinks, and also placing a 70mm fan there temporarily. Hopefully it will stop throttling then.
Update_2: The active cooling prevents throttling! It's been running prime95 at 4.2Ghz for about 10 minutes now! I'm gonna see if I can't get a custom heatsink built with the perfect dimensions to solve this heat issue. The problem is it's quite a bit of heat and the row of mosfets is only 7mm wide. A heatsink has to be either only 7mm wide (which will require active cooling) or has to be specially shaped in order to fit with the other components present (a row of chokes).
The voltage setting does seem to work. I'm now testing at 4.4Ghz, I increased the setting some more and CPU-Z reports 1.328 to 1.344V! Super PI 1M time is 8.533 sec. Getting close to 10 minutes in with Prime95, it still has not throttled. The 70mm fan I am using came from an Athlon 2800+ stock cooler, It is a jet turbine at 12V and I have it on a regulator... I will need a proper replacement. Looking at this 40mm fan
, I believe it is a Scythe Ultra Kaze Mini.
I am not sure if it will get the job done though. My little aluminum heatsinks appear to be close to 70 degrees even with good airflow.
In case anybody's wondering, (4400Mhz, 1.33V or so, Prime95 Torture Blend load) the CPU core is 60C, with the H80 fans on the silent low setting. This is at 25 minutes in.
And you thought this board wouldn't OC well.
Update 3: I'm leaving it at 4.5Ghz, I'm going to keep it under 1.35V. Got some nice freezes and BSODs, so I only have an "average" chip after all. Still, very good results from the tiny motherboard that could.