Tweaking the TWKR: LN2 Review - PAGE 2

 Obviously, we ran SuperPI, one of the most widely used benchmarks. Our objective was to go below the 11s mark, since when using dry ice I managed to do 11.234s. We worked hard to reach our goal, but I know we could have made even better. The first time I tested the TWKR, I had tweaked my memory frequency and timings beforehand, but this time I did not. I know I could have made better if I did. Oh well, that will be for next time!

To reach 10.998s, we ran the integrated memory controller at a not too shabby 4.8GHz; this is a 140% overclock! To get it to run that high, we maxed out the CPU-NB multiplier in the BIOS as well as increasing the HT link reference clock. But most importantly, we also had to max out its voltage! 1.8V on the IMC is an extra 0.7V compared to standard Phenom IIs. Fortunately below the -100C mark, the TWKR sustained this voltage without any glitch. We did not increase the cores as much though; 1.725V was enough to get a suicide screenshot at 6405Mhz.

Finally, we decided to run 3DMark Vantage to find roughly up to what frequency it would be benchmark stable.

So that is all we achieved with only 3 hours worth of liquid nitrogen. Overall, this was an awesome experience for a first time with this coolant. I am looking forward to next time, and I am sure Pier-Luc is, too. Will we still be using the TWKR? Maybe, maybe not. This piece of collection is now a few months old, and God knows on what AMD is working on at the moment. The TWKR could have it's title of great extreme overclocker taken away.

Also, no hardware was killed in the process; at the end, we let the MA790FXT-UD5P thaw and dry for a day, and luckily, it's still going strong!