[-Adrian]

4x AMD Opteron 8439SE - 2.80 GHz (24 cores) - OC to 3.1 GHz: 3m02s / 3171.85

Beast!

The time/benchmark discrepancy you mentioned shuffles the order at the top a bit, but i won't worry about that too much. In general, benchmark seems to be the preferable sort column, because if a node barely misses the target SL in an iteration (e.g. 14.92 of 15.00), it'll do considerable extra time, while the score will already have determined calculations/time rather accurately

[hal07]

I'm having some difficulties understanding why your rendering test is default at low priority. I think this cannot reflect a true CPU test, as this may vary as Windows like to run its scheduled tasks when no processes are running in normal (or higher) priority. This also do not reflect a real working situation, as one might start working at another program whilst maxwell render at low priority. If you shut down all other apps and then start maxwell, Windows will then detect the operating system as free for its own maintenance tasks, as no applications are set to normal or higher.

Especially Windows Vista and 7 have more tasks involved, and antivirus and other services may be installed to detect such "hey, no processes are running, lets do our maintenance!".

[dmeyer]

Adrian - wondering if there are any updates to the benchwell site in the works?

Willing to help out if you've lost time/energy....