Hi,

we're in transition so still a mix of the old short and new so will probably take another 3-4 days to see the full impact on the averages. The new ones on my Quad went from 0.75-1.50 per unit to between 5.5 and 7.0 hours, so in fact it looks like it's going to settle at a factor 5-6.... just keep watching the graph as it is updated every few days. All the supplemental graphs are collected in a Start Here forum post for easy perusal.

Not all PCs being equal and one better fit than the other, you'll have to let a good dozen pass thru before a stable indication of your new average will become clear.

The techs advised on multiple WU's packed in one and went for the extra heavy to stress test, but with the 12 day deadline instead of 7. Think the number was like 4 or 5 of the old. The upload result got 5x larger if that's an indication.

ciao

A bump for the special interest group that's not saying much. The run times average is now hitting > 4 hours exactly as my quad. The points per hour are still the star with an average of 76 from previously 72. Page faulting is moderate compared to the pre-recode HCC jobs and kernel time is minimal (important!!!!!).... 1:40 minutes on a 7.55 hour run time it took on the slow 1.8 C2D.

ttyl

There are hints in the performance chart that the WCG head torque-wrench operator has released the bolt somewhat and let more DDDT units flow into the system... 18 TeraFlop daily now and running 21 days 16 TeraFlops pointing up. And this is still only Phase I.

We did hear from Dr. Stan Watowich quite recenty to include that a 5th target, leishmaniasis was added to the DDDT project. For those that count the years though, a milestone was passed yesterday, accumulating 10,000 of the CPU year kind on this project. AND, the project continues to gain steam.