hmmm, can I pass you my soft autho key... then you can register them under my userID :P

300 cores, nahalem will add 20% to the current daily CEP2 output... WELCOME, see you already had silver in no-time. Need 12 more days for gold, real days as the stock hardware is not efficient when all 4 cores run CEP2 concurrent, so running a mix with the very light HCMD2 and the somewhat heavier HFCC.

Crunch On and welcome to the club, you may check out from, but can never leave.

Hi Jim 1348,

Simple, select these sciences or any other of choice in either the My Grid > My Projects or in the My Grid > Device Manager > Device Profile > Profile. WCG more or less provides a quasi mix of tasks, so when your client asks for one, the feeder gives one from Science X, then next time it gives something from Science Y and so on. With a multicore randomly there's seemingly always a CEP2 job or several running, and HFCC/HCMD2 on the other cores, but rarely 4 CEP2 jobs at the same time without need for manual intervention. When 2 run, the difference between Elapsed and CPU time is at least here quite a bit less than running 4 concurrent. It's not only with CEP2. Running mixes in general is more efficient because somehow these applications are less likely to compete for the same CPU feature or other resources. I know one project, not WCG, where performance actually goes backwards. Run 2 and it about manages to do them in say 30 minutes. Run 4 concurrent and suddenly 40 minutes are needed i.e. at the end of the day less is done than running 2 concurrent. The all compete for the same specific integer functions in the CPU... major bottlenecking. It's not like a quad has 4 of everything. With Hyperthreading that effect is even more pronounced. 1 Core each servicing 2 science threads.

I learn something every day from you, Sekerob!

Is there any advantage to running a mix of two projects when both are very heavy (CEP2 and Dengue), rather than all one or the other? Maybe athough both are very demanding, they demand somewhat different things?

Can't say DDDT2 is particularly heavy in io. It runs nicely side by side with CEP2 on the stock Q6600 quad and shows opposed to CEP2 negligible loss between Elapsed (wallclock) and CPU time. BOINCtasks logs remote 99+% of efficiency, which CEP2 rarely achieves. Still reconfirm, Bluetooth and WIFI somehow takes allot of time away from crunching as does FireFox. It was up for 48 hours and ate 750MB of memory... it *leaks*, still, so as noted now use latest Opera on Linux.

... but this thread was about the production progress... bet you that we'll have passed the 250 CPU years long before the Woz/Mac versions "comes alive" (Frampton :-)

18% already, jeez. based on current rate if this project hits windows, then this can be done within a month.