I am just wondering how many of you have moved to Poem for CASP?

Maybe this thread will give the team an idea of how many CPU's they are losing.

I have moved most of of the share to Poem after giving Rosetta a few weeks as well.

---

I'm 100% with Rosetta and doing some Orbit@home WUs when they have some.

---

I suspended Rosetta this morning. The real issue for me was the t405 problem whereby the wu stopped advancing, and just "stayed" there using it's quota but doing nothing - I had cores sitting idle which could easily have been doing something else. I first noticed something was odd last night as my CPU temperatures were down, but didn't really check it. I guess some of these stuck wu's had been so for 12 hours or more. On my big quads, that is a lot of wastage.

As for POEM, I have been crunching POEM since the project started. I haven't moved Rosetta's quota, just suspended, which means that all the other projects on the machine are getting more.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

I'm doing both. ;)

pete.

---

Sticking with ~ 99% share for Rosie throughout casp...

I am committed to Rosetta for the duration.

---

Rosie, Rosie, she's our gal,
If she can't do it, no one shall!

I wouldn't of suspended if it simply meant a bit of nurse-maiding, but my machines are distributed over a geographical area. If I suspect something is stuck, it is not just the cycles and electricity wasted, it is several liters of road fuel to drive to the site and press a button.

People with setup's like that need reliable apps. I've had a few problems with Rosetta recently, although the t405 "stick" is by far the biggest worry.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

highly recommend logmein.com for remote management...

---

I've got 99% for POEM and 1% for R@H. Since CASP is about improvements through competition, it seems only fair to try to reduce the FLOPS disparity between the two projects.

In the thread here: http://boinc.fzk.de/poem/forum_thread.php?id=4

wenzel seems to suggest that Rosetta is a fragment/homology based project, but I don't believe that to be correct??? I know fragments are sometimes used, but I believe only as an option?

Anyway, I don't think there can be too many protein modelling projects, but unfortunately they're pulling largely from the same CPU pool...

---

DB answered this in this thread.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.