New BOINC project: QMC@Home!

[kb3edk]

There is a new project for all you BOINC fans out there - Quantum Monte Carlo @ Home, run out of the University of Muenster in Germany, has just begun today. The project aim is to simulate quantum molecular interactions for theoretical use in chemistry and life sciences. It appears to be a "canonical" BOINC project since a link to it appears right on the BOINC home page.

Our good friend and TeAmmate ingleside has already formed our TeAm's branch there and I have just joined myself (though at a limited level for now, as I am mostly dedicated to TAS these days). Here is a useful link if you would like to join us -

http://qah.uni-muenster.de/create_account_form.php?teamid=7

Happy crunching!

 

[petrusbroder]

This will get to the poll for TAS next month!
Thanks kb3edk!

 

[TAandy]

Don't have the resources unfortunately

 

[BlackMountainCow]

Hehe, the second project from Germany after SIMAP. Looks like we're catching up!

 

[Spacehead]

/huddles in corner, muttering to self... "no, no, not another one..."

from the QMC homepage:
Beta test phase starts 4/2006.

 

[Fardringle]

This sounds like an interesting project. Another DNA/Protein/molecule study project but using "Quantum Chemistry" this time. I joined up on my test machine just to check it out.

If you do join this project, be aware that at least so far for me, the work units take a LONG time to complete. When my client first pulled 3 WU's from the project it estimated 10 hours 16 minutes for completion of each unit (this system is a Sempron 64 running at 2.0GHz), but so far it has been working on the first WU for 10.5 hours and says that it still has 7 hours to go. Long work units aren't necessarily a bad thing, but the report deadline is only three days from when I first enabled the project so this definitely wouldn't be a good secondary project that only gets occassional CPU time.

 

[TAandy]

Originally posted by: Spacehead
/huddles in corner, muttering to self... "no, no, not another one..."




LMAO

 

[kb3edk]

Originally posted by: Fardringle
If you do join this project, be aware that at least so far for me, the work units take a LONG time to complete. When my client first pulled 3 WU's from the project it estimated 10 hours 16 minutes for completion of each unit (this system is a Sempron 64 running at 2.0GHz), but so far it has been working on the first WU for 10.5 hours and says that it still has 7 hours to go.

Yeah Fardringle, I've seen what you're observing, I had one unit get "stuck" on 57% for half an hour... then instantly it finished! And it wasn't a failed WU, I got credit. I guess weird behavior like this is par for the course when running alpha-level projects...

 

[Fardringle]

Well, I'm more than halfway through my second QMC work unit (I'm going to let this computer run the three it downloaded). The first one took just under 21.5 hours to complete and gave me 373.30 points (nice PPD rating there ). The second one is looking like it will take 23 hours to finish. The third one (that hasn't started yet) is showing an estimated processing time of 20.5 hours. I actually kind of like the way these work units run (long run times means less clutter in the "Work" window, but they're definitely taking too much time away from Rosetta so I'll probably put the project on suspend for a while after these three units finish. I'll probably let more run in a week or two but not until I manage to break my goal of 5K points per day in Rosetta...