Project News - MilkyWay@Home

[BF]

From the MilkyWay@Home news feed:

Project News: CUDA Application Released

Wednesday, 26 August 2009 10:00 AM

The CUDA application has been released on BOINC, please see this thread for more information.  The Milkyway@Home team appreciates your patience in reaching this milestone.  This code has been implemented and tested on the GeForce GTX 285 donated by NVIDIA to the Milkyway@Home research team.

[BF]

Please take note of the hardware and driver requirements of the CUDA application!

From the MilkyWay@Home boards:

The CUDA application for 32 bit Windows has been deployed on BOINC. In terms of hardware a NVIDIA GPU supporting Compute Capability 1.3 is required. The following GPUs: GeForce GTX 295, 285, 280, 260, Tesla S1070, C1060, Quadro Plex 2200 D2, Quadro FX 5800, 4800 are known to have CUDA 1.3 support. The GPUs also need to have 256 MB of Video RAM and NVIDIA Driver 190.xx or higher needs to be installed. If these prerequisites are not met the CUDA application will not be downloaded through BOINC.

[Dataman]

The requirements are horrible.  Few, I think, will run them under these conditions. I am going to wait and see how it sorts out before running any.  ???

[kashi]

GTX 275 should support Compute Capability 1.3, although it is not listed in the nVidia documentation for some reason. CUDA-Z would probably show a GTX 275 has Compute Capability 1.3.

[Dataman]

GTX 275 should support Compute Capability 1.3, although it is not listed in the nVidia documentation for some reason. CUDA-Z would probably show a GTX 275 has Compute Capability 1.3.

I'm sure you are correct as the 260 meets them. From the subsequent posts in their thread, it looks like another app that "eats" your machine. I hope we are not entering an era where you have to give the app the entire machine to run it.  :cry2:  Ok for farms but very unfriendly to the casual cruncher who wants to run several sciences.   But as I said, I'll just see how is sorts over the next few days.

[WikiWill]

You regular Milkywayers would know, but their server is currently down for maintenance and not issuing work.  Just thought I should mention it leading into the long weekend (for those who get it  biggrin)

[Dataman]

FYI: Milkyway is now the most active BOINC project followed by Collatz, SETI, WCG, Einstein & GPUGrid. I never thought I'd see the day SETI and WCG would be knocked out of 1st & 2nd respectively.

[veebee]

It's got to be the credit rate.... at just under 4 minutes for 314 credits (4840 cr/ hr - 116K + a day !) why WOULDN'T anyone crunch there ?!?!


I have a problem though... I get work easily for my two ATI cards, yet my GTX 275 keeps getting the "no work sent" message..... is there an optimised app for cuda cards ?

[Dataman]

MW is down due to "Snowmageddon". They have a blizzard going on. I switched to GPUGrid.

[Furlozza]

And in light of those adjustments, I am now getting wus sitting in pending and also being invalidated...... fur some unknown reason.

The exception that proves the rule?: Love it when I am invalidated for wus with too many successful results LMAO

[kashi]

Might take them a while to fix this, if you don't wish to have perhaps half your completed work marked invalid then it may be time to give Collatz a go. Or DNETC if it works alright for you.

I don't know how this happened, I would have thought they would have been checking at least a small sample of the returned results against their computers at Rensselaer Polytechnic. I could understand them not catching the relatively small number that had been done by a rogue NVIDIA single precision app but missing the massive number being returned by 58xx and 5970 cards seems surprising.

[kashi]

Yep, as I predicted above it's 58xx and 5970 that are the fly in the ointment. This red wine is great stuff, it must make you psychic. 

[Furlozza]

WEG

I still reckon it's the other cards what is wrong cause ATI fixed a little bug in latest cards WEG

Anyways, am now having a run at DNETC...... 4800+ crs for 24-25 mins work ain't nowt to sneeze at.

[Dataman]

2010-12-03: MilkyWay@home - OpenCL for Nvidia available for testing
The OpenCL application for Nvidia GPUs is ready for testing for Windows and Linux x86_64. I'm particularly interested in the performance / responsiveness tradeoff on mid-low range GPUs.
Many thanks to cncguru for donating his GTX 480. If I hadn't had it, it would be about 30% slower than it is.
http://milkyway.cs.rpi.edu/milkyway/download/test/milkyway_separation_0.48_x86_64-pc-linux-gnu__cuda_opencl.tar.gz
http://milkyway.cs.rpi.edu/milkyway/download/test/milkyway_separation_0.48_windows_intelx86__cuda_opencl.zip
Extract these to the project directory. On Windows this is something like C:\ProgramData\BOINC\projects\milkyway.cs.rpi.edu_milkyway
On Ubuntu for me, this is /var/lib/boinc-client/projects/milkyway.cs.rpi.edu_milkyway

[Dataman]

2011-01-24: MilkyWay@home - bypassing server set cache limits
While we appreciate everyone wanting to crunch more MilkyWay@Home by increasing their cache limits; this is part of the reason why we've had so many server problems lately with an unresponsive validator. Mainly, our machine/database is not fast enough to keep up with the additional amount of workunits this is causing in the database. So if anyone is modifying their BOINC client to artificially increase their cache we're asking you to stop so the project will be more stable (until we can further improve our hardware). A few of the really offending clients (who have cached 1k+ workunits) are being banned as they haven't responded to us, and they're hurting everyones ability to crunch the project as a whole.

So in short, we need you guys to work with us as we're working with limited hardware that can't handle more than 500k+ workunits at a time -- our cache is low partially for this reason. Second, as we've said in a bunch of previous threads in the past, due to the nature of the science we're doing at the project we need a low cache because this really improves the quality of the work you guys report.

As you (hopefully) know by now, we search for structure and try to optimize parameters to fit that structure within the Milky Way galaxy. And lately we've been also doing N-Body simulation of the formation of those structures. What your workunits are doing is trying to find the optimal set of parameters for those N-Body simulations to end up best representing our sky survey data or to fit those different structures (like dwarf galaxy tidal streams) from that data.

To do this, we use strategies which mimic evolution. The server keeps track of a population of good potential solutions to these problems, and then generates workunits by mutating some solutions, and using others to create offspring. You guys crunch the data and return the result -- if it's a good one we insert it into the population which improves as a whole. Over time, we get very very good solutions which aren't really possible using other deterministic approaches.

If people have large caches, that means the work they're crunching can come from very old versions of those populations which have since evolved quite a bit away from where they were when the user filled up their cache. When they return the results there's a lower chance for the results to improve the population of results we're currently working with.

So that's why our cache is so low, and we'd really appreciate it if you worked with us on this. There are other great BOINC projects out there which can help fill in missing crunch time when we go down, and the BOINC client can definitely handle running more than one at a time. So it might not be too bad to explore some of the other great research going on out there. :)

Thanks again for your time and understanding,
--Travis