It is running smoothly I think doc Travis.

He didn't use any of your code, and on top of it at the beginning Travis code was not GPL if I remember well.

That's not true; there is most definitely code from milkyway@home in there. If it wasn't GPL, and if the code didn't have a license he couldn't use it at all. He even mentions in this thread http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=576 that he's violating it.

now that they're releasing the new app, maybe it would be useful if you Gipsel release the whole source code of your old app... So that they can implement the best way of doing some calculations!

This won't happen and it won't be useful. He never gave us the source, so I had to rewrite it. I'm pretty sure there's a GPL violation here also. I am not at all happy about it. I've spent a great many hours of my life that could have been spent on something new.

At some point after I get the new ATI application ready things will start breaking for old stuff.

Is it already known if the new ATI application will have the same requirements as the old one i.e. will it run on all ATI GPUs listed in the GPU Requirements Thread?

Yes, it should.

By the way: GeForce 8800 GTS (driver 26099, CUDA version 3020, compute capability 1.0).

That GPU doesn't have doubles and won't work. It needs at least compute capability 1.3.

Thanks for the information! But why did milkyway sent me any WUs then?

If you tried manually installing this, it will try sending the workunits to it. You shouldn't get sent the Nvidia applications since you don't have doubles.

Does anyone run MW on a Radeon 6850 successfully?
just upgraded and wanted to know if the ATI 58xx fix also comply to this family of cards, as i cant run the Gpu vers of MW (just SSE2)

Collatz is running smoothly though but perhaps they use other techs to get things running!?

From the message boards I understand that this application might also be ported to the GPUs later, but for the time being, wouldn't it be a good idea to add checkboxes to the account / preferences where one could deselect the 'old' application when selecting the nbody app?

Otherwise, if 'use CPU' is checked, I guess, one would also get the cpu applicaton for the other Milkyway@home application. A mere waste of cpu cycles, given the outstanding performance of the ATI cards...

Regards

Sorry, I am scrambling to fix some (non-BOINC) code right now; it's putting me a bit behind. I am planning on finishing the write up this week, however. Thank you for your patience!

-Matthew

I'm seeing lots of errors on my new 5970. Thought it might have been something to do with using an old optimised app so I renamed my app_info file and let BOINC download the latest app and there is still errors.

http://milkyway.cs.rpi.edu/milkyway/results.php?hostid=154949&offset=0&show_names=0&state=4

It's not consistent on which GPU the errors are occurring. Not seeing this on my 4870 in the same machine. I increased the fan speed a bit (man it's noisy) with no change (temp did drop a little). Increased the frequency from stock with no change. Any ideas?

Any news yet?
When comes the new application MW3 into production?
My fermi-cards and i are still waiting???

Ton (ftpd) Netherlands

ATI's HD 3850, 3870, 4850, 4870, 4890, 5850, 5870.

The 5830 does Double Precision as is listed Here. So technically It should be able to do MW and other projects that need DP as It does have 1120 Stream Processors.

What we have in the works is running n-body simulations. Essentially we'll simulate a bunch of stars, with different masses and velocities, and simulate their movement around a simulated milky way. We'll be trying to find what initial parameters generate a galaxy that looks like our observed data of the Milky Way. This should generate some really interesting results :)

Anyways, at least until we (or someone else) come out with a super fast GPU version of that, it will be CPU only, giving the CPUs something new and interesting to crunch.

Depending on what you are going to do exactly, but n-body simulations with a lot of stars scale acceptable on GPUs. Maybe not that extremely well like the current code, but still. The downside is that it will be a harder port to GPUs. Maybe one should consider OpenCL (quite similar to CUDA) for it to have a single application for nvidia and ATI (and I think there is even an OpenCL compiler for Cell).

I'm not seeing very many errors on my end. Also, the number of invalids the server is seeing is down to < 2% (including results that just error out). Swapping to the new application should drop this even farther.