WTF! We processed them...we should be given credit for them. Absolute horrible way to shut down an app!

It's just credit, it's meaningless, it's not like they stole money from you for goodness sake.

Will all submitted tasks be validated, and points given for work completed, or will any unvalidtaed tasks just be left in limbo?

Hate to say it, but a lot of users will not stay/come back to just run CPU on this project as they are used to what their GPU can do. So expect a large drop off of users after.

However, maybe spend sometime optimizing the N-Body GPU app so that maybe they will stay/come back? Especially if you can find a way to utilize NVDIA CUDA more to bring those users in?

You're right, we expect a large drop-off in the number of users.

During our benchmark tests, other GPU N-body codes performed similarly to our GPU code. I don't think that it is an issue with optimization of the GPU code implementation, I think that there is just too much overhead with building the particle tree on the GPU that there is no substantial speed-up compared to running on CPU. It's possible that there are improvements that could be made, but it isn't really a feasible thing to support with our current group and infrastructure.

is the GPU n-body code on your github page? we have a very competent GPU developer on our team who is great at GPU optimization. we've produced optimized GPU applications for several other projects.

if it's not on your github, would you be willing to share what code you have already so that we can take a look?

*Petri batsignal shines in the distance*

I've been running a RX 580 and a HD 7970 GHz edition. I recently bought a couple of second-hand GPUs to muck around with - a Quadro K6000 and a RX 5700XT.

The 7970 is a star, it was doing a task every 45-48 secs, slightly overclocked at 1125MHz. The 580 does a unit in about 100 sec at stock. The 5700 XT was a disappointment, slower than both, and drawing 90%+ load, so I'm not going to try running multiple tasks.

The Quadro - basically the same as a Titan Black, slightly down-clocked - takes 250 sec per task using about 15% GPU load. I've now got it happily running 6 tasks at a time, using 85-90% GPU in the same amount of time, so in the same league as the 7970. Alas, I thought it'd be better. I could probably get 7 units without a significant slowdown, but I'll stick with it as is. Unfortunately I can't increase the core speed, it seems locked at 900 MHz, though I can speed up the vram.

I've changed the 7970 to processing 2 units, and it is now doing 2 in 70-ish seconds, instead of 1 in 48.

<cc_config>
  <options>
    <start_delay>30</start_delay>
    <report_results_immediately>1</report_results_immediately>
    <max_file_xfers>20</max_file_xfers>
    <max_file_xfers_per_project>20</max_file_xfers_per_project>
    <use_all_gpus>1</use_all_gpus>
    <allow_multiple_clients>1</allow_multiple_clients>
    <mw_low_water_pct>1</mw_low_water_pct>
    <mw_high_water_pct>16</mw_high_water_pct>
    <mw_wait_interval>256</mw_wait_interval>
  </options>
</cc_config>