Folding Forum

This is the first time I've seen this happen, but maybe it's more common than I think.

https://apps.foldingathome.org/wu#proje ... e=12&gen=0

Summary: I got assigned a WU that was returned OK over three days ago.

Was the protein returned to the Work Server or the Collection Server?

For that project, it's not particularly important. P13415 is a test project designed to uncover crashes in the (relatively new) FAHCore_22. In fact, the error reports are more important than the completed WU. Those error reports are examined very carefully and lead to corrections that are being incorporated into the next revision. FAHCore_22 v0.0.10 has been released to the genearl public, but V0.0.11 will soon be distributed and I expect that V0.0.12 will soon follow it. P13415 will continue to be distributed since those WUs are parituclarly challenging. You'll also notice that it's not taking very long to return the error reports.

Well, that explains it...

Any idea why neither my Vega 64/HD 7970 have failed a single one of these, then? Can I call AMD and tell them their drivers are superior?

Should I do something on my end to force them to crash so I can get a 'real' WU more quickly?

I don't think so. It's a matter of (A) knowing how to fix the FAHCore bugs associated with ATI or (B) excluding assignments to those GPUs.

bruce wrote:I don't think so. It's a matter of (A) knowing how to fix the FAHCore bugs associated with ATI or (B) excluding assignments to those GPUs.

I think you read me wrong. My cards have a 100% success rate with 13415.

Or maybe I'm reading you wrong. I don't know anymore. It's been a long day in front of a screen.

Right. It was easy to fix the bugs you had quickly. NV's were more difficult or more intermittent.