I think https://foldingathome.org/ uses double precision.

several projects "use" DP, like Einstein and Asteroids. but it's not a large portion of the total computation time like Milkyway Separation is/was. so even though a P100 has decent FP64 performance, it's relatively low FP32 performance makes it fall behind in these hybrid FP32/FP64 situations. far more worth it to go with a card with better FP32 since that's where most of the time is spent.

Whatever is causing the validation queue (CPU speed?) needs upgraded. As I've said before, many of us are willing to throw money at the project. Just ask.

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

.No Separation tasks available

There are plenty of Separation tasks. 21K currently.

You need to script a project update to get tasks every scheduler connection.

If BOINC receives a no task reply for a scheduler connection, it immediately backs off for 1 hour, then progressively adds more backoff hours to get to 24 hours.

If you don't intercede, no wonder you aren't getting work.

On the server status page everything is running .