I think probably not. I suspect the difference is down to programming preferences - the 'user' in the "Wait:UserRequest" case being the programmer, rather than the local operator.

The Einstein app is distinctive, in needing very little CPU support, but wanting it NOW. Hence the big difference shown for thread priorities: Einstein users found that this app performs well when run at 'real time' priority. I notice a small hiccup if I'm using the machine at task end / next start, but no other side effects.

I do run two NVidia cards on the same machine, with a full CPU core supporting each: and a less computationally intensive CPU project on any core not otherwise occupied. I doubt the interaction in enough to completely stall the WCG app, but I can test that next time.

I first noticed this problem during the first beta here: since then, I've updated the drivers to the latest available from intel (version 10.18.14.5162): no change here, but I've noticed the Einstein app has become faster but less accurate. No response yet to https://einsteinathome.org/content/gamma-ray-...rch-1-gpus#comment-184328

This is a FWIW post and may not apply to people having issues here and only comes from observing my GTX 1060. The 1060 requires 1 CPU per GPU task to run efficiently. Task CPU runtime is always 98%+ of the total runtime. If you're running CPU tasks and GPU tasks at the same time and have not assigned 1 CPU per GPU task then you will probably have issues. From observing my AMD 270x the CPU runtime is less than 20% of the total runtime. Running GPU and CPU tasks is not an issue as long as each GPU task has at least .33 CPUs assisting. Of course all of this depends on hardware and YMMV. Hope this helps someone.

@Richard
Come to think of it. I downloaded a couple of months ago from Intel, the most recent version and installed it. It was something with that driver I didn't like (I do not remember what it was, but it wasn't a crunching issue since I didn't crunch on the GPU's then). So, I reverted back to my original driver 10.18.10.3907 WHQL.

I really should have written down, what it was that I didn't like with the new driver.

I don't know what VC stands for, but I have a GTX 660M in my ASUS G75VW ROG, and it bombs out in this Beta test. I haven't tried it in this latest test, since it failed on every other OPN1 GPU Beta test here. The GTX660M is probably not totally OpenCL 1.2 capable, only the later drivers are, but not the card itself.

I've given up on it, especially since the production one could expect from the 660M, is not much to write home about

Ah, too easy