Is there any way on the same account to keep Einstein off my computer with a Nvidia GPU but allow it onto one with an Intel GPU? When I say no pulsar work for the account, which lately has been generating continual faults on my Nvidia GPU but runs OK on the other computer with the plain Intel GPU, then work units are stopped for both machines, even though the gravitational ones seem to work OK on both. The recent history is that I noticed about 40 some work units ago that every pulsar work unit on the Nvidia computer was faulting at the end--work unit runs to completion but then the display driver stalls, has to be restarted by Windows 10, the BOINC log indicates that the result is absent in the work unit, and Science United records another failed work unit, but the scheduler keeps sending new units. I'd rather do something more useful than generate failed work units or not use a GPU. Obviously, I can set up another account for the Nvidia computer, but is there a simpler way on one account to do this?

All my Einstein GPU tasks, both GW and GRP complete and validate on my GTX 1650 Nvidia board, now engaged in long GPUGRID tasks, only for Nvidia boards (cuda 101).
Tullio

---

It's probably the age of my board that is the original problem-it is a GT 720, and the GPUGRID task would also bomb on it every time. Since I'm not going to upgrade it on a 7-year old desktop, however, I'm looking for workarounds that make the best of the available resources.