There's 255 negative and 255 positive

https://boinc.mundayweb.com/wiki/index.php?ti...d_with_code_255_(0xff,_-1)

Something like that, yes. An example of what could have happened was an out-of-bounds error like this one here.
If we had the source, we could've fixed it if there was an indication of what happened, but there isn't much of a clue:

<stderr_txt>
INFO:[18:00:34] Start AutoGrid...

autogrid4: Successful Completion.
INFO:[18:01:05] End AutoGrid...
INFO:[18:01:05] Start AutoDock for ZINC000100037032-ACR3.71_RX1--4mm3_001_tyr264-HO--CYS112_wcgsplit4.dpf(Job #0)...
INFO: In AutoDock main_autodock()
Beginning AutoDock...

</stderr_txt>

We have identified the issue and any new batches built should not have the problem ligands.

Thanks,
armstrdj

Try running just one workunit and see if that works. Gradually kick it up one work unit at a time. This may help you to see what is choking the machine. The other thing you might do is try Linux on that machine. Good luck.
Cheers

I had tried 80% power, but it still blew off and crashed. I suppose it makes sense to just start with one thread and work up, now that I have more time. I have never run Linux and don't need it elsewhere.

I am getting close. It will run 7 threads. It crashed after 8 threads started. I was able to change the utility in time to say "no new tasks" and suspended all of the tasks and then started them one by one.

Given that, I told the computer to use 59% of the CPU, which covers the 58.3% for 7/12 of the threads. I started the 8th thread, which correctly said "waiting", but then it crashed anyway. It was one that had started before, so was still in memory. BOINC is using 909 MB which isn't a huge amount. The computer has 16 GB of memory. The hard drive is a small at 119 GB, but it should be big enough. I will add a large solid state drive sometime. I may have to take this machine to the shop and have them do some diagnostics. I think I will abort the suspended tasks and let the queue start over. with only 7 threads allowed.

After the last reboot, one of the 7 said it had a computation error, so I will cut back to 51% covering the 50% needed for 6/12 threads. I aborted all of the suspended tasks waiting. It hasn't yet decided to add anything to the queue.

It remains happily running 6 threads out of 12. Two finished and are ready to report. One waiting in the queue said "ready to start" so that is one that hadn't had any computer power yet. A diagnostic is in its future...next month.