There are still lots of the newer batches available to run, but we identified some of the older batches that didn't build correct (and therefore didn't run on the grid), so I went back and built those correctly so we can get results back for them. Once these older batches are complete, we'll return to the higher batch numbers.

Seippel

Thanks for the information. Out of complete curiosity I would be interested to know how many of the Vina tasks are returned per day?

In association with automating affinity, process lasso is being mentioned on the forum. Set which cores an app is allowed to use 2-7 for instance and set your priority programmes to 0-1. Maybe the child processes follow. You can set yield levels too. In boinc that's done by tellng it to pause if the user apps use n percent. Also in boinc you can set to pause when a game is loaded, an exclusive app. Many ways to skin the cat.

Depends how you interpret 1:4, but it was meant as 1 in 4, so your calculation was right.

The fahv took a big jump in runtime, the last set of 4 for x3vq5_a_in_ledgfb are 7.5 hours running and barely at 50%, where the set before of x1sp5-b were doing 1.5-2.5 hours, on linux. Checkpointing regularly. Boinctasks is showing them as being at 51-60 checkpoints. If the packages have not changed, 140 in a task, they'll be running about 9 hours more. Should these arrive on my tablet, the runtimes will turn monstrous, and the fail chance horrendous. Have 1 in every 48 hours on there, only signal 11, the overload type, so good chance a 60+ hour job is going southwards. Looking in buffer, 12 tasks deep, all seem from earlier batches. Contemplating to pre-emptively abort those from the tablet should the runtimes persist on the pc.

I got one of those long running ones also: 15.94 hours.

Neat