Is there a 'queue' concept in proof?

Dear Proof expert:

When several users run proof at the same time, it would be very slow and sometimes proof could crash. I am wondering if there is a queue concept in proof so that the whole system will not get overloaded?



Something experimental exists.


xpd.schedparam default mxsess:10

should allow at most 10 sessions (i.e. TProof::Open) at a given time on the cluster. Additional attempts would just fail.


xpd.schedparam default mxrun:10 queue:fifo

should allow at most 10 concurrent sessions running (i.e. no limit on TProof::Open but at most 10 sessions processing); additional process requests are queued and processed asynchronously as soon as the running ones are done.

Let me know if you try.

G. Ganis