leeping
commented
1 year ago @j-wags This is totally on me for failing to review it after several months. Is this code still correct? Do we still want this in a future release?
Open SimonBoothroyd opened 2 years ago
leeping
commented
1 year ago @j-wags This is totally on me for failing to review it after several months. Is this code still correct? Do we still want this in a future release?
Description
This PR adds the ability for the Evaluator target to restart any failed calculations up to a specified number of times before failing. This adds some resilience in cases of rare failures due to e.g. calculations landing on a badly configured compute node.