Skip to content

ConcurrencyFlags: move setting from AvalancheScheduler to HiveWhiteBoard

Frank Winklmeier requested to merge fwinkl/Gaudi:concurrency_flags into master

Move the setting of the number of stores in ConcurrencyFlags from the scheduler to the whiteboard. This is more logical and also makes the correct number of slots available before the scheduler is initialized (e.g. in the ATLAS HLT we initialize the scheduler only after the start transition to avoid creating thread pools before forking).

Merge request reports