...
Another new feature in Qube! 6.5 is tracking "cumulative time." , and this is stored as cpu-seconds. This will show This is displayed in WranglerView and ArtistView as "CPU-minutes", and it is stored internally in CPU-seconds.
If a job instance or frame runs for 60 seconds, and has been allocated 3 slots, the "timecumulative" value is set to 180 when the instance or frame is complete/failed/killed. If the same job or instance is retried and is allocated 8 slots this time around, and runs for 30 seconds, the "cumulative time" is increased by (8 * 30) = 240 cpuCPU-seconds, for a new value of 180 + 240 = 420 cpuCPU-seconds.
We feel that cpuCPU-minutes is a much more accurate metric for "How much of the farm did I get?" that than trying to track cores or processors.
...