Hi Fred,
�I've noticed a difference in how BoincTasks JS calculates the CPU % column compared to the original desktop BoincTasks.
�In the original BoincTasks, the CPU % was calculated as (CPU Time / Wall Time). If I had an LHC Theory task assigned to 3 CPUs but it only utilized ~1.1 cores (as verified in Windows Task Manager), it would correctly show about 33-38%.
�However, in BoincTasks JS, the column shows 100% regardless of actual internal VM utilization, as if it is reporting "Slot Reservation" rather than "Actual CPU Time Efficiency." Non-VM tasks (like MilkyWay) still seem to report correctly (~98%).
�Could you consider bringing back the (CPU Time / Wall Time) calculation for the CPU % column in the JS version? It helps me identify when a VM is being "sluggish" or under-utilizing its allocated cores, which is very helpful for system utilisation.
�Thanks for all your work on the JS version!
In BoincTasks Clasic The calculation is the same, but removing the check at CPU % Long time average.
The interval is much shorter.
What I will do is add and optional calculate the delta between two reads.
That way you see an immediate result
What I added >>> Readings between two intervals.
>>> cpuTime = cpuTime - cpuTimeOld;
>>> elapsedTime = elapsedTime - elapsedTimeOld;
let cpu = 0;
if (cpuTime == 0 || elapsedTime == 0)
{
resultItem.cpu = 0;
}
else
{
cpu = (cpuTime/elapsedTime) * 100;
if (cpu > 100) cpu = 100;
resultItem.cpu = cpu;
}
Old is the previous reading;