Quote:
Originally Posted by Heliocon
The way around this would be to break up a specific task or pipeline and run it on more than one thread, instead of just assigning other threads minor tasks. (for example in HOI3 FTM the performance boost was no where near what it should of been because it maxes out a thread which requires all other threads to wait on it almost like a single thread).
|
That is interesting, if one of the threads is making the other three wait, that might explain why cores 2, 3 & 4 always seem to show the same (equal) level of lower usage while core 1 runs at maximum.