The industry has pretty much settled on the calculation for GC efficiency to be 1 - ( total GC pause time / application run time). It's a reasonable measure of how much CPU the parallel collectors are taking away from application (otherwise known
↧
