On Wed, 2003-09-10 at 12:43, Jim Sibley wrote: > I have seen the improvement under VM with the clocking > stopped. I don't anyone here who has looked at the > LPAR case with the clocking stopped; its always been > on and no-one has complained, even the performance > people who buts large loads on the OSA cards. Its > doesn't seem to be a big deal under LPAR in that all > the attendant VM wakeup and paging for each EC is not > necessary.
More than that--you're not even going to notice the overhead of a 100Hz timer interrupt until you have several dozen machines running. In the LPAR case there's currently no way to get that density of machines-popping-timers on the box in the first place. Adam
