14 Mar
2007
14 Mar
'07
3:31 p.m.
On Wed, 2007-03-14 at 09:34 -0500, Jeremy Fowler wrote:
However, I have to correct you on your statement. Preemptable kernel configurations are hardly a week step when you consider a twenty fold increase in process-level responses in some cases.
I don't dispute that the preemptable kernel is valuable -- especially for cases where you are doing audio or video capture -- but surely we can agree that the different between a 2ms wake-up latency and a 0.1ms wake-up latency is not going to be noticed by the end user in the vast majority of user-space applications?