On Tue, Jan 08, 2013 at 12:55:27PM +0200, Alexander Motin wrote: > On 06.01.2013 18:20, Luigi Rizzo wrote: ... > > I think that for the time being it would be useful to run at least > > one set of tests with kern.timecounter.alloweddeviation=0 so we can > > tell how close we get to the required timeouts > > May be just to be sure, because it should not significantly affect > results of the 1us tests, as 5% of 1us is much less then numbers we see > there. to clarify - i don't mind if we are 50-100us (absolute error) off the requested timeout for short intervals, but i want to be sure that this error can be achieved also for large requests. cheers luigiReceived on Tue Jan 08 2013 - 10:16:57 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:33 UTC