At 04:07 AM 11/17/2008, Daryl wrote: >Hi ! > >I'm trying to get a 1000FPS stable server, with pain.. I've followed many >recommendations on this thread like : >- RT patched kernel (from Ingo Molnar)
RT helps out with scheduler latency.. >- with and without HRT Support Required at least. >- Tickrate 1000 and pingboost 2 pingboost uses select() for timing of frames. the problem is that select uses jiffies resolution, not hrtimers. another thing is, FPS is measured by a couple of syscalls, namely gettimeofday() and nanosleep(). There's also a bug in the timespec_to_jiffies code where it automatically adds 3 jiffies to prevent timers firing off on all CPUs. >I've posted my configuration here : http://pastebin.com/f7c4dff1e > >here's the result : around 950 FPS Unstable. If I try Tickrate 10000 I get >1000FPS some times... but not stable and with an accelerated game :( Never going to get it all the time, no matter what. Only way to do is is to make gettimeofday coarse based. That will round the usec field to the last stored value, IIRC. G. "Monk" Stanley <gary at summit-servers dot com | gary at DragonflyBSD dot org> http://leaf.dragonflybsd.org/~gary "There currently are 7 different ways to get time from a computer. All of them can't agree on how long a second is supposed to be" -Me _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
