On Tue, 2002-06-04 at 00:16, Christian Jaeger wrote: > > 2.) I am now running 2.4.18 with > http://penguinppc.org/~eb/files/2.4.18-elf-fix.patch and > 2.4.18-rc1-low-latency.patch from Andrew Morton's Website > (zipworld.com.au something) (and also the attached one). It runs, > also mol is no problem; however: I've compiled it with the sysctl in > /proc, and tried switching with > echo 1 > /proc/sys/kernel/lowlatency > or echo 0 > ... > but haven't noticed any difference whatsoever yet. I've mainly tried > untarring kernel tarballs and copying around with or without nice and > at the same time dragged a window around in X. The window is sluggish > either way during phases where it seems the CPU is working and not > the disk. Now I would have expected that since X runs with elevated > priority (-10) it should profit from lower latency.
Beware that the window manager (and possibly other clients when their windows get exposed) is also involved when moving windows. > I guess lowlat only helps for realtime priority processes? Maybe, or I guess it rather decreases worst-case latencies but doesn't necessarily help in the average case. > Maybe the O(1) schedular would help more in that case? I doubt it, I understand it mostly makes a difference with lots of processes running on multiple CPUs. But why not try? :) -- Earthling Michel D�nzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
