Run a compile job, or a video encode... then try web browsing.
Perhaps you could time clicks to open new tabs in your browser. Or switch virtual spaces.
With the background work, you should notice no or very little difference. Since the interactive work should get highest priority(this is what you are aiming for).
You can try these tests without bfs patches too. You can see how the default kernel/X11/driver is not optimal - and how things can be improved.
It's still much nicer than OSX, and windows however.
Yeah, I've done the standard "nice make -j 4 bzImage" and noticed no detectable slowdown in normal browsing tasks, window movement, chromium/flashplayer etc. It does feel snappier as CK was going for but since I know I'm running bfs I'm somewhat afraid of confirmation bias.
Perhaps you could time clicks to open new tabs in your browser. Or switch virtual spaces.
With the background work, you should notice no or very little difference. Since the interactive work should get highest priority(this is what you are aiming for).
You can try these tests without bfs patches too. You can see how the default kernel/X11/driver is not optimal - and how things can be improved.
It's still much nicer than OSX, and windows however.