/srv/irclogs.ubuntu.com/2009/01/22/#ubuntu-boot.txt

jan_herehello?13:10
KeybukStyXman: so the problem is to create an efficient readahead list19:06
Keybukwe've got an inefficient one today19:06
Keybukreplacing that with a differently ineffeicient one doesn't really help things in the long-term19:06
Keybukideally we'd like to know, on average, which executables request which blocks of which files19:06
Keybukand at what time they do it19:06
Keybukthen we can page them into the cache in order, in advance of when they're needed19:06
Mirvare you planning on directly tackling some application problems, similar to this extremely great work done by Behdad: http://mces.blogspot.com/2008/12/improving-login-time-part-3.html ?19:07
StyXmanisn't too much low level? I mean, normally booting programs just read a file, maybe seq'ly, maybe randomly, and all you need is to have it already in ram, or do I miss something?19:08
KeybukMirv: eventually19:08
KeybukStyXman: the page cache works on blocks, not files19:08
Mirvyeah, I know, there is just so many even more low-hanging fruits, while the upstreams might actually fix something too19:08
StyXmanKeybuk: again, why so lo level? why not just read the file and let the kernel take care of page cache?19:09
KeybukStyXman: that's what we do today19:09
StyXmanif you're low in ram, you can make slices19:09
StyXmanKeybuk: ok19:09
Mirvhmm, I could put a mental note somewhere to check whether compiz git fixes the 5s delay (no disk activity) I've seen, it'd good for jaunty if 0.8.0 does not come out19:09
Keybukbut we read too much as a result19:09
Keybukin practice, when you execute a binary, you don't read all of it into memory19:09
Keybukyou just read certain bits19:09
StyXmanKeybuk: how is the read/size ratio?19:10
StyXmanI mean, how much do you read?19:10
KeybukStyXman: I don't have numbers on that yet19:11
StyXmanKeybuk: ok19:11
StyXmanKeybuk: do you know if what readahead does (being so low level) is a consequence of already trying that? how was the approach selected?19:12
Keybukas you can probably tell from the graph changes, my focus atm is on the kernel, udev and modprobe load times19:12
KeybukStyXman: readahead() is just a syscall19:12
Keybukman readahead gives you the docs on it19:12
StyXmanKeybuk: ah, ok19:14

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!