Encouraged by Brendan Gregg, I tried to see how much time Firefox
spends on the CPU during the first 20s of startup, as well as how much
IO goes on during that time.
cpu.d gives me 1555 samples which I assume means ~1.5s on the CPU. I'm
on a dual-core MacBook Pro so the maximum I should be seeing is 20k or
40k samples. I gather that 90% of the startup time is spent in disk
and other IO.
There shouldn't be much other IO as Firefox barely gets a chance to
come up before getting killed. The sum of disk IO timings rolled up by
file name and read/write op [1] doesn't come close to the difference,
though. Is there an explanation for this?
I'm sampling cold startup (after 'sync && purge') and using scripts
from my Git repo [1]. Time is in microseconds.
Thanks, Joel
[1] http://gist.github.com/175733
[2] http://github.com/wagerlabs/firefox-startup/tree/master
---
fastest mac firefox!
http://wagerlabs.com
_______________________________________________
dtrace-discuss mailing list
dtrace-discuss@opensolaris.org