Re: dyld shared cache: firefox vs safari
Re: dyld shared cache: firefox vs safari
- Subject: Re: dyld shared cache: firefox vs safari
- From: Greg Parker <email@hidden>
- Date: Wed, 26 Aug 2009 18:49:37 -0700
On Aug 26, 2009, at 2:06 PM, Joel Reymont wrote:
I wrote io.d, a DTrace script and a derivative of iosnoop, to roll
up disk io time during the first 20s of the Firefox startup [1].
I already have a 'cold startup' shell script that I use to trace
Firefox, e.g.
---
#!/bin/bash
cmd="./Minefield.app/Contents/MacOS/firefox-bin -no-remote -
foreground -P 2"
sync && purge && dtrace -x dynvarsize=64m -x evaltime=exec -c "$cmd"
-wZs $1
---
I cold-started and traced both Firefox [2] and Safari [3]. I now
have a hunch that the slow Firefox startup may have a lot to do with
the fact that most of the Safari bits are pulled from the dyld
shared cache.
Exercise: run both Safari and Firefox with environment variable
DYLD_SHARED_REGION=avoid. This will turn off the dyld shared cache,
removing that variable from your startup time measurements. If Safari
is still faster to launch, it's not the shared cache's fault.
DYLD_PRINT_STATISTICS may also be useful in identifying where your
startup time is going.
--
Greg Parker email@hidden Runtime Wrangler
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden