On Aug 27, 2008, at 5:57 , Bradford Winfrey wrote:
Yea, I figured it wouldn't make an enormous difference since the
vast majority of "power" is left to Erl. Interesting still!
2x speed for 10x mem consumption on my ad-hoc,
non-scientific testing. The keeping-memory-down
methods of Spidermonkey either are not in place yet
or I was doing it wrong. But i didn't spend a whole lot
of time one it, so take this with a grain of salt.
Yes, I had to try that immediately :)
Cheers
Jan
--
PS: Sorry for the empty email earlier.
----- Original Message ----
From: Chris Anderson <[EMAIL PROTECTED]>
To: [email protected]
Sent: Tuesday, August 26, 2008 10:40:51 PM
Subject: Re: Any benefit in the pipe from TraceMonkey?
On Tue, Aug 26, 2008 at 8:26 PM, Bradford Winfrey
<[EMAIL PROTECTED]> wrote:
I was just curious about CouchDB's performance (not that I have any
problem with it at the moment). As the buzz surrounding FireFox's
javascript support amplifies with each day they get closer to the
3.0.1 release, does this mean that there will be some way to slip
TraceMonkey into the mix and reap any benefits? Just curious - as
I will have a success story to share with the masses using CouchDB
as a backend very, very soon.
In my experience, Javascript view evaluation is not a large part of
the time spent in view generation. Of course a faster Javascript
engine will speed things up, but the work that is currently being done
to make CouchDB take advantage of more cores may have a greater
effect.
That said, I'm not sure what roadblocks (if any) there are to using
TraceMonkey.
Chris
--
Chris Anderson
http://jchris.mfdz.com
