Subject: Re: [Boost-build] Major engine changes
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2011-11-29 21:49:27
On 11/29/2011 05:57 PM, Rene Rivera wrote:
> On 11/29/2011 6:02 PM, Steven Watanabe wrote:
>> I've just committed a major rewrite of the
>> jam interpreter. So, far, I'm seeing a 2x speed
> Using which metric?
Elapsed wall-clock time.
> And which test case?
bjam -n -d0 in $BOOST_ROOT/status
on Linux. The initial state was
what this command leaves behind,
so repeated calls would run the
same updating actions.
> Did memory usage change also?
> (This is one area where I know we are slow as a consequence of the
> enormous mem use.)
Apparently so, although I wasn't particularly
trying to reduce memory usage.
On Windows compiling with MSVC 10.0 Express,
with the same command:
before: 231 seconds, 790 MB
after: 150 seconds, 665 MB
In $BOOST_ROOT with bjam --build-type=complete -n -d0:
before: 38 seconds, 110 MB
after: 27 seconds, 88 MB
>> Here's a summary of all the behavior changes
>> that I am aware of. Some of these are just
>> bug fixes, but I'm including them for completeness.
> OK.. But I really would like to see a summary of the performance changes.
Most of the speed up comes from replacing the
insane variable expansion algorithm.
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk