develooper Front page | perl.perl5.porters | Postings from February 2003

Re: TODO (was Re: [PATCH] Re: change #9754: 5 x slowdown for perl -d)

Thread Previous | Thread Next
From:
Jarkko Hietaniemi
Date:
February 1, 2003 12:43
Subject:
Re: TODO (was Re: [PATCH] Re: change #9754: 5 x slowdown for perl -d)
Message ID:
20030201204239.GB30134@kosh.hut.fi
On Sat, Feb 01, 2003 at 11:36:50AM +0000, Nicholas Clark wrote:
> On Mon, Jan 27, 2003 at 01:08:07AM +0200, Jarkko Hietaniemi wrote:
> 
> > > So there's nothing big to come out :-(
> > > Any speed improvements have to be found from new ideas?
> > 
> > Finding out hot spots, writing fast paths for them, inlining them.
> > Oh, sorry, you said *new* ideas...
> 
> Is there a list of known hot spots?
> By "known", I mean known down to the accuracy to paths through particular
> functions that are slow, with benchmarking test cases?

Not that I know of.  We can use profiling tools but statistic
profiling is not that useful unless it gives us also at least
caller-callees.  With basic block profiling we get usually that,
and also better per-line accuracy.  That way we can recognize at
least any "hot lines".

-- 
Jarkko Hietaniemi <jhi@iki.fi> http://www.iki.fi/jhi/ "There is this special
biologist word we use for 'stable'.  It is 'dead'." -- Jack Cohen

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About