From paul at guermonprez.org  Fri Oct 27 10:24:33 2006
From: paul at guermonprez.org (Paul Guermonprez)
Date: Fri, 27 Oct 2006 16:24:33 +0200
Subject: [emboss-dev] Request For Comments : optimisation for water-emboss
	algo.
Message-ID: <74d4a1e10610270724s4a068299nd66d9b232716b9f5@mail.gmail.com>

Hello,


I am a bioinformatician working for intel in france,
I have been working a few days on the water algo included in emboss.
the performance speedup is x11.8 on a typical xeon setup.
the speedup is not only good with intel compiler but with GCC too.

if some of you could read it and give me their opinion,
i would be glad to hear it.

please find the pdf of the slides here :
http://www.guermonprez.net/projects/emboss/emboss-EBI_06_06.pdf
and the sources here (need emboss 3.0.0 to work) :
http://www.guermonprez.net/projects/emboss/emboss-runsh.tar.gz

do you think that sort of optimization could be included in
the main EMBOSS distribution ?
of course, spending more time would bring better serial results,
but an interesting target would be to use OpenMP pragmas
to parallelise the execution. what do you think about openmp pragmas
(not seen by GCC yet) in emboss code ?


regards, paul.