You probably won't get the fastest executable possible by releasing
something you prebuild, and I would really want to check out the licensing
issues carefully too. I guess I have always thought that amber was
intrinsically a source product to support folks playing with the code. I do
think that the incredible noise in the configuration space caused by all the
junk that intel has been doing is very unfortunate, and that does make using
things like -fast desirable, but just because there is a -fast flag, it does
not mean that you will get the best performance on all codes (got to
remember, every executable is at least slightly, if not wildly different,
than others in terms of how best to optimize it). My view has always been
that since pmemd is being built pretty much purely for a combination of
speed and reliability (okay, I have occasionally added functionality that I
think is nice...), then you have to put a little more effort into compiler
options. I have dropped out of the fray for this last round of development
due to funding; if funding arrangements occur that allow me to work on this
stuff again, I will try to get it all done correctly again for pmemd 12, but
there are no guarantees on funding at the moment.
Regards - Bob
----- Original Message -----
From: "Mengjuei Hsieh" <mengjueh.uci.edu>
To: "AMBER Developers Mailing List" <amber-developers.ambermd.org>
Sent: Tuesday, February 02, 2010 4:40 AM
Subject: Re: [AMBER-Developers] questions about compiler flags
> On Mon, Feb 01, 2010 at 03:18:40PM -0500, case wrote:
>> Two compiler options questions:
> [libc issues deleted]
>> 2. Do we really need the "-ip -O3 -xHost" flags for optimized ifort
>> compiles?
>> Is this better/worse/the same as "-fast" (used for pmemd)? Does sander
>> need
>> anything more than -O3? Is there a good compromise between "fastest of
>> all
>> 2^N possible combinations of flags on this particular machine" and
>> "something
>> that sacrifices a few percent in performance but which is likely to both
>> work
>> now and continue working after Intel puts out its next release"?
>
> IIRC, "-ip" doesn't really optimized the speed too much although it
> does theoretically reduce program loading time. As to -xHost, isn't
> it for replacing SSE1/2/3 ? I like SSE* arguments, but I am not sure
> if that gives you dramatic performance boost.
>
> If performance is really a concern for certain programs, how about
> we release some well-optimized executables for them on some major
> platforms?
>
> Best,
> --
> Mengjuei
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Feb 02 2010 - 05:30:02 PST
