[gmx-developers] real in C++

Szilárd Páll szilard.pall at cbr.su.se
Mon Feb 25 16:37:06 CET 2013

One thing to note is that we would like to have an additional mixed
precision scheme, hopefully in 5.0. The idea is that part of the code, like
force calculations, would run in single precision with fixed precision
accumulators/reduction and double for whatever it makes sense/is useful.
Using real everywhere could make achieving this quite difficult, right?


On Mon, Feb 25, 2013 at 3:50 PM, Mark Abraham <mark.j.abraham at gmail.com>wrote:

> I'm not aware of a need to change the way we implement this.
> Berk and I have mooted the desirability of switching our naming scheme to
> "double-" vs "mixed-precision" on the grounds that aspects of the virial
> computation have been in double precision for a long time. We may as well
> avoid people leaping to the conclusion that "single-precision" GROMACS must
> be dodgy because you need double precision for some aspects of the
> algorithm.
> Mark
> On Sun, Feb 24, 2013 at 10:09 PM, David van der Spoel <
> spoel at xray.bmc.uu.se> wrote:
>> Hi,
>> did we decide on how to treat real variables in C++?
>> There is nothing stopping us from keeping the same
>> #ifdef GMX_DOUBLE
>> typedef double real;
>> #else
>> typedef float real;
>> #endif
>> that we have use for decades. Or is there reason to change it?
>> Cheers,
>> --
>> David van der Spoel, Ph.D., Professor of Biology
>> Dept. of Cell & Molec. Biol., Uppsala University.
>> Box 596, 75124 Uppsala, Sweden. Phone:  +46184714205.
>> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
>> --
>> gmx-developers mailing list
>> gmx-developers at gromacs.org
>> http://lists.gromacs.org/**mailman/listinfo/gmx-**developers<http://lists.gromacs.org/mailman/listinfo/gmx-developers>
>> Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-developers-request@**gromacs.org<gmx-developers-request at gromacs.org>
>> .
> --
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20130225/15797511/attachment.html>

More information about the gromacs.org_gmx-developers mailing list