[gmx-users] intel compiler for AMD CPU?

Szilárd Páll pall.szilard at gmail.com
Fri Oct 10 21:19:43 CEST 2014

On Fri, Oct 10, 2014 at 8:51 PM, Albert <mailmd2011 at gmail.com> wrote:
> I doubt this. According to my own experiences, intel c++ plus intel MPI make
> the best efficiency.......

Reference please? Are you referring to best efficiency i) with code
some code X, ii) with any arbitrary code or ii) with GROMACS? The
first two can be (made) true, but I assume we are discussing about the
latter here.

I've been advising people to stop believing that the the Intel
compiler is faster just because it's Intel. I did that based on:
i) comparisons of compilers *on GROMACS* that I do regularly; since
v4.6 or 4.7 gcc has been generating faster GROMACS code than Intel and
icc 13+ even has some regressions wrt 12; Just recently I compared
again intel 12,13,15 with gcc 4.6,4.7,4.8,4.9 and clang 3.3, 3.4, 3.5
and gcc 4.8/4.9 was fastest in all cases, icc got often very close,
but it was never faster (note that this was an extensive, but not a
exhaustive comparison!)
ii) I've heard the same experience from several others in the HPC community.


> but I am not so sure for this under AMD CPU....
> On 10/10/2014 08:47 PM, Johnny Lu wrote:
>> The mailling list said that for intel xeon, gcc+fftw3 make faster gromacs
>> than intel c/c++ compiler + intel math kernel library.
> --
> Gromacs Users mailing list
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a
> mail to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list