[gmx-developers] random123 library portability issues

[Alexey Shvetsov]

Roland Schulz писал 01-07-2014 03:59:
> HI Alexey ,
> 
> On Fri, May 30, 2014 at 6:52 AM, Alexey Shvetsov
> <alexxy at omrb.pnpi.spb.ru> wrote:
> 
>> Hi!
>> 
>> I tryed to build 5.0 branch on ARM board (dual core cortex-a7). It
>> builds (if
>> i define arm as allowed arch in features/gccfeatures.h) but some
>> regression
>> tests fails.
> 
> What compiler did you use? Which tests fail?

Hi!

I tryed with gcc-4.8.2. Spc and methanol test failed with SIGBUS

There were 2 notes
Reading file 
/home/alexxy/Develop/gromacs_build/none/src/programs/mdrun/tests/Testing/Temporary/CanWrite_Trajectories_ThatDifferInNstxout_0.tpr, 
VERSION 5.0-beta2-dev-20140211-9a31310 (single precision)
Changing nstlist from 10 to 25, rlist from 1.022 to 1.076

Using 1 MPI thread
starting mdrun 'spc-and-methanol'
6 steps,      0.0 ps.
Bus error


I'll rebuild current master with gcc-4.9 to check if there are same 
issue.

> 
> Roland
> 
>> В письме от 30 мая 2014 12:43:29
>> пользователь Szilárd Páll написал:
>> 
>>> Hi,
>>> 
>>> A few days ago I did some sanity checks by compiling and
>>> regressiontesting the 5.0 branch on ARM. This exercise made me
>> realize
>>> that the recently included Random123 library is now the single
>> most
>>> limiting factor when it comes to portability. In fact, as it is
>> not
>>> optional, v5.0 is ATM practically limited to x86* and Powerpc64
>> (BGQ)
>>> architectures -- based on the docs (see http://goo.gl/YdKofG [1])
>>> Random123 has not been tested (let alone optimized) on anything
>> else.
>>> Additionally, note that the code won't even compile on
>> non-supported
>>> architectures, so in its current state GROMACS 5.0 won't compile
>> on
>>> anything else but x86* and PPC64, e.g. with gcc compilation will
>> be
>>> aborted by the "#error" in features/gccfeatures.h:38. Removing
>> these
>>> sanity checks does let the code compile on ARM and
>> regressiontests do
>>> pass, though.
>>> 
>>> I've exchanged a couple of emails with the developers and they
>> suggest
>>> that in practice Threefry should work on ARM (and perhaps other
>>> architectures too), but testing/validation is definitely
>> necessary.
>>> Performance is not optimized on unsupported architectures, I
>> assume,
>>> so some optimization may not hurt too - especially if there are
>> cases
>>> where we generate many PRNGs per step (BTW has the potential
>> overhead
>>> been assessed?).
>>> 
>>> Cheers,
>>> --
>>> Szilárd
>> 
>> --
>> Best Regards,
>> Alexey 'Alexxy' Shvetsov, PhD
>> Department of Molecular and Radiation Biophysics
>> FSBI Petersburg Nuclear Physics Institute, NRC Kurchatov Institute,
>> Leningrad region, Gatchina, Russia
>> mailto:alexxyum at gmail.com
>> mailto:alexxy at omrb.pnpi.spb.ru
>> 
>> --
>> Gromacs Developers mailing list
>> 
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List [2]
>> before posting!
>> 
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists [3]
>> 
>> * For (un)subscribe requests visit
>> 
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers
>> [4] or send a mail to gmx-developers-request at gromacs.org.
> 
> --
> ORNL/UT Center for Molecular Biophysics cmb.ornl.gov [5]
> 865-241-1537, ORNL PO BOX 2008 MS6309
> 
> Links:
> ------
> [1] http://goo.gl/YdKofG
> [2] http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List
> [3] http://www.gromacs.org/Support/Mailing_Lists
> [4] 
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers
> [5] http://cmb.ornl.gov

-- 
Best Regards,
Alexey 'Alexxy' Shvetsov, PhD
Department of Molecular and Radiation Biophysics
FSBI Petersburg Nuclear Physics Institute, NRC Kurchatov Institute,
Leningrad region, Gatchina, Russia
mailto:alexxyum at gmail.com
mailto:alexxy at omrb.pnpi.spb.ru