So,

This is slightly aside from the installation of the openMM (cuda), but is CUDA 
related.

Does anyone know the state of the CUDA to OpenCL porting the AMD (CUDA teams 
said they were going to do this) and then further all CUDA work as OpenCL, as 
no one has mentioned anything on this since 2010 (the CUDA web site).  There 
are however several useful code bits such as openCLpython, JAVAopenCL, etc...

Also, How much for an ATI emulator that works with CUDA?  As I understand it 
the only main problem is how the 8 by 8 GPU compute units are arranged on the 
device, and not even the actual code, so one could just easily write the ATI 
device specs into a CUDA code or vise versa?

Just interested but lacking any time to even start to attempt such things, but 
anything of this nature would make my life and many I know much easier.

Sincerely,

Stephan Lloyd Watkins

-------- Original-Nachricht --------
> Datum: Tue, 31 Jan 2012 21:50:53 +0300
> Von: "Андрей Гончар" <[email protected]>
> An: Discussion list for GROMACS users <[email protected]>
> Betreff: Re: [gmx-users] Installing GMX-GPU 4.5.5

> You should try this:
> http://verahill.blogspot.com/2012/01/debian-testing-64-wheezy-compiling_20.html
> 
> 2012/1/31 Efrat Exlrod <[email protected]>:
> > Hi,
> >
> > I'm compiling gromacs 4.5.5 with gcc compiler (v 4.5.3), cmake (2.8.7)
> and
> > OpenMM 3.1.1 on Linux (Red Hat release 5.7). I have followed the
> > installation instructions.
> >
> > The configuration seems to work well.
> >
> >> ~/progs/cmake-2.8.7/bin/cmake -DGMX_OPENMM=ON
> >> -DCUDA_TOOLKIT_ROOT_DIR:PATH=/opt/cuda
> >> -DCMAKE_C_COMPILER:FILEPATH=/private/gnss/local/bin/gcc
> >> -DCMAKE_INSTALL_PREFIX=/private/gnss/Gromacs_455
> >
> > But, when I run make mdrun I get the following error:
> >
> >>make mdrun
> > [  0%] Building NVCC (Device) object
> > src/kernel/gmx_gpu_utils/./gmx_gpu_utils_generated_memtestG80_core.cu.o
> > cc1plus: error: unrecognized command line option
> "-fexcess-precision=fast"
> > CMake Error at
> > CMakeFiles/gmx_gpu_utils_generated_memtestG80_core.cu.o.cmake:198
> (message):
> >   Error generating
> >
> >
> /private/gnss/Gromacs_Install_455/gromacs-4.5.5/src/kernel/gmx_gpu_utils/./gmx_gpu_utils_generated_memtestG80_core.cu.o
> >
> >
> > make[3]: ***
> >
> [src/kernel/gmx_gpu_utils/./gmx_gpu_utils_generated_memtestG80_core.cu.o]
> > Error 1
> > make[2]: *** [src/kernel/gmx_gpu_utils/CMakeFiles/gmx_gpu_utils.dir/all]
> > Error 2
> > make[1]: *** [src/kernel/CMakeFiles/mdrun.dir/rule] Error 2
> > make: *** [mdrun] Error 2
> >
> > When I run make mdrun after deleting the 2 occurences of
> > "-fexcess-precision=fast" from CMakeCache.txt the compilation works.
> >
> > What could be the problem?
> >
> > Thanks, Efrat
> >
> > --
> > gmx-users mailing list    [email protected]
> > http://lists.gromacs.org/mailman/listinfo/gmx-users
> > Please search the archive at
> > http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> > Please don't post (un)subscribe requests to the list. Use the
> > www interface or send it to [email protected].
> > Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> 
> 
> 
> -- 
> 
> Андрей Гончар
> -- 
> gmx-users mailing list    [email protected]
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to [email protected].
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

-- 
NEU: FreePhone 3-fach-Flat mit kostenlosem Smartphone!                          
        
Jetzt informieren: http://www.gmx.net/de/go/freephone/
-- 
gmx-users mailing list    [email protected]
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at 
http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
Please don't post (un)subscribe requests to the list. Use the 
www interface or send it to [email protected].
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

Reply via email to