Hi Jason, thanks.
Yes we did that and the configuration seems to be working OK, now we seem to have a compilation error see this:
Hi Fabian,
The compilation of amber with cuda failed for this version (cude 6.5).
Previously it compiled properly with the previous cude 5.5.
Please sent the compilation error below to the amber people.
ptxmacros.h(69): error: function "__shfl(unsigned int, int, int)" has already been defined
ptxmacros.h(100): error: function "__shfl(double, int, int)" has already been defined
2 errors detected in the compilation of "/tmp/tmpxft_00004ce7_00000000-12_kForcesUpdate.compute_30.cpp1.ii".
make[5]: *** [kForcesUpdate.o] Error 2
make[5]: Leaving directory `/usr/local/amber-gpu/src/pmemd/src/cuda'
make[4]: *** [cuda/cuda.a] Error 2
make[4]: Leaving directory `/usr/local/amber-gpu/src/pmemd/src'
make[3]: *** [cuda_parallel] Error 2
make[3]: Leaving directory `/usr/local/amber-gpu/src/pmemd'
make[2]: *** [cuda_parallel] Error 2
make[2]: Leaving directory `/usr/local/amber-gpu/src'
make[1]: [cuda_parallel] Error 2 (ignored)
make[1]: Leaving directory `/usr/local/amber-gpu/AmberTools/src'
make[1]: Entering directory `/usr/local/amber-gpu/src'
Starting installation of Amber12 (cuda parallel) at Thu Sep 18 14:12:53 IDT 2014.
cd pmemd && make cuda_parallel
make[2]: Entering directory `/usr/local/amber-gpu/src/pmemd'
make -C src/ cuda_parallel
make[3]: Entering directory `/usr/local/amber-gpu/src/pmemd/src'
make -C ./cuda
make[4]: Entering directory `/usr/local/amber-gpu/src/pmemd/src/cuda'
/usr/local/cuda/bin/nvcc -use_fast_math -gencode arch=compute_13,code=sm_13 -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -DCUDA -DMPI -DMPICH_IGNORE_CXX_SEEK -Duse_SPFP -I/usr/local/cuda/include -IB40C -IB40C/KernelCommon -I/usr/local/intel/impi/4.0.3.008/intel64/include -c kForcesUpdate.cu
nvcc warning : The 'compute_11', 'compute_12', 'compute_13', 'sm_11', 'sm_12', and 'sm_13' architectures are deprecated, and may be removed in a future release.
ptxmacros.h(69): error: function "__shfl(unsigned int, int, int)" has already been defined
ptxmacros.h(100): error: function "__shfl(double, int, int)" has already been defined
2 errors detected in the compilation of "/tmp/tmpxft_00004d13_00000000-12_kForcesUpdate.compute_30.cpp1.ii".
make[4]: *** [kForcesUpdate.o] Error 2
make[4]: Leaving directory `/usr/local/amber-gpu/src/pmemd/src/cuda'
make[3]: *** [cuda/cuda.a] Error 2
make[3]: Leaving directory `/usr/local/amber-gpu/src/pmemd/src'
make[2]: *** [cuda_parallel] Error 2
make[2]: Leaving directory `/usr/local/amber-gpu/src/pmemd'
make[1]: *** [cuda_parallel] Error 2
make[1]: Leaving directory `/usr/local/amber-gpu/src'
make: *** [install] Error 2
Amihai
HPC Support
CIS - Technion
_______________________________
Fabian Glaser, PhD
Head of the Structural Bioinformatics section
Bioinformatics Knowledge Unit - BKU
The Lorry I. Lokey Interdisciplinary
Center for Life Sciences and Engineering
Technion - Israel Institute of Technology
Haifa 32000, ISRAEL
fglaser.technion.ac.il
Tel: +972 4 8293701
Fax: +972 4 8225153
On Sep 18, 2014, at 2:40 PM, Jason Swails <jason.swails.gmail.com> wrote:
> On Thu, 2014-09-18 at 11:39 +0300, Fabian Glaser wrote:
>> Hi all,
>>
>> We have updated all the following drivers on ReadHat 6.4 (see below),
>> and still running pmemd.cuda I get a different error this time, sander
>> works normally:
>>
>> -bash-4.1$ pmemd.cuda -O -i md.in -p RIT_6mol_h_wat.prmtop -c
>> RIT_6mol_h_wat_eq5.rst -ref RIT_6mol_h_wat_eq5.rst -o
>> RIT_6mol_h_wat_md1.out -x RIT_6mol_h_wat_md1.mdcrd -r
>> RIT_6mol_h_wat_md1.rst
>>
>> pmemd.cuda: error while loading shared libraries: libcurand.so.5.5:
>> cannot open shared object file: No such file or directory
>>
>> Is this something known to you?
>> Looking into different forums, our technical people don't find the
>> specific library....
>
> The libcurand.so.5.5 library is part of the CUDA toolkit distribution.
> If you run "./configure -cuda gnu", the end of the configure step you
> should see a message like this:
>
> --------------------------------------------------------------------------------
> Environment resource files are provided to set the proper environment
> variables to use AMBER and AmberTools.
>
> If you use a Bourne shell (e.g., bash, sh, zsh, etc.), source the
> /home/swails/build_amber/amber/amber.sh file in your shell. Consider adding the line
> test -f /home/swails/build_amber/amber/amber.sh && source /home/swails/build_amber/amber/amber.sh
> to your startup file (e.g., ~/.bashrc)
>
> If you use a C shell (e.g., csh, tcsh), source the
> /home/swails/build_amber/amber/amber.csh file in your shell. Consider adding the line
> test -f /home/swails/build_amber/amber/amber.csh && source /home/swails/build_amber/amber/amber.csh
> to your startup file (e.g., ~/.cshrc)
>
> If you have not already done so, you may need to add /opt/cuda/5.5/lib
> and/or /opt/cuda/5.5/lib64 to LD_LIBRARY_PATH using the command
> export LD_LIBRARY_PATH="/opt/cuda/5.5/lib:${LD_LIBRARY_PATH}" (bash, sh)
> setenv LD_LIBRARY_PATH "/opt/cuda/5.5/lib:${LD_LIBRARY_PATH}" (tcsh, csh)
> --------------------------------------------------------------------------------
>
> The last piece of advice -- namely adding $CUDA_HOME/lib and/or
> $CUDA_HOME/lib64 to your LD_LIBRARY_PATH -- is the root cause of your
> error. On my computer, CUDA_HOME happens to be set to /opt/cuda/5.5.
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 18 2014 - 06:00:04 PDT
