Dear Experts,
I am trying to install amber22 in pmemd.cuda.MPI. I do the following steps.
module load compiler/cuda/11.0 compiler/cmake/3.15.4
compiler/openmpi/4.1.4 compiler/gcc/8.3.0
Then I install nccl using the following command.
git clone
https://github.com/NVIDIA/nccl.git
cd nccl
git checkout ‘git tag | tail -n1‘
make src.build CUDA_HOME=/opt/ohpc/pub/cuda-11.0/
I also set a path for nccl.
export PATH="/home/21cy91r03/nccl/build/include:$PATH"
export LD_LIBRARY_PATH="/home/21cy91r03/nccl/build/lib:$LD_LIBRARY_PATH
And add -DNCCL=TRUE in the run cmake file.
But the following error appears during compilation.
Building the GTI version of pmemd.cuda
CMake Error at cmake/PMEMDCompilerFlags.cmake:239 (message):
NCCL is selected for inter-GPU communications but was not found.
Call Stack (most recent call first):
CMakeLists.txt:212 (include)
-- Configuring incomplete, errors occurred!
See also "/home/21cy91r03/amber22_src/build/CMakeFiles/CMakeOutput.log".
See also "/home/21cy91r03/amber22_src/build/CMakeFiles/CMakeError.log".
--
*With regards,*
*Dulal Mondal,*
*Research Scholar,*
*Department of Chemistry,*
*IIT Kharagpur, Kharagpur 721302.*
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Jan 02 2024 - 22:00:02 PST
