how to install gpu version of genesis on group server

Install cuda 11 environment with miniconda

Because the server only have cuda-12 and cuda-12 will make the genesis get wrong results. Therefore, we need use miniconda to install cuda-11 enviroment. We use below command to make a environment include cuda-11.8.

1	conda create -n cuda_11 -c nvidia cudatoolkit=11 cuda-nvcc=11 cuda-cudart-dev=11 cuda-nvtx=11

You can activate the environment by conda activate cuda_11.

Then, we source the intel compiler and add the cuda library to the LD_LIBRARY_PATH:

1 2	source /opt/intel/oneapi/setvars.sh export LD_LIBRARY_PATH=/your/miniconda/envs/cuda_11/lib:$LD_LIBRARY_PATH

Compile genesis

Then we can compile genesis use below command

1
2
3

./configure --enable-gpu --enable-mixed --program-suffix='-intel-mixed-cuda11-conda'
make -j 8
make install

After that, the genesis can preform correct calculation on cell.