[ENH]: Add support for pbsa.cuda

[marioernestovaldes]

In which tool?

No response

New Feature

• [ ] Add support for pbsa.cuda

Description

Install pbsa.cuda as follows:

Line 37 run.cmake file in amber20_src/build:

#  Assume this is Linux:

  cmake $AMBER_PREFIX/amber20_src \
    -DCMAKE_INSTALL_PREFIX=$AMBER_PREFIX/amber20 \
    -DCOMPILER=MANUAL -DCMAKE_C_COMPILER=gcc-6 -DCMAKE_CXX_COMPILER=g++-6 -DCMAKE_Fortran_COMPILER=gfortran-7\
    -DMPI=FALSE -DCUDA=TRUE -DINSTALL_TESTS=TRUE \
    -DDOWNLOAD_MINICONDA=TRUE -DMINICONDA_USE_PY3=TRUE \
    2>&1 | tee  cmake.log

See that -DCUDA=TRUE and we specify MANUAL for -DCOMPILER with all the routes to different compilers

if we want to test this in hpc, run this command to get a temporal assignation of resources:

salloc --nodes=1 --ntasks-per-node=1 --mem-per-cpu=2G --time=5 --gpus=2 --account='name_of_the_project'

Relevance

low

Difficulty to implement

medium

[sherwoodyao]

+1 excited to see this support coming soon!