Amber And AmberTools Cheatsheet

Running simulation:

# serial:
pmemd -O -i md.input -o md.output -p prmtop -c  last_run.rst \
-r this_run.rst -x md.mdcrd -inf md.mdinfo

# parrallel:
mpirun -n nproc pmemd.MPI -O -i md.input -o md.output -p prmtop -c  last_run.rst \
-r this_run.rst -x md.mdcrd -inf md.mdinfo

# single GPU:
pmemd.cuda -O -i ...

# two single GPU runs:
export CUDA_VISIBLE_DEVICES=0
cd run_1
pmemd.cuda -O -i ... &

export CUDA_VISIBLE_DEVICES=1
cd run_2
pmemd.cuda -O -i ... &

# dual GPUs:
mpirun -np 2 pmemd.cuda.MPI -O -i ...

Restart to pdb:

ambpdb -p foo.prmtop -c bar.rst > output.pdb
ambpdb -p foo.prmtop -c bar.inpcrd > output.pdb

mdcrd to dcd:

cpptraj -p foo.prmtop -y bar.mdcrd -x output.dcd