ID=267

• Dataset: 9naw
• Software: PMEMD.cuda.mpi (amber/24.3-gofbc-2023.2.415.126.-avx512)
• Resource: 8 tasks, 1 cores, 1 nodes, 8 GPUs, with NVLink
• CPU: Xeon Platinum 8570 (Emerald Rapids), 2.1 GHz
• GPU: NVidia-H100-SXM5-80GB, 14 cores/GPU
• Simulation speed: 30.13 ns/day
• Efficiency: 23.6 %
• Site: Nibi
• Date: Aug. 26, 2025, 12:32 p.m.
• Submission script:

  !/bin/bash
  #SBATCH -A def-svassili
  #SBATCH --nodes=1 --gpus-per-node=8 --ntasks=8 --cpus-per-task=14 --mem-per-cpu=4000 --time=1:0:0
  
  module --force purge
  module load  StdEnv/2023  gcc/12.3  openmpi/4.1.5  cuda/12.6 amber-pmemd/24.3
  
  # Run simulation 
  srun pmemd.cuda.MPI -O -i pmemd_prod.in -p prmtop.parm7 -c restart.rst7

• Notes:

• Simulation input file:

  Production 
  &cntrl 
  irest=1, ntx=5, 
  cut=9.0, 
  dt=0.002,
  ntt = 3, gamma_ln = 1.0, temp0=300,
  ntp = 1, barostat = 1, pres0 = 1, taup = 1.0, 
  ntc = 2, ntf = 2,
  ntpr=1000,                                  ! Print energies every ntpr steps
  ntwx=10000, ntwr=100000,                    ! coordinates, restart
  nstlim=200000
  /
   &ewald
    nfft1=336,
    nfft2=336,
    nfft3=336,
  /