Add new machine perlmutter

configs/perlmutter.config

@@ -0,0 +1,36 @@
+# NERSC Perlmutter (HPE Cray Compiler Wrappers)
+
+### Caution!! Load the module below before compiling
+### module load gcc-native/12.3
+
+CUDA_PATH       /opt/nvidia/hpc_sdk/Linux_x86_64/23.9/cuda/12.2/
+FFTW2_PATH
+FFTW3_PATH      /opt/cray/pe/fftw/3.3.10.6/x86_milan/
+MPI_PATH        /opt/cray/pe/craype/2.7.30/  # This is actually the path to CC (just for perlmutter).
+HDF5_PATH       /opt/cray/pe/hdf5/default/gnu/12.3
+#HDF5_PATH       /opt/cray/pe/hdf5-parallel/1.12.2.9/gnu/12.3/
+GRACKLE_PATH
+GSL_PATH
+LIBYT_PATH
+CUFFTDX_PATH
+
+# compilers
+CXX     CC
+CXX_MPI CC                # MPI compiler (The MPI_PATH/bin/ will be combined automatically)
+
+# flags
+CXXFLAG -target-accel=nvidia80   # Also available in the environment variable
+CXXFLAG -g
+CXXFLAG -O2
+
+OPENMPFLAG -fopenmp
+
+#LIBFLAG
+
+NVCCFLAG_COM -O3 -I/opt/cray/pe/mpich/8.1.28/ofi/gnu/12.3/include
+#NVCCFLAG_COM -use_fast_math
+NVCCFLAG_FLU -Xptxas -dlcm=ca -prec-div=false -ftz=true
+NVCCFLAG_POT -Xptxas -dlcm=ca
+
+# gpu
+GPU_COMPUTE_CAPABILITY 800    # NVIDIA A100

example/queue/submit_perlmutter_cpu.job

@@ -0,0 +1,21 @@
+#!/bin/bash
+#SBATCH --account=YOUR_ACCOUNT
+#SBATCH --job-name=GAMER_CPU
+#SBATCH --qos=debug
+#SBATCH --time=0:30:00
+#SBATCH --nodes=1
+#SBATCH --mail-type=BEGIN,END,FAIL
+#SBATCH --mail-user=YOUR_EMAIL
+
+########## Do NOT modify this block ##########
+#SBATCH --constraint=cpu
+#SBATCH --ntasks-per-node=1
+#SBATCH --cpus-per-task=256
+export OMP_PROC_BIND=spread
+export OMP_PLACES=threads
+##############################################
+
+module load gcc-native/12.3
+module list  # Record the loaded modules
+
+srun ./gamer

example/queue/submit_perlmutter_gpu.job

@@ -0,0 +1,22 @@
+#!/bin/bash
+#SBATCH --account=YOUR_ACCOUNT_g
+#SBATCH --job-name=GAMER_GPU
+#SBATCH --qos=debug
+#SBATCH --time=0:30:00
+#SBATCH --nodes=1
+#SBATCH --mail-type=BEGIN,END,FAIL
+#SBATCH --mail-user=YOUR_EMAIL
+
+########## Do NOT modify this block ##########
+#SBATCH --constraint=gpu
+#SBATCH --ntasks-per-node=4
+#SBATCH --cpus-per-task=32
+#SBATCH --gpus-per-task=1
+export OMP_PROC_BIND=spread
+export OMP_PLACES=threads
+##############################################
+
+module load gcc-native/12.3
+module list  # Record the loaded modules
+
+srun ./gamer