diff --git a/systems/Perlmutter/config-command b/systems/Perlmutter/config-command
index b399c535..b62704dc 100644
--- a/systems/Perlmutter/config-command
+++ b/systems/Perlmutter/config-command
@@ -1,9 +1,13 @@
+DIR=`pwd`
+PREFIX=$DIR/../Prequisites/install/
 ../../configure \
     --enable-comms=mpi \
     --enable-simd=GPU \
     --enable-shm=nvlink \
     --enable-gen-simd-width=64 \
     --enable-accelerator=cuda \
+    --disable-accelerator-cshift \
+    --with-gmp=$PREFIX \
     --disable-fermion-reps \
     --disable-unified \
     --disable-gparity \
diff --git a/systems/Perlmutter/dwf4.slurm b/systems/Perlmutter/dwf4.slurm
index ba198595..f6d6a2a9 100644
--- a/systems/Perlmutter/dwf4.slurm
+++ b/systems/Perlmutter/dwf4.slurm
@@ -1,24 +1,20 @@
 #!/bin/bash
-#SBATCH -A mp13
+#SBATCH -A m3886_g
 #SBATCH -C gpu
-#SBATCH -q regular
-#SBATCH -t 0:20:00
-#SBATCH -n 16
+#SBATCH -q debug
+#SBATCH -t 0:10:00
+#SBATCH -n 4
 #SBATCH --ntasks-per-node=4
 #SBATCH -c 32
 #SBATCH --exclusive
 #SBATCH --gpus-per-task=1
-#SBATCH --gpu-bind=map_gpu:0,1,2,3
+#SBATCH --gpu-bind=none
 
 export SLURM_CPU_BIND="cores"
-export MPICH_RDMA_ENABLED_CUDA=1
 export MPICH_GPU_SUPPORT_ENABLED=1
-srun ./benchmarks/Benchmark_comms_host_device --mpi 2.2.2.2  --accelerator-threads 8 > comms.4node
-
-OPT="--comms-overlap --comms-concurrent --shm-mpi 0"
-srun ./benchmarks/Benchmark_dwf_fp32 --mpi 2.2.2.2 --grid 64.64.64.64 --accelerator-threads 8 --shm 2048 $OPT > dwf.64.64.64.64.4node.opt0
-srun ./benchmarks/Benchmark_dwf_fp32 --mpi 2.2.2.2 --grid 48.48.48.48 --accelerator-threads 8 --shm 2048 $OPT > dwf.48.48.48.48.4node.opt0
-
+export MPICH_RDMA_ENABLED_CUDA=0
+export MPICH_GPU_IPC_ENABLED=0
+export MPICH_GPU_EAGER_REGISTER_HOST_MEM=0
+export MPICH_GPU_NO_ASYNC_MEMCPY=1
 OPT="--comms-overlap --comms-concurrent --shm-mpi 1"
-srun ./benchmarks/Benchmark_dwf_fp32 --mpi 2.2.2.2 --grid 64.64.64.64 --accelerator-threads 8 --shm 2048 $OPT > dwf.64.64.64.64.4node.opt1
-srun ./benchmarks/Benchmark_dwf_fp32 --mpi 2.2.2.2 --grid 48.48.48.48 --accelerator-threads 8 --shm 2048 $OPT > dwf.48.48.48.48.4node.opt1
+srun ./benchmarks/Benchmark_ITT --mpi 2.1.1.2 --grid 64.64.64.64 --accelerator-threads 8 --shm 2048 $OPT > ITT.log
diff --git a/systems/Perlmutter/sourceme.sh b/systems/Perlmutter/sourceme.sh
index 9359dea9..6d09b1c9 100644
--- a/systems/Perlmutter/sourceme.sh
+++ b/systems/Perlmutter/sourceme.sh
@@ -1,4 +1,4 @@
 
 export CRAY_ACCEL_TARGET=nvidia80
 
-module load PrgEnv-gnu cpe-cuda cuda
+module load PrgEnv-gnu cpe-cuda cudatoolkit/11.4