使用一个节点提交32个核心程序可以正常运行,使用两个节点提交64核心就会出现上图错误,请问这个问题怎么解决?
试下加入节点列表,mpirun -np $np --hostfile hostfile-dl-$SLURM_JOB_ID --bind-to none,请参考/public/software/apps/DeepLearning/Benchmark/PyTorch/detection_vision/mpi_slurm.sbatch
试下加入节点列表,mpirun -np $np --hostfile hostfile-dl-$SLURM_JOB_ID --bind-to none,请参考/public/software/apps/DeepLearning/Benchmark/PyTorch/detection_vision/mpi_slurm.sbatch