想要调用64个进程，每个进程分配给一个核，一共平均分配给2个numa。作业脚本内容为：

export OMP NUM THREADS=1 

np=64 

export UCX RC VERBS ROCE LOCAL SUBNET=Y 

export UCX UD VERBS ROCE LOCAL SUBNET=Y 

export OMP PROC BIND=close 

export OMP PLACES=cores 

export KML BLAS THREAD TYPE=omp 

export BLAS VERBOSE=2 

mpirun -np 64 --map-by ppr:32:numa:PE=1 $EXE

运行会报错：

request was made to bind to that would result in binding more processes than cpus on a resource: 

Bind to: CORE

#processes: 2

#cpus: 1

You can override this protection by adding the "overload-allowed"option to your binding directive.