Introduction to nPB Installation and Operation

NPB is called NAS parallel benchmark. Like linpack, NPB is the most commonly used benchmark program for Parallel computers. From
Http://www.nas.nasa.gov/Resources/Software/npb.html
Download the desired version. This article takes npb2.4 as an example to briefly introduce the installation and operation of nPB.
1. Download npb2.4.tar.gz and decompress it.
2. Cp conf/make. Def. template CONF/make. Def
3. Modify the conf/make. Def file:
Line 40: fmpi_lib =-L/data2/software/mpich-gcc/lib-lmpich # the connector will find the mpich library here.
Row 45: fmpi_inc =-I/data2/software/mpich-gcc/include # MPIF. h under this path
Row 80: mpicc = mpicc
Row 82: Clink = mpicc
Row 87: cmpi_lib =-L/data2/software/mpich-gcc/lib/-lmpich

Row 93: cmpi_inc =-I/data2/software/mpich-gcc/include # MPI. h under this path
4. Compile
Take the is program as an example, execute in the decompressed NPB2.4-MPI directory:
Make is class = A nprocs = 16 # The problem scale is a, and the number of processors is 16
The is. A.16 executable file is generated under the NPB2.4-MPI/bin/directory. You can also compile all programs at one time.
5. Run
If torque (mpiexec or mpirun) is installed on the machine (if not, the script is named test. Sh), write the following script ):
#! /Bin/sh

# PBS-N is_test // specify the Task Name
# PBS-O is_test.log // specify the standard output
# PBS-e is_test.err // specified standard error
# PBS-Q batch // specify the task queue
# PBS-l nodes = 16 // specify the number of processes

CD/home/XX/npb2.4/NPB2.4-MPI/bin
Echo Time is 'date'
Echo directory is $ pwd
Echo this job runs on following nodes:
Cat $ pbs_nodefile
Nprocs = 'wc-L <$ pbs_nodefile'
Echo this job has been allocated $ nprocs nodes
Mpirun-machinefile $ pbs_nodefile-NP $ nprocs./is. A.16 // run

Run: qsub test. Sh
Run: qstat to view the task status.
After the execution is complete, open is_test.log to get the result.