Sample 1 - IMB-MPI1 PingPong Allreduce

The following example shows the results of the PingPong and Allreduce benchmark:

<..>  np 2 IMB-MPI1 PingPong Allreduce
#---------------------------------------------------
# Intel ® MPI Benchmark Suite V3.2, MPI1 part
#---------------------------------------------------
# Date                  : Thu Sep  4 13:20:07 2008
#  Machine               : x86_64
#  System                : Linux
#  Release               : 2.6.9-42.ELsmp
#  Version               : #1 SMP Wed Jul 12 23:32:02 EDT 2006
#  MPI Version           : 2.0
#  MPI Thread Environment: MPI_THREAD_SINGLE
 
 
#  New default behavior from Version 3.2 on:
#  the number of iterations per message size is cut down
# dynamically when a certain run time (per message size sample)
# is expected to be exceeded. Time limit is defined by variable
#  SECS_PER_SAMPLE (=> IMB_settings.h)
#  or through the flag => -time
 
#  Calling sequence was:
 
#  ./IMB-MPI1 PingPong Allreduce
 
#  Minimum message length in bytes:   0
#  Maximum message length in bytes:   4194304
#
#  MPI_Datatype                   :   MPI_BYTE
#  MPI_Datatype for reductions    :   MPI_FLOAT
#  MPI_Op                         :   MPI_SUM 
#
#
 
#  List of Benchmarks to run:
 
#  PingPong
#  Allreduce
 
#---------------------------------------------------
#  Benchmarking PingPong
#  #processes = 2
#---------------------------------------------------
 #bytes  #repetitions      t[μsec]   Mbytes/sec
      0         1000         ..       .. 
      1         1000        
      2         1000        
      4         1000        
      8         1000        
     16         1000        
     32         1000        
     64         1000        
    128         1000        
    256         1000       
    512         1000       
   1024         1000       
   2048         1000       
   4096         1000       
   8192         1000       
  16384         1000        
  32768         1000       
  65536          640      
 131072          320      
 262144          160      
 524288           80      
1048576           40     
2097152           20     
4194304           10     
#-------------------------------------------------------
#  Benchmarking Allreduce
#  ( #processes = 2 )
#-------------------------------------------------------
 #bytes  #repetitions  t_min[μsec]  t_max[μsec]   t_avg[μsec]
      0         1000         ..          ..          ..
      4         1000       
      8         1000       
     16         1000       
     32         1000       
     64         1000       
    128         1000       
    256         1000       
    512         1000       
   1024         1000       
   2048         1000       
   4096         1000      
   8192         1000      
  16384         1000      
  32768         1000     
  65536          640     
 131072          320     
 262144          160     
 524288           80    
1048576           40    
2097152           20   
4194304           10    
 
# All processes entering MPI_Finalize