grid-14-virt-clusters

Applications of on-demand virtual clusters to high performance computing
git clone https://git.igankevich.com/grid-14-virt-clusters.git
Log | Files | Refs

imb.4.out (3326B)


      1 #------------------------------------------------------------
      2 #    Intel (R) MPI Benchmarks 4.0, MPI-1 part    
      3 #------------------------------------------------------------
      4 # Date                  : Tue Sep  2 19:54:05 2014
      5 # Machine               : x86_64
      6 # System                : Linux
      7 # Release               : 3.10.0-123.4.2.el7.x86_64
      8 # Version               : #1 SMP Mon Jun 30 16:09:14 UTC 2014
      9 # MPI Version           : 2.1
     10 # MPI Thread Environment: 
     11 
     12 # New default behavior from Version 3.2 on:
     13 
     14 # the number of iterations per message size is cut down 
     15 # dynamically when a certain run time (per message size sample) 
     16 # is expected to be exceeded. Time limit is defined by variable 
     17 # "SECS_PER_SAMPLE" (=> IMB_settings.h) 
     18 # or through the flag => -time 
     19   
     20 
     21 
     22 # Calling sequence was: 
     23 
     24 # /opt/IMB-MPI1 -npmin 16 Exchange
     25 
     26 # Minimum message length in bytes:   0
     27 # Maximum message length in bytes:   4194304
     28 #
     29 # MPI_Datatype                   :   MPI_BYTE 
     30 # MPI_Datatype for reductions    :   MPI_FLOAT
     31 # MPI_Op                         :   MPI_SUM  
     32 #
     33 #
     34 
     35 # List of Benchmarks to run:
     36 
     37 # Exchange
     38 
     39 #-----------------------------------------------------------------------------
     40 # Benchmarking Exchange 
     41 # #processes = 16 
     42 #-----------------------------------------------------------------------------
     43        #bytes #repetitions  t_min[usec]  t_max[usec]  t_avg[usec]   Mbytes/sec
     44             0         1000        93.42        94.46        93.74         0.00
     45             1         1000        95.19        96.00        95.56         0.04
     46             2         1000        94.81        95.88        95.21         0.08
     47             4         1000        94.28        95.24        94.57         0.16
     48             8         1000        96.14        96.80        96.54         0.32
     49            16         1000        95.72        97.06        96.17         0.63
     50            32         1000        91.76        92.58        92.12         1.32
     51            64         1000        91.43        93.96        91.85         2.60
     52           128         1000        99.72       100.55       100.12         4.86
     53           256         1000        99.30        99.89        99.61         9.78
     54           512         1000       102.01       102.72       102.29        19.01
     55          1024         1000       116.22       117.00       116.49        33.39
     56          2048         1000       165.38       166.65       165.76        46.88
     57          4096         1000       271.67       272.10       271.90        57.42
     58          8192         1000       364.81       366.07       365.49        85.37
     59         16384         1000       481.13       482.86       481.70       129.44
     60         32768         1000      1398.96      1401.24      1400.14        89.21
     61         65536          640      9540.67      9556.17      9549.20        26.16
     62        131072          320      6603.38      6630.64      6617.28        75.41
     63        262144          160     15067.02     15216.04     15138.35        65.72
     64        524288           80     23230.17     23887.74     23599.77        83.72
     65       1048576           40     40223.08     41961.70     41140.18        95.33
     66       2097152           20     74094.55     80883.94     77595.19        98.91
     67       4194304           10    143267.89    165591.60    156278.59        96.62
     68 
     69 
     70 # All processes entering MPI_Finalize
     71