Flipper performance

Scalapack xdlu

TIME     N  NB   P   Q  LU Time   Sol Time  MFLOP/S Residual  CHECK
---- ----- --- --- --- --------- --------- -------- -------- -------
WALL 21000  40   2   8   1395.52      2.09  4418.03 0.000213 PASSED

Streams2

 Smallest time delta is   0.00999999978
     Size  Iter     FILL      COPY     DAXPY       SUM
       30     5    800.0    1600.0    2400.0     320.0   33333.0
       43     5    800.0    1600.0    2400.0     320.0   23255.5
       61     5    800.0    1600.0    2399.9     320.0   16393.0
       88     5    800.0    1600.0    2400.0     320.0   11363.5
      126     5    800.0    3200.0    2400.0     320.0    7936.5
      180     5    533.3    3200.0    2400.0     320.0    3703.7
      258     5    799.9    3199.7    2399.7     320.0    3875.5
      368     5    799.9    3199.5    2399.7     320.0    2717.0
      527     5    800.0    3199.9    2400.0     320.0    1897.5
      754     5    533.2    1599.7    2399.5     319.9     884.0
     1079     5    799.8    1599.5    1599.5     319.9     926.5
     1545     5    799.7     639.8     799.7     319.9     647.0
     2210     5    532.8     639.3     799.1     319.7     301.3
     3163     5    533.1     639.7     799.6     319.8     210.7
     4525     5    399.1     638.6     798.2     319.3     110.3
     6475     5    398.9     638.2     797.7     319.1      77.0
     9266     5    531.3     637.5     796.9     318.7      71.7
    13258     5    530.3     636.4     681.8     318.2      50.0
    18971     5    531.2     531.2     597.6     318.7      35.0
    27146     5    396.3     396.3     475.6     317.1      18.3
    38844     5    396.2     317.0     365.7     317.0      12.7
    55582     5    389.1     259.4     291.8     222.3       8.7
    79532     5    265.1     227.2     280.7     198.8       4.2
   113802     5    221.1     221.1     273.1     154.8       2.4
   162840     5    223.3     208.4     275.9     156.3       1.7
   233008     5    213.0     213.0     279.6     149.1       1.1
   333411     5    222.3     205.2     285.8     148.2       0.8
   477079     5    218.1     218.1     269.4     152.7       0.6
   682653     5    218.5     198.6     273.1     136.5       0.4
   976810     5    195.4     208.4     275.8     156.3       0.3
  1397720     5    186.4     203.3     279.5     159.7       0.2
  2000000     5    200.0     200.0     266.7     145.5       0.1

Bandwidth

Ping test (P0 -> P1) --- blocking standard, typesize=2
---------

Length(bytes)    elapsed(s)  rate(Mbytes/s)   latency   iterations

            2     3.217843       0.326         6.137549   524288
            4     3.055362       0.686         5.827640   524288
            8     1.529763       1.371         5.835582   262144
           16     2.160834       1.941         8.242929   262144
           32     0.961611       4.362         7.336510   131072
           64     1.424276       5.890        10.866364   131072
          128     0.808042      10.381         0.000000    65536
          256     0.975313      17.202         0.000000    65536
          512     0.659568      25.437         0.000000    32768
         1024     0.785272      42.730         0.000000    32768
         2048     0.592653      56.617         0.000000    16384
         4096     0.977825      68.631         0.000000    16384
         8192     0.881250      76.152         0.000000     8192
        16384     0.822966      81.545         0.000000     4096
        32768     0.793670      84.555         0.000000     2048
        65536     0.784351      85.560         0.000000     1024
       131072     0.779129      86.133         0.000000      512
       262144     0.910226      73.728         0.000000      256
       524288     0.837323      80.147         0.000000      128
      1048576     0.818512      81.989         0.000000       64
      2097152     0.804363      83.431         0.000000       32
      4194304     0.797825      84.115         0.000000       16
      8388608     0.795604      84.350         0.000000        8
     16777216     0.790831      84.859         0.000000        4
Elapsed(s)  0.602128 Ping latency(us) for 0 byte message   5.742378
Elapsed(s)  0.601759 Ping latency(us) for 0 byte message   5.738853
Elapsed(s)  1.344159 Ping_Pong/2 latency(us) for 0 byte message   6.409487


Ping_Pong test (P0 -> P1 -> P0) --- blocking standard, typesize=2
---------

Length(bytes)    elapsed(s)  rate(Mbytes/s)   latency   iterations

Elapsed(s)  1.344054 Ping_Pong/2 latency(us) for 0 byte message   6.408986
            4     7.444004       0.282        14.198311   524288
            8     7.447085       0.563        14.204187   524288
           16     3.756296       1.117        14.329132   262144
           32     5.134555       1.634        19.586772   262144
           64     2.439353       3.439        18.610787   131072
          128     3.705299       4.528         0.000000   131072
          256     2.106637       7.964         0.000000    65536
          512     2.538962      13.216         0.000000    65536
         1024     1.737593      19.311         0.000000    32768
         2048     2.176442      30.834         0.000000    32768
         4096     1.534133      43.744         0.000000    16384
         8192     2.508369      53.508         0.000000    16384
        16384     2.316788      57.933         0.000000     8192
        32768     2.199389      61.025         0.000000     4096
        65536     2.135401      62.854         0.000000     2048
       131072     2.115161      63.455         0.000000     1024
       262144     2.100988      63.883         0.000000      512
       524288     1.793060      74.854         0.000000      256
      1048576     1.691116      79.366         0.000000      128
      2097152     1.647841      81.451         0.000000       64
      4194304     1.611978      83.263         0.000000       32
      8388608     1.596738      84.057         0.000000       16
     16777216     1.588280      84.505         0.000000        8
     33554432     1.582374      84.820         0.000000        4

Updated: 2024-03-21, 12:31