[[wg:dynamo:Performance_results|Back to performance benchmark lists]] \\

===== compile options =====

F90OPTFLAGS = -r8 -i4 -ftz -IPF_fma -IPF_fltacc -WB -O3 -xhosts

===== Definition of columns =====

^  name  ^    ^
|  # of Cores  |  Number of used CPU cores  |
|  # of Processes  |  Number of MPI processes  |
|  # of Threads  |  Number of threads for each process  |
|  $N_{c}$  |  Truncation lavel for Chebyshev polynomials  |
|  $l_{max}$  |  Truncation lavel for spherical harmonincs  |
|  $(N_{r},N_{\theta},N_{\phi})$  |  Nuber of grids in spherical coordinate  |
|  Elapsed  |  Elapsed (wall clock time) for one time step  |
|  Legendre  |  Elapsed (wall clock time) for Legendre transform  |
|  Implicit  |  Elapsed (wall clock time) for linear calculation  |
|  Efficiency  |  Parallel efficiency  |
|  SUs  |  Service unit for $10^{4}$ time steps (Core hours)  |

===== Single Processor Result =====

^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^  Elapsed  ^  Legendre  ^  Implicit  ^  LUdecomp  ^  SUs  ^
|  71  |  47  |  (73,72,144)  |  0.96659  |  0.57970  |  0.13313  |  0.010014  |  2.6849  |

===== Strong Scaling Results =====

^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^
|  191  |  255  |  (193,384,768)  |

^  # of Cores  ^  # of Processes  ^  # of SMP  ^   Elapsed  ^  Legendre  ^  Implicit  ^  Efficiency  ^  SUs  ^
|  16  |  16  |  1  |  7.8559  |  3.1132  |  1.0993  |  1.0  |  349.151  |
|  32  |  32  |  1  |  4.4581  |  1.5484  |  0.67073  |  0.881082  |  396.276  |
|  64  |  64  |  1  |  3.4032  |  0.77098  |  0.68621 |  0.577097  |  605.013  |
|  128  |  128  |  1  |  1.0696  |  0.37643  |  0.14921  |  0.918089  |  380.302  |


^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^
|  255  |  511  |  (257,768,1536)  |

^  # of Cores  ^  # of Processes  ^  # of SMP  ^   Elapsed  ^  Legendre  ^  Implicit  ^  Efficiency  ^  SUs  ^
|  64  |  64  |  1  |  13.018  |  4.7327  |  1.9132  |  1.0  |  414.015  |
|  128  |  128  |  1  |  8.7973  |  2.3534  |  1.7398  |  0.555322  |  745.541  |
|  256  |  256  |  1  |  8.678  |  1.1378  |  4.3325  |  0.412058  |  1004.75  |


{{wg:dynamo:Performance_results:Gary_Latest:ucsc2_elapsed.png?480}}\\
Elapsed (wall clock) time for the strong scaling. Number of OpenMP threads are shown by the numbers. Ideal scaling is plotted by dotted line. 

{{wg:dynamo:Performance_results:Gary_Latest:ucsc2_efficiency.png?480}}\\
Parallel Efficiency for the strong scaling.


===== Weak Scaling Results =====

^  # of Cores  ^  # of Processes  ^  # of SMP  ^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^  Elapsed  ^  Legendre  ^  Implicit  ^  SUs  ^
|  2  |  2  |  1  |  255  |  63  |  (257,96,192)  |  3.3257  |  1.8103  |  0.47442  |  147.809  |
|  8  |  8  |  1  |  255  |  127  |  (257,192,384)  |  4.0801  |  1.8754  |  0.51211  |  181.338  |
|  32  |  32  |  1  |  255  |  255  |  (257,384,768)  |  5.8172  |  2.0489  |  0.8497  |  517.084  |
|  128  |  128  |  1  |  255  |  511  |  (257,768,1536)  |  9.5023  |  2.3534  |  1.7398  |  3378.6  |
|   |   |   |   |   |    |   |   |   |   |
|  3  |  3  |  1  |  255  |  63  |  (257,96,192)  |  2.3095  |  1.1811  |  0.3029  |  102.644  |
|  9  |  9  |  1  |  255  |  127  |  (257,192,384)  |  3.5408  |  1.7574  |  0.46589  |  157.369  |
|  33  |  33  |  1  |  255  |  255  |  (257,384,768)  |  5.528  |  1.7904  |  0.77925  |  737.067  |
|  129  |  129  |  1  |  255  |  511  |  (257,768,1536)  |  8.9802  |  1.9042  |  1.7902  |  3592.1  |

{{wg:dynamo:Performance_results:Gary_Latest:ucsc2_weak_sph.png?480}}\\
Elapsed time for the weak scaling in the horizontal resolution. Scaling of $O(Ncore^{1/2})$ is plotted by dotted line.

^  # of Cores  ^  # of Processes  ^  # of SMP  ^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^  Elapsed  ^  Legendre  ^  Implicit  ^  SUs  ^
|  16  |  16  |  1  |  31  |  511  |  (33,768,1536)  |  5.1154  |  2.5566  |  0.49331  |  227.351  |
|  32  |  32  |  1  |  63  |  511  |  (65,768,1536)  |  6.0620  |  2.4221  |  0.76595  |  538.844  |
|  64  |  64  |  1  |  127  |  511  |  (129,768,1536)  |  6.9065  |  2.3967  |  1.1075  |  1227.82  |
|  128  |  128  |  1  |  255  |  511  |  (257,768,1536)  |  9.5023  |  2.3534  |  1.7398  |  3378.6  |
|   |   |   |   |   |    |   |   |   |   |
|  17  |  17  |  1  |  31  |  511  |  (33,768,1536)  |  4.2943  |  2.4266  |  0.40772  |  381.716  |
|  33  |  33  |  1  |  63  |  511  |  (65,768,1536)  |  5.6936  |  2.2617  |  0.67252  |  759.147  |
|  65  |  65  |  1  |  127  |  511  |  (129,768,1536)  |  6.7351  |  2.1191  |  1.0571  |  1496.69  |
|  129  |  129  |  1  |  255  |  511  |  (257,768,1536)  |  8.9802  |  1.9042  |  1.7902  |  3592.08  |

{{wg:dynamo:Performance_results:Gary_Latest:ucsc2_weak_r.png?480}}\\
Elapsed time for the weak scaling in the radial resolution. Scaling of $O(Ncore)$ is plotted by dotted line.


[[wg:dynamo:Performance_results|Back to performance benchmark lists]] \\
[[wg:dynamo:Performance_results:Gary_Latest:files|files]]