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

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

F90OPTFLAGS = -r8 -i4 -ftz -IPF_fma -IPF_fltacc -WB -O2

===== 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  |
|  Nonlinear  |  Elapsed (wall clock time) for nonlinear terms (including communications)  |
|  Solver  |  Elapsed (wall clock time) for linear calculation  |
|  Comm.  |  Elapsed (wall clock time) for data communication  |
|  Efficiency  |  Parallel efficiency  |
|  SUs  |  Service unit for $10^{4}$ time steps (Core hours)  |

===== Two Processor Result =====

^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^  Elapsed  ^  Nonlinear  ^  Solver  ^  Comm.  ^  SU  ^
|    |    |   |   |   |   |   |

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

^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^
|  192  |  170  |  (193,256,256)  |

^  # of Cores  ^  # of Processes  ^  # of SMP  ^  Elapsed  ^  Nonlinear  ^  Solver  ^  Comm.  ^  Efficiency  ^  SUs  ^
|  32  |  32  |  1  |  0.719657  |  0.518729  |  0.200927  |  0.182378  |  1  |  63.9695  |
|  64  |  64  |  1  |  0.545835  |  0.336805  |  0.209030  |  0.21079  |  0.659226  |  97.0373  |
|  128  |  128  |  1  |  0.285552  |  0.19960  |  0.085952  |  0.174127  |  0.630058  |  101.529  |


^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^
|  256  |  341  |  (257,512,512)  |

^  # of Cores  ^  # of Processes  ^  # of SMP  ^  Elapsed  ^  Nonlinear  ^  Solver  ^  Comm.  ^  Efficiency  ^  SUs  ^
|  86  |  86  |  1  |  2.47102  |  1.88142  |  0.589604  |  0.379399  |  1  |  658.939  |
|  128  |  128  |  1  |  2.32883  |  1.91045  |  0.418383  |  0.857392  |  0.712896  |  828.03  |
|  129  |  129  |  1  |  2.09683  |  1.68765  |  0.409184  |  0.663364  |  0.785635  |  838.734  |
|  256  |  256  |  1  |  1.41293  |  1.13385  |  0.279076  |  0.675334  |  0.58751  |  1004.75  |
|  257  |  257  |  1  |  1.33908  |  1.06534  |  0.273739  |  0.621596  |  0.61991  |  1011.75  |


{{wg:dynamo:Performance_results:Gary:Gary_1_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:Gary_1_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  ^  Nonlinear  ^  Solver  ^  Comm.  ^  SUs  ^
|  4  |  4  |  1  |  256  |  42  |  (257,64,64)  |  0.322194  |  0.165516  |  0.156679  |  0.0525075  |  14.3197  |
|  16  |  16  |  1  |  256  |  85  |  (257,128,128)  |  0.406717  |  0.204322  |  0.202394  |  0.0716188  |  18.0763  |
|  64  |  64  |  1  |  256  |  170  |  (257,256,256)  |  0.743692  |  0.44779  |  0.295902  |  0.312471  |  132.212  |
|  256  |  256  |  1  |  256  |  341  |  (257,512,512)  |  1.41293  |  1.13385  |  0.279076  |  0.675334  |  1004.75  |
|    |    |    |    |    |    |    |    |    |    |
|  5  |  5  |  1  |  256  |  42  |  (257,64,64)  |  0.252394  |  0.127396  |  0.124998  |  0.0395718  |  11.2175  |
|  17  |  17  |  1  |  256  |  85  |  (257,128,128)  |  0.359973  |  0.180849  |  0.179124  |  0.0811399  |  31.9976  |
|  65  |  65  |  1  |  256  |  170  |  (257,256,256)  |  0.639472  |  0.363557  |  0.275915  |  0.227634  |  142.105  |
|  257  |  257  |  1  |  256  |  341  |  (257,512,512)  |  1.33908  |  1.06534  |  0.273739  |  0.621596  |  1011.75  |

{{wg:dynamo:Performance_results:Gary:Gary_1_weak_sph.png?480}}\\
Elapsed time for the weak scaling in the horizontal directions. Scaling of $O(Ncore^{1/2})$ (ideal scaling for Legendre transform) is plotted by dotted line.


^  # of Cores  ^  # of Processes  ^  # of SMP  ^  $N_{c}$  ^  $l_{max}$  ^  $(N_{r},N_{\theta},N_{\phi})$  ^  Elapsed  ^  Nonlinear  ^  Solver  ^  Comm.  ^  SUs  ^
|  8  |  8  |  1  |  16  |  341  |  (17,512,512)  |  1.28171  |  1.20689  |  0.0748194  |  0.372245  |  56.9651  |
|  16  |  16  |  1  |  32  |  341  |  (33,512,512)  |  1.52159  |  1.42569  |  0.0958951  |  0.315524  |  67.6261  |
|  32  |  32  |  1  |  64  |  341  |  (65,512,512)  |  1.61663  |  1.45842  |  0.158208  |  0.389299  |  143.7  |
|  64  |  64  |  1  |  128  |  341  |  (129,512,512)  |  1.75436  |  1.50559  |  0.248772  |  0.453414  |  311.887  |
|  128  |  128  |  1  |  256  |  341  |  (257,512,512)  |  2.32883  |  1.91045  |  0.418383  |  0.857392  |  828.03  |
|    |    |    |    |    |    |    |    |    |    |    |
|  9  |  9  |  1  |  16  |  341  |  (17,512,512)  |  1.05288  |  0.993848  |  0.0590345  |  0.207514  |  46.7948  |
|  17  |  17  |  1  |  32  |  341  |  (33,512,512)  |  1.03108  |  0.94969  |  0.0813889  |  0.194925  |  91.6515  |
|  33  |  33  |  1  |  64  |  341  |  (65,512,512)  |  1.25166  |  1.10257  |  0.149091  |  0.218563  |  166.888  |
|  65  |  65  |  1  |  128  |  341  |  (129,512,512)  |  1.43519  |  1.19399  |  0.241198  |  0.262446  |  318.931  |
|  129  |  129  |  1  |  256  |  341  |  (257,512,512)  |  2.09683  |  1.68765  |  0.409184  |  0.663364  |  838.734  |

{{wg:dynamo:Performance_results:Gary:Gary_1_weak_r.png?480}}\\
Elapsed time for the weak scaling in the radial direction. Scaling of $O(Ncore^{1/2})$ (ideal scaling for Legendre transform) is plotted by dotted line.


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