|
Peano
|
Data Structures | |
| class | PerformanceData |
Functions | |
| extract_grid_construction_times (performance_data_points) | |
| Returns a tuple of arrays to be plotted. | |
| extract_total_time_stepping_times (performance_data_points, max_cores_per_rank=0, verbose=False) | |
| extract_times_per_step (performance_data_points, max_cores_per_rank=0, verbose=False) | |
| Returns a tuple of arrays to be plotted. | |
| load_file_sequence (file_prefix, file_postfix="", solver_name="", verbose=False) | |
| Load all files that start with file_prefix and merge them into one instance of Dataset. | |
| postprocessing.PerformanceData.extract_grid_construction_times | ( | performance_data_points | ) |
Returns a tuple of arrays to be plotted.
Definition at line 385 of file PerformanceData.py.
| postprocessing.PerformanceData.extract_times_per_step | ( | performance_data_points, | |
max_cores_per_rank = 0, |
|||
verbose = False |
|||
| ) |
Returns a tuple of arrays to be plotted.
This is a rather simple routine. Its big USP is that it accepts an unsorted list of measurements and returns properly sorted measurements.
performance_data_points: [exahype2.postprocessing.PerformanceData] Set of measurements.
max_cores_per_rank: Integer Should be set to -1 if you analyse single node data. Should be set to 0 if you have only one core count measurement per rank count. Should be set to something bigger than 0 if you scale over both the mpi rank count and the core count.
Definition at line 432 of file PerformanceData.py.
| postprocessing.PerformanceData.extract_total_time_stepping_times | ( | performance_data_points, | |
max_cores_per_rank = 0, |
|||
verbose = False |
|||
| ) |
Definition at line 405 of file PerformanceData.py.
| postprocessing.PerformanceData.load_file_sequence | ( | file_prefix, | |
file_postfix = "", |
|||
solver_name = "", |
|||
verbose = False |
|||
| ) |
Load all files that start with file_prefix and merge them into one instance of Dataset.
Definition at line 475 of file PerformanceData.py.