MatchingData.computation_schedule#
- MatchingData.computation_schedule(matching_method='FLCSphericalMask', max_cores=1, use_gpu=False, pad_fourier=False, pad_target_edges=False, analyzer_method=None, available_memory=None, max_splits=256)[source]#
Computes a parallelization schedule for a given template matching operation.
- Parameters:
- matching_methodstr
Matching method to use, default “FLCSphericalMask”.
- max_coresint, optional
Maximum number of CPU cores to use, default 1.
- use_gpubool, optional
Whether to utilize GPU acceleration, default False.
- pad_fourierbool, optional
Apply Fourier padding, default False.
- pad_target_edgesbool, optional
Apply padding to target edges, default False.
- analyzer_methodstr, optional
Method used for score analysis, default None.
- available_memoryint, optional
Available memory in bytes. If None, uses all available system memory.
- max_splitsint, optional
Maximum number of splits to consider, default 256.
- Returns:
- target_splitsdict
Optimal splits for each axis of the target tensor
- scheduletuple
(n_outer_jobs, n_inner_jobs_per_outer) defining the parallelization schedule