Boolean MULTICORE Modifier

Dear Blender Development Community,

I would like to propose an enhancement for the Boolean modifier that aims to significantly improve its performance. As many of you are aware, the current Boolean modifier operates in two modes: FAST and EXACT. Both modes, however, suffer from performance issues due to the limitation of single-threaded execution. This is primarily because the current implementation requires the processing of the entire space of the two objects involved in the Boolean operation, which inherently restricts the ability to parallelize the computations.

To address this limitation, I suggest a novel approach leveraging a 3D Region of Interest (ROI) segmentation strategy. Here's a detailed breakdown of the proposed enhancement:

Problem Analysis Current Implementation: The Boolean operations are performed on the entire spatial domain of the objects, making it difficult to divide the workload across multiple CPU cores effectively. Single-threaded Bottleneck: The constraint of single-threaded execution results in suboptimal performance, particularly for complex models. Proposed Solution Encapsulation in Bounding Cubes:

Encapsulate both objects involved in the Boolean operation within their respective bounding cubes. Determine a common 3D ROI that encompasses the interaction space of these bounding cubes. ROI Segmentation:

Divide the 3D ROI into smaller sub-cubes (chunks). The number of chunks should ideally correspond to the number of available CPU cores to maximize parallelism. Parallel Processing of Chunks:

Assign each chunk to a separate CPU core. Each core will independently execute the Boolean operation within its assigned chunk. Aggregation of Results:

Once all cores complete their computations, aggregate the results from all chunks. Sum the partial results to form the final Boolean operation output. Implementation Details Data Structure: Use an efficient data structure to manage the division of the 3D ROI into sub-cubes and the subsequent aggregation of results. Thread Management: Implement a robust thread management system to handle the parallel execution and synchronization of chunks. Edge Handling: Ensure proper handling of edge cases where the Boolean operation spans multiple chunks to maintain the integrity of the final model. Expected Benefits Performance Improvement: Significant reduction in computation time for Boolean operations due to multi-core parallel processing. Scalability: Enhanced scalability, making it feasible to handle more complex models with intricate Boolean operations. User Experience: Improved user experience with faster and more responsive Boolean modifications. I believe this approach can leverage modern multi-core processors effectively and bring a substantial performance boost to Blender's Boolean modifier. I look forward to discussing this proposal further and collaborating with the community to refine and implement this enhancement.

Best regards, Bartosz Startek