Well, at the moment, sigma.js only handles rendering. The main two CPU-heavy parts are indexing the data into lower-level data structures that can be transferred to the WebGL programs, and the quad-tree indexation. We have no plan to move the quad-tree indexation into WebGPU right now, and the indexation part (ie. translating graphology data to GPU-compatible data) would still be an issue.
Well, at the moment, sigma.js only handles rendering. The main two CPU-heavy parts are indexing the data into lower-level data structures that can be transferred to the WebGL programs, and the quad-tree indexation. We have no plan to move the quad-tree indexation into WebGPU right now, and the indexation part (ie. translating graphology data to GPU-compatible data) would still be an issue.