Vertex layout of a GPU-simulated system's position stream (§36
simulation: "gpu", R-31 wiring): the simulation's flat x,y,z storage
buffer — wgpu-particle-simulation.ts's positions, allocated with
VERTEX for exactly this bind — read per instance at the same
@location(1) the interleaved stream feeds, 12-byte stride.
Same shader module, different plumbing: the WGSL above never changes,
because a location is a name and this layout merely re-sources it — which
is what makes the GPU variant one pipeline-cache entry
(gpuInstances: true, wgpu-pipeline-cache.ts) and zero new WGSL.
Vertex layout of a GPU-simulated system's position stream (§36
simulation: "gpu", R-31 wiring): the simulation's flat x,y,z storage buffer —wgpu-particle-simulation.ts'spositions, allocated withVERTEXfor exactly this bind — read per instance at the same@location(1)the interleaved stream feeds, 12-byte stride.Same shader module, different plumbing: the WGSL above never changes, because a location is a name and this layout merely re-sources it — which is what makes the GPU variant one pipeline-cache entry (
gpuInstances: true,wgpu-pipeline-cache.ts) and zero new WGSL.