← All resources

Kiln: WebGPU-Native Out-of-Core Volume Rendering for Multi-GB Datasets

Auto-discovered

Blog Advanced

Martin Panknin

Visit resource ↗

Description

Kiln: WebGPU-Native Out-of-Core Volume Rendering for Multi-GB Datasets is a DEV Community article by Martin Panknin introducing a browser volume renderer for data larger than GPU memory. Kiln targets scientific and graphics developers experimenting with WebGPU compute.

The system streams multi-gigabyte brick pyramids over HTTP into a fixed VRAM page cache and ray-marches with compute shaders. It supports LOD bricking, virtual texture indirection, direct volume rendering, MIP, and isosurface modes, Zarr-backed datasets, and interactive demos with published performance metrics.

Advanced developers comfortable with WebGPU, volume data layouts, and paging strategies are the audience. Large-dataset visualization experience helps when tuning cache sizes.

The article and linked open-source project are free on dev.to and GitHub. WebGPU volume rendering posts are niche but appear in scientific visualization and graphics hack discussions.