Axiom provides a GPU‑accelerated, sparse volumetric fluid solver that runs entirely on modern graphics hardware, eliminating the need for CPU fallback. Integrated as native Houdini nodes and supporting CUDA, OpenCL, Metal, and OpenVDB, it enables VFX and game artists to simulate from 25 M to 250 M voxels with reduced memory usage, near‑real‑time previews, and scalable cloud GPU farm execution.
Funding
Funding not disclosed
Founders
Product
Problem
VFX and game studios often rely on CPU‑based fluid solvers that require dense grids and extensive compute time, limiting the speed of iteration and the scale of simulations. Traditional pipelines also involve complex data transfers between simulation and rendering stages, which can hinder production efficiency.
Solution
Axiom is a sparse, GPU‑accelerated volumetric fluid solver that runs entirely on modern graphics hardware, eliminating the need for CPU fallback. It leverages industry‑standard APIs such as CUDA, OpenCL, Metal, and OpenVDB to ensure broad compatibility across workstations and cloud GPU farms. Integrated as native Houdini nodes, Axiom fits seamlessly into existing Houdini pipelines, preserving familiar workflows while delivering higher throughput. The solver’s sparse data structures enable simulations ranging from low‑resolution background effects to high‑resolution hero shots (e.g., 25 M to 250 M voxels) without excessive memory overhead. By offloading the full simulation pipeline to the GPU, artists can obtain near‑real‑time previews and iterate more rapidly while maintaining full control over simulation parameters and output quality.
Target Audience
The primary customers are VFX and game simulation artists, technical directors, and pipeline engineers who use Houdini for fluid and pyro effects in film, television, and interactive media productions.
Features
- True sparse simulation engine that computes only active fluid cells, reducing memory usage and GPU load.
- Full‑GPU execution using CUDA, OpenCL, and Metal, with OpenVDB support for seamless data exchange.
- Native Houdini node integration, preserving existing procedural networks and parameter linking.
- Scalable resolution handling from 25 M to 250 M voxels, enabling both background and hero‑level effects.
- Cloud rendering support via GridMarkets GPU farm services for on‑demand high‑performance compute.
- Real‑time preview and interactive feedback loop to accelerate artistic iteration.
- Compatibility with standard VFX pipelines, including export to Alembic, OpenVDB caches, and render engines.