AI WebGPU Lab Compute Stress

Fluid Compute Readiness

`exp-fluid-webgpu-core` records a deterministic fluid and particle-grid baseline before real WGSL pressure solve kernels and atomics-heavy updates land.

The harness fixes particle count, grid size, workgroup size, pressure iterations, dispatch timing, divergence error, atomics note, and thermal note in one schema-aligned result.

Compute Controls

Probe capability first, then run the deterministic fluid simulation to export compute-stress metrics.

Simulation Viewport

Checks

  • Keep particle count, grid resolution, workgroup size, and pressure iterations fixed.
  • Record dispatch timing, divergence error, atomics contention, and thermal note before real fluid kernels land.
  • Use this readiness surface as the seed input for later compute stress, atomics, and particle-VFX tracks.

Metrics

Environment

Activity Log

    Schema-Aligned Result Draft

    {
  "status": "pending"
}

    What This Unlocks

    • First fluid compute raw JSON and screenshot capture path
    • Stable particle/grid metadata for future WGSL pressure solve integration
    • Reusable compute stress metrics for atomics, particles, and texture-streaming benchmarks
    Open Repository Read README View Results