@plasius/gpu-renderer 0.1.15 → 0.2.1

package/CHANGELOG.md

@@ -23,16 +23,40 @@ All notable changes to this project will be documented in this file.
 - **Security**
   - (placeholder)
 
-## [0.1.15] - 2026-06-04
+## [0.2.1] - 2026-06-06
 
 - **Added**
-  - (placeholder)
+  - Added deterministic wavefront reference helpers for primary-ray creation,
+    triangle-hit evaluation, nearest-hit selection, and environment-miss
+    fallback alongside regression tests for task-level acceptance coverage.
+  - Added ADR coverage for mesh BVH wavefront path tracing as the display-quality
+    baseline.
 
 - **Changed**
-  - (placeholder)
+  - Promoted the mesh BVH wavefront renderer source into the canonical
+    `src/wavefront-compute.js` module so source, typings, tests, and package
+    builds no longer drift behind generated artifacts.
+  - Changed wavefront mesh tracing to dispatch bounce intersection and
+    surface-resolution work from GPU-authored active-ray counts instead of the
+    fixed tile capacity after queue compaction.
+  - Changed the npm CD workflow to prepare package version and changelog
+    metadata locally, then publish the validated tag while persisting release
+    metadata through an automation PR branch instead of direct protected-main
+    pushes.
 
 - **Fixed**
-  - (placeholder)
+  - Fixed wavefront source ownership so display-quality mesh BVH code is not
+    hidden behind a legacy module split.
+  - Preserved the wavefront mode/workgroup constants, shader-source helper,
+    one-shot frame helper, and `renderFrame(...)` compatibility wrapper while
+    promoting the mesh BVH renderer source.
+  - Fixed a production performance regression where compacted continuation
+    queues still scheduled full-capacity per-bounce workgroups.
+  - Fixed low-sample wavefront surface traces so terminal non-emissive surface
+    collisions use an environment-derived ambient floor instead of a flat or
+    missing fallback.
+  - Fixed release metadata versioning, demo dependencies, and one-shot
+    wavefront cleanup blockers found during PR validation.
 
 - **Security**
   - (placeholder)
@@ -287,4 +311,4 @@ All notable changes to this project will be documented in this file.
 [0.1.11]: https://github.com/Plasius-LTD/gpu-renderer/releases/tag/v0.1.11
 [0.1.12]: https://github.com/Plasius-LTD/gpu-renderer/releases/tag/v0.1.12
 [0.1.14]: https://github.com/Plasius-LTD/gpu-renderer/releases/tag/v0.1.14
-[0.1.15]: https://github.com/Plasius-LTD/gpu-renderer/releases/tag/v0.1.15
+[0.2.1]: https://github.com/Plasius-LTD/gpu-renderer/releases/tag/v0.2.1

package/README.md

@@ -141,6 +141,10 @@ const renderer = await createWavefrontPathTracingComputeRenderer({
 renderer.renderOnce();
 ```
 
+Existing consumers that still call `renderFrame(...)` or
+`renderWavefrontPathTracingComputeFrame(...)` remain supported as compatibility
+wrappers around the canonical mesh renderer.
+
 Analytic scene objects remain available for debug fixtures:
 
 ```js
@@ -203,7 +207,12 @@ sampling, temporal accumulation, and better material PDFs are hardened.
 For static mesh scenes, the GPU acceleration build is submitted once and then
 reused by subsequent frames. Per-frame tracing writes one dynamic uniform slot
 per tile/sample or post-process pass and batches tile tracing, tile output,
-optional denoise, and presentation into a single command submission. WebGPU
+optional denoise, and presentation into a single command submission. After each
+primary-ray or compaction pass, the GPU writes the active-ray workgroup count
+into the counter buffer and the encoder copies it into an indirect-dispatch
+argument buffer. Intersection and surface-resolution passes therefore scale
+with active continuation rays instead of the maximum tile capacity, while still
+avoiding CPU readback between bounces. WebGPU
 still preserves ordering between dependent bounce passes, but the renderer no
 longer forces one CPU queue submission per tile/sample.
 Environment-light portals can additionally guide and gate sky/HDRI contribution
@@ -242,6 +251,11 @@ renderer.bindXrManager(xr, {
 - `createRayTracingRenderPlan(options)`
 - `createWavefrontPathTracingComputeRenderer(options)`
 - `createWavefrontPathTracingComputeConfig(options)`
+- `createWavefrontPathTracingComputeShaderSource(options?)`
+- `renderWavefrontPathTracingComputeFrame(options)`
+- `createWavefrontReferenceRay(config, options?)`
+- `intersectWavefrontReferenceTriangle(ray, triangle, options?)`
+- `traceWavefrontReferenceTriangles(config, ray, triangles, options?)`
 - `normalizeWavefrontMesh(input)`
 - `createWavefrontGpuMeshSource(meshes)`
 - `createWavefrontBvhSortStages(itemCount)`
@@ -251,6 +265,9 @@ renderer.bindXrManager(xr, {
 - `packWavefrontSceneObjects(sceneObjects, capacity?)`
 - `packWavefrontTriangles(triangles, capacity?)`
 - `packWavefrontBvhNodes(nodes, capacity?)`
+- `rendererWavefrontComputeMode`
+- `rendererWavefrontComputeWorkgroupSize`
+- `rendererWavefrontComputeStatsStride`
 - `bindRendererToXrManager(renderer, xrManager, options)`
 - `defaultRendererClearColor`
 - `rendererDebugOwner`
@@ -260,6 +277,11 @@ renderer.bindXrManager(xr, {
 - `rendererWorkerProfileNames`
 - `rendererWorkerManifests`
 
+The reference helpers mirror the renderer WGSL camera and triangle-hit math in
+deterministic JavaScript so tests and downstream tooling can validate primary
+ray generation, barycentrics, nearest-hit selection, and environment misses
+without standing up a WebGPU device.
+
 ## Demo
 
 Run the demo server from the repo root:
@@ -290,6 +312,8 @@ npm run pack:check
 ## Files
 
 - `src/index.js`: WebGPU renderer runtime and XR binding helper.
+- `src/wavefront-compute.js`: Canonical WebGPU mesh BVH wavefront renderer,
+  debug scene-object fixtures, and deterministic reference helpers.
 - `src/index.d.ts`: public API typings.
 - `tests/package.test.js`: unit tests for renderer lifecycle behavior.
 - `docs/design/worker-manifest-integration.md`: renderer frame-stage DAG model.