WebGPU RT: wavefront tracer core (GENERATE/PREP/TRACE/SHADE/RESOLVE)

Replace the megakernel @compute entry with five wavefront kernels sharing
one module, connected by GPU ray/hit/payload buffers and a GPU-driven
indirect bounce loop:

  GENERATE -> (PREP -> TRACE -> SHADE) x maxDepth -> RESOLVE

- TRACE contains zero user code (pure _rtwTraverseTlas/Blas, opaque-only).
- PREP publishes dispatchWorkgroupsIndirect args from the live ray count;
  the indirect-args buffer lives in its own bind group so it is never
  bound read-write in the same dispatch that consumes it as INDIRECT.
- New emit/accumulate API: rtEmitPrimaryRay / rtEmitRay / rtAccumulate,
  plus an optional user Resolve stage (tonemap hook; identity by default).
- Per-pass WfParams via a dynamic-offset uniform ring (curIsA/bounce vary
  between passes within one submit).
- Payload-typed wfPayload binding emitted in the codegen region after the
  user's struct Payload; payload travels with each ray (2*W*H slots).
- Request maxBufferSize / maxStorageBufferBindingSize / maxComputeWorkgroups
  PerDimension so the W*H-sized work buffers fit past the 128MB baseline.

VulkanTriangle ported to the new API and renders bit-identical to the
megakernel baseline at maxDepth=1.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

examples/VulkanTriangle/closesthit.wgsl

@@ -1,6 +1,9 @@
-// WebGPU port of closesthit.glsl. Library concatenates this BEFORE the
-// library helpers, so `Payload` declared here is visible to traceRay,
-// runClosestHit, the mega-switch, and the user's raygen source.
+// Payload declared here so the WGSL assembler sees it before the wfPayload
+// binding, the SHADE dispatch, and the raygen source.
+//
+// Wavefront model: closesthit_main runs in SHADE and accumulates the
+// pixel's color directly (rtAccumulate) instead of writing a payload that
+// raygen reads back.
 
 struct Payload {
     color: vec3<f32>,
@@ -8,5 +11,5 @@ struct Payload {
 
 fn closesthit_main(ray: RayDesc, hit: HitInfo, payload: ptr<function, Payload>) {
     let bary = vec3<f32>(1.0 - hit.attribs.x - hit.attribs.y, hit.attribs.x, hit.attribs.y);
-    (*payload).color = bary;
+    rtAccumulate(bary);
 }

examples/VulkanTriangle/miss.wgsl

@@ -1,5 +1,6 @@
-// WebGPU port of miss.glsl.
+// Wavefront miss: runs in SHADE for rays that hit nothing. Accumulate the
+// white background directly.
 
 fn miss_main(ray: RayDesc, payload: ptr<function, Payload>) {
-    (*payload).color = vec3<f32>(1.0, 1.0, 1.0);
+    rtAccumulate(vec3<f32>(1.0, 1.0, 1.0));
 }

examples/VulkanTriangle/raygen.wgsl

@@ -1,11 +1,12 @@
-// WebGPU port of raygen.glsl. Mirrors the pinhole camera setup — the
-// Payload type is declared in closesthit.wgsl (concatenated earlier).
+// WebGPU wavefront raygen. Runs in GENERATE: compute the pinhole camera
+// ray and emit it as the pixel's primary ray. Shading happens later in
+// SHADE (closesthit/miss). The Payload type is declared in closesthit.wgsl.
 
 fn raygen_main(gid: vec3<u32>) {
-    if (gid.x >= hdr.surfaceW || gid.y >= hdr.surfaceH) { return; }
+    if (gid.x >= wfParams.surfaceW || gid.y >= wfParams.surfaceH) { return; }
 
     let pixel = vec2<f32>(f32(gid.x), f32(gid.y));
-    let resolution = vec2<f32>(f32(hdr.surfaceW), f32(hdr.surfaceH));
+    let resolution = vec2<f32>(f32(wfParams.surfaceW), f32(wfParams.surfaceH));
     let uv  = (pixel + vec2<f32>(0.5)) / resolution;
     let ndc = uv * 2.0 - vec2<f32>(1.0);
 
@@ -23,17 +24,11 @@ fn raygen_main(gid: vec3<u32>) {
     var payload: Payload;
     payload.color = vec3<f32>(0.0);
 
-    traceRay(
-        0u,            // tlasIdx (unused)
-        0u,            // ray flags
-        0xFFu,         // cull mask
-        0u, 0u, 0u,    // sbtRecordOffset, sbtRecordStride, missIndex
+    rtEmitPrimaryRay(
         origin, 0.001,
         direction, 10000.0,
-        &payload,
-    );
-
-    textureStore(outImage,
-                 vec2<i32>(i32(gid.x), i32(gid.y)),
-                 vec4<f32>(payload.color, 1.0));
+        0u,            // ray flags
+        0xFFu,         // cull mask
+        0u, 0u,        // sbtRecordOffset, missIndex
+        payload);
 }