perf(noise): batch the three cave-warp channels through the SIMD evaluator -- exact, ~4% faster

The remove-one ablation, re-confirmed on the current build, puts the cave warp at 11.0% static /
11.8% moving of worker time: three independent Perlin3D calls per sample. VoxelNoise::Perlin3D_x3
evaluates the three warp channels together through the existing SIMD evaluator (Codex; reviewed
by me). It is wired into the generator, op-stack, classifier and room/player-fit unwarp paths.

Exact: the field is byte-identical. 0/841 tiles changed, 0 triangle delta, 0 voxels of maximum
surface displacement; canonical export 07C14005 / b3e5f4c3..9b377; capability 20,830 / 10,909;
strate-crossing and sealed-proof tests pass. The seam test (VoxelForgeMazeSeamTest, extended) has
48 neighbouring-tile and 24 cross-LOD comparisons, 0 mismatches. Before/after renders from a
camera aimed at the real arrival tunnel are byte-identical.

Cross-machine determinism: Perlin3D_x3 adds no new intrinsics, no AVX, no FMA and no runtime CPU
dispatch. It reuses the evaluator behind the compile-time gate
`#if PLATFORM_ENABLE_VECTORINTRINSICS && PLATFORM_CPU_X86_FAMILY`, so every x64 machine runs the same
instruction path; its output matches the scalar path bit for bit.

Final DEFAULT build vs the committed-tip DLL (two interleaved runs each):
  static  worker 30.83 -> 29.47 s (-4.4%)   LOD0 generation p50/p95 56.3/99.3 -> 53.3/95.3 ms
  moving  worker 67.28 -> 64.56 s (-4.0%)   LOD0 generation p50/p95 55.9/99.0 -> 53.4/96.4 ms
Ledger: every launch passed; two harness false negatives (the seam test on the game-only
validator, and a trace ended by intentional forced termination) were rerun. No crash folders.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TaFdXhTdYMkW1xrV6ApT6Y

Source/VoxelForge/Public/VoxelNoise.h

@@ -357,6 +357,33 @@ FORCEINLINE void Perlin3D_x4(const float* Xs, const float* Ys, const float* Zs,
 }
 #endif
 
+// Three independent channels are the hot cave-warp shape.  On x86, evaluate them as three lanes
+// of the existing four-wide implementation; the unused fourth lane is deliberate.  The scalar
+// fallback keeps the exact three-call behavior on platforms without the SIMD path.  This helper is
+// the owner of the cave-warp channel evaluation so generator, op-stack, and room-fit callers cannot
+// drift into different noise formulas or channel offsets.
+FORCEINLINE void Perlin3D_x3(
+    float X0, float Y0, float Z0,
+    float X1, float Y1, float Z1,
+    float X2, float Y2, float Z2,
+    float& Out0, float& Out1, float& Out2)
+{
+#if VF_NOISE_USE_SIMD
+    const float Xs[4] = { X0, X1, X2, X0 };
+    const float Ys[4] = { Y0, Y1, Y2, Y0 };
+    const float Zs[4] = { Z0, Z1, Z2, Z0 };
+    float Out[4];
+    Perlin3D_x4(Xs, Ys, Zs, Out);
+    Out0 = Out[0];
+    Out1 = Out[1];
+    Out2 = Out[2];
+#else
+    Out0 = Perlin3D(X0, Y0, Z0);
+    Out1 = Perlin3D(X1, Y1, Z1);
+    Out2 = Perlin3D(X2, Y2, Z2);
+#endif
+}
+
 //=============================================================================
 // fBm / Ridged — octaves evaluated 4 at a time through Perlin3D_x4.
 // Accumulation stays scalar in octave order, so the result is independent of