Merge pull request #33 from AdaWorldAPI/claude/continue-session-0mAVa

feat(hpc): complete Pumpkin shopping list — 5 gap fills + JIT noise codegen

Author: AdaWorldAPI

src/hpc/aabb.rs

@@ -292,6 +292,15 @@ unsafe fn aabb_intersect_batch_sse41(query: &Aabb, candidates: &[Aabb]) -> Vec<b
 /// assert!(!hits[1]);
 /// ```
 pub fn ray_aabb_slab_test_batch(ray: &Ray, aabbs: &[Aabb]) -> (Vec<bool>, Vec<f32>) {
+    #[cfg(target_arch = "x86_64")]
+    {
+        if is_x86_feature_detected!("avx512f") && aabbs.len() >= 16 {
+            // SAFETY: avx512f detected, enough AABBs for batch processing.
+            unsafe {
+                return ray_aabb_slab_test_avx512(ray, aabbs);
+            }
+        }
+    }
     ray_aabb_slab_test_scalar(ray, aabbs)
 }
 
@@ -320,6 +329,128 @@ fn ray_aabb_slab_test_scalar(ray: &Ray, aabbs: &[Aabb]) -> (Vec<bool>, Vec<f32>)
     (hits, t_values)
 }
 
+/// AVX-512 batch ray-AABB slab test: processes 16 AABBs per iteration.
+///
+/// Broadcasts ray origin and inv_dir per axis, gathers candidate min/max
+/// coords into SoA arrays, computes slab intervals with `_mm512_min_ps` /
+/// `_mm512_max_ps`, and combines masks with `_mm512_cmp_ps_mask`.
+///
+/// # Safety
+/// Caller must ensure AVX-512F is available.
+#[cfg(target_arch = "x86_64")]
+#[target_feature(enable = "avx512f")]
+unsafe fn ray_aabb_slab_test_avx512(ray: &Ray, aabbs: &[Aabb]) -> (Vec<bool>, Vec<f32>) {
+    use core::arch::x86_64::*;
+
+    let mut hits = Vec::with_capacity(aabbs.len());
+    let mut t_values = Vec::with_capacity(aabbs.len());
+
+    // Broadcast ray origin and inv_dir per axis
+    let orig_x = _mm512_set1_ps(ray.origin[0]);
+    let orig_y = _mm512_set1_ps(ray.origin[1]);
+    let orig_z = _mm512_set1_ps(ray.origin[2]);
+    let inv_x = _mm512_set1_ps(ray.inv_dir[0]);
+    let inv_y = _mm512_set1_ps(ray.inv_dir[1]);
+    let inv_z = _mm512_set1_ps(ray.inv_dir[2]);
+    let zero = _mm512_set1_ps(0.0);
+
+    // Process 16 AABBs at a time
+    let chunks = aabbs.len() / 16;
+    for c in 0..chunks {
+        let base = c * 16;
+
+        // Gather min/max coords for 16 AABBs into SoA arrays
+        let mut a_min_x = [0.0f32; 16];
+        let mut a_max_x = [0.0f32; 16];
+        let mut a_min_y = [0.0f32; 16];
+        let mut a_max_y = [0.0f32; 16];
+        let mut a_min_z = [0.0f32; 16];
+        let mut a_max_z = [0.0f32; 16];
+
+        for i in 0..16 {
+            let aabb = &aabbs[base + i];
+            a_min_x[i] = aabb.min[0];
+            a_max_x[i] = aabb.max[0];
+            a_min_y[i] = aabb.min[1];
+            a_max_y[i] = aabb.max[1];
+            a_min_z[i] = aabb.min[2];
+            a_max_z[i] = aabb.max[2];
+        }
+
+        // SAFETY: arrays are 16-element, avx512f checked by caller.
+        let v_min_x = _mm512_loadu_ps(a_min_x.as_ptr());
+        let v_max_x = _mm512_loadu_ps(a_max_x.as_ptr());
+        let v_min_y = _mm512_loadu_ps(a_min_y.as_ptr());
+        let v_max_y = _mm512_loadu_ps(a_max_y.as_ptr());
+        let v_min_z = _mm512_loadu_ps(a_min_z.as_ptr());
+        let v_max_z = _mm512_loadu_ps(a_max_z.as_ptr());
+
+        // X axis: t1 = (min - origin) * inv_dir, t2 = (max - origin) * inv_dir
+        let t1_x = _mm512_mul_ps(_mm512_sub_ps(v_min_x, orig_x), inv_x);
+        let t2_x = _mm512_mul_ps(_mm512_sub_ps(v_max_x, orig_x), inv_x);
+        let t_near_x = _mm512_min_ps(t1_x, t2_x);
+        let t_far_x = _mm512_max_ps(t1_x, t2_x);
+
+        // Y axis
+        let t1_y = _mm512_mul_ps(_mm512_sub_ps(v_min_y, orig_y), inv_y);
+        let t2_y = _mm512_mul_ps(_mm512_sub_ps(v_max_y, orig_y), inv_y);
+        let t_near_y = _mm512_min_ps(t1_y, t2_y);
+        let t_far_y = _mm512_max_ps(t1_y, t2_y);
+
+        // Z axis
+        let t1_z = _mm512_mul_ps(_mm512_sub_ps(v_min_z, orig_z), inv_z);
+        let t2_z = _mm512_mul_ps(_mm512_sub_ps(v_max_z, orig_z), inv_z);
+        let t_near_z = _mm512_min_ps(t1_z, t2_z);
+        let t_far_z = _mm512_max_ps(t1_z, t2_z);
+
+        // t_enter = max(t_near_x, t_near_y, t_near_z)
+        let t_enter = _mm512_max_ps(_mm512_max_ps(t_near_x, t_near_y), t_near_z);
+        // t_exit = min(t_far_x, t_far_y, t_far_z)
+        let t_exit = _mm512_min_ps(_mm512_min_ps(t_far_x, t_far_y), t_far_z);
+
+        // hit = t_enter <= t_exit AND t_exit >= 0
+        // _CMP_LE_OQ = 18, _CMP_GE_OQ = 29 (ordered, quiet)
+        let m_le = _mm512_cmp_ps_mask::<{ _CMP_LE_OQ }>(t_enter, t_exit);
+        let m_ge = _mm512_cmp_ps_mask::<{ _CMP_GE_OQ }>(t_exit, zero);
+        let hit_mask = m_le & m_ge;
+
+        // Clamp t_enter to 0 for origins inside box
+        let t_enter_clamped = _mm512_max_ps(t_enter, zero);
+
+        // SAFETY: 16-element array matches __m512 lane count.
+        let mut t_arr = [0.0f32; 16];
+        _mm512_storeu_ps(t_arr.as_mut_ptr(), t_enter_clamped);
+
+        for i in 0..16 {
+            let hit = (hit_mask >> i) & 1 != 0;
+            hits.push(hit);
+            t_values.push(if hit { t_arr[i] } else { f32::MAX });
+        }
+    }
+
+    // Scalar tail for remainder
+    for i in (chunks * 16)..aabbs.len() {
+        let aabb = &aabbs[i];
+        let mut t_enter = f32::NEG_INFINITY;
+        let mut t_exit = f32::INFINITY;
+
+        for axis in 0..3 {
+            let t1 = (aabb.min[axis] - ray.origin[axis]) * ray.inv_dir[axis];
+            let t2 = (aabb.max[axis] - ray.origin[axis]) * ray.inv_dir[axis];
+            let t_near = t1.min(t2);
+            let t_far = t1.max(t2);
+            t_enter = t_enter.max(t_near);
+            t_exit = t_exit.min(t_far);
+        }
+
+        let hit = t_enter <= t_exit && t_exit >= 0.0;
+        hits.push(hit);
+        t_values.push(if hit { t_enter.max(0.0) } else { f32::MAX });
+    }
+
+    (hits, t_values)
+}
+
 /// Expand all AABBs in-place by `(dx, dy, dz)` in both directions per axis.
 pub fn aabb_expand_batch(aabbs: &mut [Aabb], dx: f32, dy: f32, dz: f32) {
     #[cfg(target_arch = "x86_64")]
@@ -722,4 +853,28 @@ mod tests {
         assert!(approx_eq(ray.inv_dir[0], 0.5));
         assert!(ray.inv_dir[1].is_infinite());
     }
+
+    // ---------- AVX-512 ray-AABB parity ----------
+
+    #[test]
+    fn test_ray_aabb_avx512_parity() {
+        // 100 AABBs to exercise AVX-512 + tail
+        let ray = Ray::new([0.0, 0.5, 0.5], [1.0, 0.0, 0.0]);
+        let aabbs: Vec<Aabb> = (0..100)
+            .map(|i| {
+                let f = i as f32;
+                Aabb::new([f, 0.0, 0.0], [f + 1.0, 1.0, 1.0])
+            })
+            .collect();
+        let (hits_batch, ts_batch) = ray_aabb_slab_test_batch(&ray, &aabbs);
+        let (hits_scalar, ts_scalar) = ray_aabb_slab_test_scalar(&ray, &aabbs);
+        assert_eq!(hits_batch, hits_scalar, "ray AVX-512 hit parity");
+        for i in 0..100 {
+            assert!(
+                approx_eq(ts_batch[i], ts_scalar[i]),
+                "ray AVX-512 t parity at {i}: {} vs {}",
+                ts_batch[i], ts_scalar[i]
+            );
+        }
+    }
 }