feat(simd): add Sub/Div for F32x8 + F64x4 (AVX2 arithmetic completeness)

F32x8 and F64x4 only had Add + Mul. AVX2 fallback for F32x16 needs
all four arithmetic ops on the 256-bit types.

Additive only — no existing code changed.

https://claude.ai/code/session_01BTATTRUACijvsK4hqmKUBR

Author: claude

src/simd_avx512.rs

@@ -1259,6 +1259,36 @@ impl MulAssign for F32x8 {
     }
 }
 
+impl Sub for F32x8 {
+    type Output = Self;
+    #[inline(always)]
+    fn sub(self, rhs: Self) -> Self {
+        Self(unsafe { _mm256_sub_ps(self.0, rhs.0) })
+    }
+}
+
+impl SubAssign for F32x8 {
+    #[inline(always)]
+    fn sub_assign(&mut self, rhs: Self) {
+        self.0 = unsafe { _mm256_sub_ps(self.0, rhs.0) };
+    }
+}
+
+impl Div for F32x8 {
+    type Output = Self;
+    #[inline(always)]
+    fn div(self, rhs: Self) -> Self {
+        Self(unsafe { _mm256_div_ps(self.0, rhs.0) })
+    }
+}
+
+impl DivAssign for F32x8 {
+    #[inline(always)]
+    fn div_assign(&mut self, rhs: Self) {
+        self.0 = unsafe { _mm256_div_ps(self.0, rhs.0) };
+    }
+}
+
 impl fmt::Debug for F32x8 {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "F32x8({:?})", self.to_array())
@@ -1360,6 +1390,36 @@ impl MulAssign for F64x4 {
     }
 }
 
+impl Sub for F64x4 {
+    type Output = Self;
+    #[inline(always)]
+    fn sub(self, rhs: Self) -> Self {
+        Self(unsafe { _mm256_sub_pd(self.0, rhs.0) })
+    }
+}
+
+impl SubAssign for F64x4 {
+    #[inline(always)]
+    fn sub_assign(&mut self, rhs: Self) {
+        self.0 = unsafe { _mm256_sub_pd(self.0, rhs.0) };
+    }
+}
+
+impl Div for F64x4 {
+    type Output = Self;
+    #[inline(always)]
+    fn div(self, rhs: Self) -> Self {
+        Self(unsafe { _mm256_div_pd(self.0, rhs.0) })
+    }
+}
+
+impl DivAssign for F64x4 {
+    #[inline(always)]
+    fn div_assign(&mut self, rhs: Self) {
+        self.0 = unsafe { _mm256_div_pd(self.0, rhs.0) };
+    }
+}
+
 impl fmt::Debug for F64x4 {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "F64x4({:?})", self.to_array())