Fixed-width SIMD (128-bit)

SIMD enables lane-wise parallel operations on 128-bit vectors.

(module
  (func (export "add4") (param $x v128) (param $y v128) (result v128)
    (i32x4.add (local.get $x) (local.get $y)))
)

• Construct vectors with v128.const, i32x4.splat, etc.
• Load/store with v128.load / v128.store.

Instruction Reference

• SIMD Instructions — all v128, i8x16, i16x8, i32x4, i64x2, f32x4, f64x2 operations