Available on
target_family="wasm" and target feature simd128 only.Expand description
Lane-wise multiply signed 16-bit integers in the two input vectors and add adjacent pairs of the full 32-bit results.