use floatToRawIntBits and doubleToRawLongBits for encoding float values as normalization is not needed and much faster