Manual
Programming Reference
r32_float
Ordinary format with one component stored as 32-bit (single-precision) floating-point number.