The Matrix Multiply supports a matrix of elements of integer type (int16, cint16, int32 or cint32) multiplied by a matrix of elements of integer type. It also supports a matrix of elements of float type (float, cfloat) multiplied by a matrix of elements of float type. However, a mix of integer types and float types is not supported.
The Matrix Multiply for AIE-ML has support of integer types (int16, int32, cint16, and cint32) but does not support floating-point types (float, cfloat).