The input of altfp_mult should be a floating-point representation. That is to say, when you want to simulate 3.3* 1. 1, the input should be the floating-point representation of 3.3 and 1. 1. You can use matlab to convert values in advance.

It should be noted that the output of this IP has a delay of several cycles compared with the input, and 6 is still 12, I forgot. You should see it in the software.