Single and Double Precision Floating Point Modeling

[Hybr1dz]

I'm not looking for synthesizable code. I'm just modeling some floating point instructions for a CPU.

Ultimately, I want to take a 32-bit std_logic_vector(0 to 31), interpret it as a standard single precision floating point number, use +,-,* operators and then store the result as a 32-bit std_logic_vector(0 to 31).

I also want to do the same for a 64-bit std_logic_vector(0 to 63) and interpret it as a double precision floating point number.

I was thinking of using the float_pkg package in ModelSim. I haven't looked into it too deeply yet but I was wondering if that package should be sufficient enough for me.

Any input is appreciated. Thanks in advance!

 

[JohnDuq]

Hmm, this has come up before without getting answered:

http://www.velocityreviews.com/forums/t497993-experience-of-ieeefloatpkg.html

The translations should be fairly straightforward for the libraries; MMUs do this all the time. Your best bet is to set up a test bench and test the floating point extremes to make sure you get the expected results.

John