FPGA Floating Point Multiplier Design

If your taste leans more towards mathematical rigour than practical examples, the 2nd edition of Israel Cohen's "Computer Arithmetic Algorithms" is an excellent reference. The majority of the book deals with floating point algorithms, including multiply & divide etc.

Don't get me wrong about the style - the information (in terms of process and design) is all there and extremely useful, but you do have to put some effort into translating the maths into usable verilog/vhdl code.

Simon.

And on the opposite side I was almost tempted to buy a text on FPU design with all the schematics given, I forget the title, pretty much all gate level, pity its not been brought up to date in HDL. IIRC it was a German -> English book to supplement a DLX design tought in schools. I figure the math books would be more usefull so the above ref, I'll look into also. The HW books though could be usefull in covering some of the thorny details.

Also most of the comp arch textbooks usually include some chapters on FPU implemention but not at the detail of either of the two extremes.

You will no doubt be using ready made 18b muls so it will be interesting.

Also there are some opensource designs out there, 1 at opencores.org

regards

johnjakson_usa_com

Hi friends

You can try "Digital Signal Processing with FPGA" Second Edition by Uwe Meyer-Baese (Springer 2004), there is subchapter about floating-point arithmetic implementation. Very nice book for person with VHDL and com arch experience.

I think that the name of this book is "Computer Architecture-Complexity and Correctness" by S.Mueller and W.J.Paul, also from Springer, with chapters about fp arithmetics and data paths.

Both of them discuss fp multipliers. These two books should be enought for your design.

Michal

Indeed it was, thanks. If I see it again I may buy it this time.

regards

johnjakson_usa_com