User manual | Lab5

Laboratory work 5 for TSTE18 Digital Arithmetic
Floating-Point Arithmetic
Oscar Gustafsson
September 30, 2014
The purpose of this laboratory work is to realize different arithmetic operations with the
help of a computer. The recommended programming languages to use are Matlab, VHDL,
and Verilog. However, if you feel more comfortable with using another language feel free to do
so, but make sure you have confirmed with the course responsibles that they can understand
that programming language and have a fair chance of running your instances. It is OK to use
different programming languages for different parts/labs.
It is required that you use a digit-based representation format for all your data, i.e., arrays,
rather than integer types. Feel free to write conversion functions for easy use, but the main
realization should use a digit-based representation in the appropriate radix.
The laboratories are nominally individual. However, as there are a few more students
than can fit in a lab, it is OK to use the smallest number of pairs possible per lab. Note that
each student is only allowed to work in pairs a minimum number of times. Accordingly, even
though you happen to be working in a pair, you must make sure that you understand everything
you submit. Also, note that the pairings will be decided in the lab, based on the number of
students attending. Hence, there is no point in planning the pairs beforehand as some students
may choose to do the labs at home. All source codes will be cross-correlated, so please
write your own and do not get “inspired”.
On each sheet of paper write name, personal id number, and student-id, as well
as the consecutive number assigned to you.
The reporting should consist of (printed and emailed to [email protected]):
• Source code
• Example run, showing the usage of the realization
• Some non-trivial examples showing the correctness of the realization
• Where applicable, make sure that all relevant intermediate results are also shown in the
examples
• Where applicable, drawings (hand-written is OK and no need to email those)
For Matlab, an .m-file with the runs and the output log would be fine. For VHDL/Verilog
a .do-file setting up the windows etc and providing stimuli or a testbench may be appropriate.
It is OK to have separate functions for the different cases.
1
1
Floating-point addition
Realize a floating-point addition with IEEE 754 single precision inputs. Note that you can
base your design on the sign-magnitude adder code from the first lab. It is enough to realize
one rounding mode, but specify how the rounding is done.
2
Floating-point multiplication
Realize a floating-point multiplication with IEEE 754 single precision inputs. Note that you
can base your design on the unsigned binary multiplier code from the third lab. The same thing
holds for rounding as for the adder.
2

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project