Test Plan for toFloat Module
Aaron Stewart, 24 July 2001

(1) How to test the toFloat module

    Run the shell script in byucc/jhdl/Logic/Modules/FloatingPoint/test/

	test_tofloat

    This is all that is really needed.

If you are not running Unix or Linux, then manually run the commands in this file --- it is just a list commands to run the testbench program with different parameters.
    (Of course, the script must be an actual file in the file system --- this will not work if the script is packed away in a JAR file.)

(2) How to run custom tests of the toFloat module
    The testbench is a good way to familiarize yourself with the toFloat module (although a BETTER way is the example in FloatingPoint/examples/ .)
    (The testbench must be compiled for this to work.  At the time of writing, it isn't clear how this will happen in the release --- so it may be necessary to compile it manually.  It is in byucc/jhdl/Logic/Modules/FloatingPoint/tests/tb_toFloat.java .  Make sure the CLASSPATH is set appropriately too.)

    To run a custom test, type:

	java tb_toFloat help

    This will give a list of options you can send to the toFloat testbench.  For instance,
    
	java tb_toFloat in.width=32 signed out.width=32 out.expwidth=8 pipelined=false

will produce a combinational int-to-float converter which takes a standard 32-bit signed int and converts it to a standard single-precision float.

    Alternately, (if this feature is supported in the release), you can bring up the testbench in Jab *with* parameters, using a similar form.  Here, the double dash ("--") separates the testbench parameters from the browser parameters.

	java jab tb_toFloat -- in.width=32 signed out.width=32 out.expwidth=8 pipelined=false

That's all!

