Matt Heck wrote:

>>>    - very poorly suited to developing floating-point functions,
>>>      especially untested ones
>>
>> The XCASM assembler supports floating point. So does the associated
>> XCSIM simulator.
>
>
> Sweet-- I'll have to look into it.  Do you happen to know if this
> thing's IEEE compliant?  Converting values between two floating-point
> packages to test things is an extra headache I'd like to avoid.
>
> Thank you, Sergio!
>
> --Matt


Yep. IEEE 754 32 bit single format stored little endien order.

The assembler uses .df .dsf and .pic_df directives to initialise memory
and associate labels with floating point quantities (association allows
the simulator to display, monitor and edit multibyte quantities such as
floats, longs and words according to their type). The .pic_df directive
generates retlw statements suitable for PIC table lookup.

The floating point library supports IEEE 754 single format quite NaNs
(including +/- Infinity) and rounding is by simple truncation.

The assembler allows statements of the form:

fred    .dsf        0
jack    .dsf        0
bert    .dsf        0

       .let    fred = bert / jack

Here the assembler will generate all the code necessary to invoke the
floating point divide function with parameters bert and jack, and then
store the result to fred.

The assembler also allows embedded user functions
e.g.

       .let    fred = SIN(theta) * bert

where SIN is a user defined function

Regards
Sergio Masci

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics