Re: looking for assertion language for instruction set verification

Alan Lehotsky <lehotsky@tiac.net> wrote:
>Summary: I'm looking for existing research on an assertion language
> suitable for testing an instruction set architecture.


I've been interested in a similar problem for some time, but I haven't
gotten anywhere. If you're curious about the work, you can check out
http://www.cs.virginia.edu/zephyr/csdl/lrtlindex.html. You might also
check out http://www.eecs.harvard.edu/~nr/pubs/checker-abstract.html
for some ideas on automatic generation of test *cases* (but *lacking*
the postconditions).


>I'd like to augment this with another set of statements that would
>make assertions about the pre and post-conditions that are relevant
>to the instruction. These statements would be used to automatically
>generate test programs.


There might be some leverage in some more recent unpublished work I'm
doing, e.g., the algebraic laws mentioned in
http://www.eecs.harvard.edu/~nr/opmsg.html.


I'd be delighted to talk with you on the phone about some of these
problems...




Norman

Return to the comp.compilers page.
Search the comp.compilers archives again.