Re: Specifying time limits in source code ?

hat@se-46.wpa.wtb.tue.nl wrote:


> For the designer, it would be enough to be able to specify an upper
> time limit for some code, and the compiler should accept the program
> if it is able to produce code which meets the dead line.


> - Also, current processors do prefetching and out of order execution
> which is very unpredictable (this may be considered as a form of
> run-time optimization).


Cache misses are a big problem, too.


> Therefore
> Do these beasts exist ?
> Is there work being done on this kind of compilers ?
> Is it just a silly idea of mine ?


I would look for pointers from Richard Gerber's work; he's not the
only person working in this area, but it seems to me that his work is
likely to overlap your interests.


        http://www.cs.umd.edu/~rich/


I think this is an excellent topic to work on, and wish that I had the
time to do so myself. It was my impression, about five years ago,
that the people in the real-time field were getting by with the
software-tools equivalent of rocks and sticks, and that they had been
largely ignored by the program-analysis world, and vice-versa. That
is, then, it looked like there was plenty of low-hanging fruit; I
don't know if that fruit is still there, or if other people have
already snagged the easy stuff. I would beware of getting in too deep
in the area of analyzing exact performance of sequences of machine
instructions, unless that really appeals to you, because


(1) that is a mess, and it could change as fast as you can
        write your dissertation.


(2) you ought to be able to treat some of that as a subroutine;
        advanced optimizing compilers already figure that stuff out
        for scheduling purposes.
--
David Chase -- chase@naturalbridge.com
NaturalBridge LLC -- http://www.naturalbridge.com

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