NP-complete problem in automatic compiler parallelization?

I have heard of a proof having to do with compilers trying to
automatically parallelize code with general dependencies. The gist was
that there is an NP-complete problem somewhere in there, making
prospects for automatic parallelization look bad.


I recently saw a statement of what was proved somewhere in an online
book or paper on the WWW, but I can't remember where! I don't really
need, and probably couldn't understand the proof, but I would like to
find a correct statement of what was proved.


Please post/email a statement of this if you know it, or pointers to a
book/paper *likely* to have it (I've already spent a bit of time
stabbing in the dark at the library myself).


Thanks,


--


O======================================================================O
| Tim Williams (tjw@nersc.gov 510-423-6278) |
O======================================================================O


--

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