This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Open Projects and GCC 3.0


> Value Range Propagation				John Wehle

I've been sitting on some newer VRP code which uses less memory and
converges faster.  I switched off to do some smaller things that were
more likely to be included in the short term and simply haven't gotten
back to VRP.

> 1. Inclusion in GCC 3.0

No one ever mentioned to me that it was a candidate for GCC 3.0 (assuming
I could beat it into shape in time).  I figured that it was sitting
behind GCC 3.0.

> 2. Estimated MTTA (Mean Time To Arrival)

This depends in part on what it means to arrival.  I believe that the
current VRP generates correct code and performs useful work.  It can
be somewhat slow (as you've noticed :-).  I'm happy to keep working
on it, however it would be useful to know what constitutes a useful
implementation that should be included in the GCC source tree (keeping
in mind that it'll continue to evolve)?  And what's the point of
diminishing returns with regards to VRP (aka ... how far should I take
it)?

-- John
-------------------------------------------------------------------------
|   Feith Systems  |   Voice: 1-215-646-8000  |  Email: john@feith.com  |
|    John Wehle    |     Fax: 1-215-540-5495  |                         |
-------------------------------------------------------------------------


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]