Re: tree-ssa branch: Call for preliminary testing

[Daniel Berlin <dberlin at dberlin dot org>]

On Monday, May 12, 2003, at 09:01  AM, Steven Bosscher wrote:

> Op ma 12-05-2003, om 14:43 schreef Diego Novillo:
> > On Sun, 2003-05-11 at 13:11, Richard Guenther wrote:
> > > cc1plus: out of memory allocating 564784160 bytes after a total of 
> > > 28152172 bytes
> > > Command exited with non-zero status 1
>
> FWIW, this is not the first report of such a message showing up.  I've
> seen it at least twice for something involving the preprocessor.
> I really can't think of _any_ situation where we would want to allocate
> 564MB in one big chunk, so this really is a bug somewhere.

Apparently you've never had the joy of working with some form of 
interference graph.
:)

Andrew, if this relationship is symmetric (ie a interferes with b 
implies b interferes with a), you can halve the size and represent it 
as a triangular bitmatrix like we do in new-regalloc.
grep for igraph and igraph_index.


> Gr.
> Steven