[GCC 4.2 Project] Omega data dependence test

[Daniel Berlin]

On Sun, 2005-08-14 at 13:14 +0200, Sebastian Pop wrote:
> Daniel Berlin wrote:
> > 
> > Sebastian,  I really think you are worrying too much.
> 
> right.
> 
> > It's pretty rare that it will take going all the way to omega to be able
> > to disambiguate two dependences.
> > 
> 
> for dependence tests we exercise only a limited part of omega, but now
> suppose that we'll use omega for other purposes, for example cache
> miss equations, 
Most cache miss equations are using distance numbers, which we don't
always need to go all the way to omega to get.

> scheduling, 

Dependence again :)

> regalloc, 

Dependence again :)

> etc.  One of these users will
> sooner or later end by triggering the exponential bullet.

Only if you use such a big hammer for every small nail :)