Re: Support for full redundancy elimination and VN of virtual operands in GVN-PRE

[Daniel Berlin <dberlin at dberlin dot org>]

On Jun 29, 2004, at 10:30 PM, Diego Novillo wrote:

> On Tue, 2004-06-29 at 22:05, Daniel Berlin wrote:
>
> > The below patch should work fine, assuming it's tested (it's what i 
> > was
> > going to do eventually anyway, after i fixed the current memory use on
> > 20001226-1.c), and as a nice side effect, reduce the number of values
> > we generate.
> OK, thanks.  Should I remove all traces of 'c' checks and replace them
> with is_gimple_min_invariant?  That would also let me remove the
> TREE_CONSTANT and TREE_CHAIN setting we do in add_to_value.

Yup.
You might also want to short circuit phi_translate at the top and just 
return expr if is_gimple_min_invariant is true. There will never be a 
phi for them, and it'll just create needless expressions.
I can take care of this change separately and test it if you don't have 
time right now.



> Diego.