IRA copy heuristics

Jeff Law
Fri Sep 5 03:19:00 GMT 2008

Peter Bergner wrote:
> On Thu, 2008-09-04 at 20:28 -0400, David Edelsohn wrote:
>> On Thu, Sep 4, 2008 at 7:39 PM, Vladimir Makarov <> wrote:
>>> Meanwhile I am going to submit your second patch with an added
>>> comment.  The patch permits gcc to generate the same quality code as
>>> before your first patch.
>> Why?
>> As Richard said before:
>> "... it changes
>> the heuristics _without any explanation of why this is necessary_.
>> IMO, that's unacceptable for our shiny, new (and generally very nice)
>> register allocator.  And I think it's unacceptable even if it happens
>> to fix a performance regression."
> I have to agree with Richard and David here.  I find it troubling that
> allocation order affects performance by anything other than a small
> amount due to heuristic noise.  It might be in the end there is a 
> valid reason on why Richard's patch has a positive benefit, but until
> we know why, I'd rather wait.
> Peter
Agreed.  Let's at least have a reasonable explanation of why this is 
affecting performance so much.  In theory, this kind of change ought to 
be in the noise.


More information about the Gcc mailing list