This is the mail archive of the 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]
Other format: [Raw text]

Re: PR 23551: why should we coalesce inlined variables?

On Jul  9, 2007, Andrew MacLeod <> wrote:

> The burden of proof is on the code going in, not what is already
> there.

The distinction that is already there was justified as needed for
major performance improvements.  The distinction no longer provides
that major performance improvement.  It actually hurts in a number of
cases.  Therefore, the distiction is no longer justified.

> I think this is a temporary solution which doesnt
> address the real issue.

This is true.  It was never meant as a solution.  It was meant as
"hey, why are we doing this, it looks wrong?"  "because it made for
better code back then" "well, it doesn't any more"  How do we get from
this to:

> I say back it out.


> I think there is a better way to retain the information.

Yes.  So, we should remove the preservation of user variables
entirely, as in the follow-up patch I posted, right?

Alexandre Oliva
FSF Latin America Board Member
Red Hat Compiler Engineer   aoliva@{,}
Free Software Evangelist  oliva@{,}

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