On Nov 11, 2023, Thomas Schwinge wrote:
> ACK -- but why not likewise "fix" the 'gcc.dg/tree-ssa/scev-4.c' FAIL?
I have evidence from earlier compiler version bumps that there's some
correlation and that scev-4.c also failed in the past, but it wasn't
failing for me this time.
--
Alexandre Oli
Hi!
On 2023-11-08T13:01:47-0300, Alexandre Oliva wrote:
> These gimplefe tests never got the desired optimization on ia32, but
> they only started visibly failing when the representation of MEMs in
> dumps changed from printing 'symbol: a' to '&a'.
ACK -- but why not likewise "fix" the 'gcc.dg/t
On Wed, 8 Nov 2023, Alexandre Oliva wrote:
>
> These gimplefe tests never got the desired optimization on ia32, but
> they only started visibly failing when the representation of MEMs in
> dumps changed from printing 'symbol: a' to '&a'.
>
> The transformation is not considered profitable on ia3
These gimplefe tests never got the desired optimization on ia32, but
they only started visibly failing when the representation of MEMs in
dumps changed from printing 'symbol: a' to '&a'.
The transformation is not considered profitable on ia32, that's why it
doesn't take place. Maybe that's a bu
