[Bug target/84200] [8/9 Regression] r256888 causes 30% performance regression of 519.lbm_r at -Ofast generic tuning on Zen/9

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84200

Martin Liška  changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 Resolution|--- |WONTFIX

--- Comment #10 from Martin Liška  ---
I can't confirm that on a Ryzen 5 machine, it's actually faster on trunk:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=154.477.0=65.477.0=82.477.0=35.477.0;

Out of all configurations, the benchmark is actually faster on trunk than on
GCC-8. Thus I'm closing that.

[Bug target/84200] [8/9 Regression] r256888 causes 30% performance regression of 519.lbm_r at -Ofast generic tuning on Zen/9

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84200

Richard Biener  changed:

   What|Removed |Added

   Target Milestone|9.0 |10.0

--- Comment #9 from Richard Biener  ---
deferred.  Also needs re-confirmation, but I guess also a WONTFIX is an
appropriate action for this case.

[Bug target/84200] [8/9 Regression] r256888 causes 30% performance regression of 519.lbm_r at -Ofast generic tuning on Zen/9

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84200

Richard Biener  changed:

   What|Removed |Added

  Component|middle-end  |target

--- Comment #8 from Richard Biener  ---
Re-classifying as target (alignment / stride-prefetching(?)) issue, sub-target
Zen only actually.