[Bug tree-optimization/88533] [9 Regression] Higher performance penalty of array-bounds checking for sparse-matrix vector multiply

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88533

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
So we do rely on loop header copying (our single "loop rotation" pass) to hoist
invariant conditions out of loops.

As said determining whether a condition is invariant or not is a bit awkward
at the point of CH.  But reverting the revision in question and
re-instantiating the effect on the added testcase by adding heuristic to
should_duplicate_loop_header_p should hopefully work.