[Bug rtl-optimization/92264] [10 Regression] Compile time hog in 521.wrf_r with -Ofast -march=znver2 -g since r276318

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92264

--- Comment #20 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #17)
> Created attachment 48061 [details]
> cache base term
> 
> I wonder if we could simply cache the base terms in elt_loc_list?  Does that
> make a difference?

Yes:

$ time gfortran module_configure.fppized.f90 -c -march=znver2 -std=legacy
-fconvert=big-endian -fno-openmp -Ofast -march=znver2 -g
real    1m43.067s
user    1m42.410s

$ perf report --stdio | head -n20
...
     6.26%  f951      f951                         [.] get_ref_base_and_extent
     5.85%  f951      f951                         [.] bitmap_set_bit
     5.72%  f951      f951                         [.] find_base_term
     3.73%  f951      f951                         [.]
pre_expr_reaches_here_p_work
     2.77%  f951      f951                         [.] fsm_find_thread_path
     2.62%  f951      f951                         [.] rtx_equal_for_memref_p
     2.32%  f951      f951                         [.] build_object_conflicts
     2.17%  f951      f951                         [.] memrefs_conflict_p
     1.88%  f951      f951                         [.] ira_build_conflicts