https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
Andrew Macleod changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #7 from CVS Commits ---
The master branch has been updated by Andrew Macleod :
https://gcc.gnu.org/g:0a38f677463ff8a4fb61b049263aa596ef6471a7
commit r14-275-g0a38f677463ff8a4fb61b049263aa596ef6471a7
Author: Andrew MacLeod
Date:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #6 from Andrew Macleod ---
(In reply to Aldy Hernandez from comment #5)
> (In reply to Richard Biener from comment #3)
> > But yes, your observation about m_has_cache_entry is correct - if that has
> > any value (it makes reset_path
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #5 from Aldy Hernandez ---
(In reply to Richard Biener from comment #3)
> But yes, your observation about m_has_cache_entry is correct - if that has
> any value (it makes reset_path "cheap"), then it should be also relied on
> upon
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #4 from Aldy Hernandez ---
Created attachment 54420
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54420&action=edit
reuse path_range_query SSA cache for all of back_threader class
Off of the top of my head (i.e. probably very
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #3 from Richard Biener ---
But yes, your observation about m_has_cache_entry is correct - if that has any
value (it makes reset_path "cheap"), then it should be also relied on upon
growth. Maybe make this bitmap an optional feature
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #2 from Richard Biener ---
Created attachment 54418
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54418&action=edit
replace vector with hash_map
it's the clearing that shows up on the profile, I've experimented with the
attac
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108697
--- Comment #1 from Aldy Hernandez ---
Sharing a cache with say the ranger is a no go, because the path_range_query's
cache is specific to the path, but perhaps we could share the path's cache
between constructions providing a constructor that t
