public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
From: Martin Liska <marxin@gcc.gnu.org>
To: gcc-cvs@gcc.gnu.org
Subject: [gcc(refs/users/marxin/heads/loop-unswitch-improvement-v3)] Add more comments.
Date: Mon, 29 Nov 2021 11:13:49 +0000 (GMT)	[thread overview]
Message-ID: <20211129111349.85C61386F437@sourceware.org> (raw)

https://gcc.gnu.org/g:277832038c353f154ba7bc670f6e6d0751102b51

commit 277832038c353f154ba7bc670f6e6d0751102b51
Author: Martin Liska <mliska@suse.cz>
Date:   Mon Nov 29 12:04:28 2021 +0100

    Add more comments.

Diff:
---
 gcc/tree-ssa-loop-unswitch.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)

diff --git a/gcc/tree-ssa-loop-unswitch.c b/gcc/tree-ssa-loop-unswitch.c
index a41c25784f3..aa9f25e1907 100644
--- a/gcc/tree-ssa-loop-unswitch.c
+++ b/gcc/tree-ssa-loop-unswitch.c
@@ -285,11 +285,9 @@ is_maybe_undefined (const tree name, gimple *stmt, class loop *loop)
   return false;
 }
 
-// FIXME
 /* Checks whether we can unswitch LOOP on condition at end of BB -- one of its
    basic blocks (for what it means see comments below).
-   RANGER is gimple ranger used in this pass and unswitch_predicate is returned
-   if there is an opportunity for unswitching.  */
+   All candidates all filled to the provided vector CANDIDATES.  */
 
 static void
 find_unswitching_predicates_for_bb (basic_block bb, class loop *loop,
@@ -409,10 +407,8 @@ evaluate_control_stmt_using_entry_checks (gimple *stmt,
   return NULL_TREE;
 }
 
-/* Find all unswitching predicates for a LOOP that contains BBS.
-   TRUE_EDGE distinguish which PARENT_PREDICATE should be used when
-   asking RANGER infrastructure.  Return unswitch_predicate in the provided
-   CANDIDATES vector.  */
+/* Simplify LOOP based on PREDICATE_PATH where dead edges are properly
+   marked.  */
 
 static bool
 simplify_loop_version (class loop *loop, predicate_vector &predicate_path)
@@ -450,12 +446,11 @@ simplify_loop_version (class loop *loop, predicate_vector &predicate_path)
 }
 
 /* Evaluate how many instructions will be executed if we unswitch
-   LOOP (with BBS) based on PREDICATE.  TRUE_EDGE distinguishes if
-   we calculate taken or not taken edge when asking RANGER.
+   LOOP (with BBS) based on PREDICATE_PATH.
    REACHABLE_FLAG is used for marking of the basic blocks.  */
 
 static unsigned
-evaluate_insns (class loop *loop,  basic_block *bbs,
+evaluate_insns (class loop *loop, basic_block *bbs,
 		predicate_vector &predicate_path,
 		auto_bb_flag &reachable_flag)
 {
@@ -528,7 +523,7 @@ evaluate_insns (class loop *loop,  basic_block *bbs,
 }
 
 /* Evaluate how many instruction will we have if we unswitch LOOP (with BBS)
-   based on CANDIDATE predicate (using RANGER infrastructure).  */
+   based on PREDICATE predicate (using PREDICATE_PATH).  */
 
 static unsigned
 evaluate_loop_insns_for_predicate (class loop *loop, basic_block *bbs,
@@ -552,7 +547,8 @@ evaluate_loop_insns_for_predicate (class loop *loop, basic_block *bbs,
 
 /* Unswitch single LOOP.  NUM is number of unswitchings done; we do not allow
    it to grow too much, it is too easy to create example on that the code would
-   grow exponentially.  RANGER is gimple ranger used in this pass.  */
+   grow exponentially.  PREDICATE_PATH contains so far used predicates
+   for unswitching.  */
 
 static bool
 tree_unswitch_single_loop (class loop *loop, int num,


                 reply	other threads:[~2021-11-29 11:13 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211129111349.85C61386F437@sourceware.org \
    --to=marxin@gcc.gnu.org \
    --cc=gcc-cvs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).