public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "amonakov at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/53265] New: Warn when undefined behavior implies smaller iteration count
Date: Mon, 07 May 2012 14:09:00 -0000	[thread overview]
Message-ID: <bug-53265-4@http.gcc.gnu.org/bugzilla/> (raw)

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53265

             Bug #: 53265
           Summary: Warn when undefined behavior implies smaller iteration
                    count
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: enhancement
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: amonakov@gcc.gnu.org


As requested in PR 53128, this is an enhancement request that asks to produce a
warning when loop bound inferred from undefined behavior may be used to
transform the loop in a surprising way.

Example code:

enum {N=4};
int a[N], pfx[N];
void foo()
{
  int i, accum;
  for (i=0, accum=a[0]; i < N; i++, accum+=a[i])
    pfx[i] = accum;
}

VRP in trunk produces an infinite loop from the above. By the way, with
-fno-tree-vrp one would expect that cunroll unrolls just 3 iterations of the
loop, but in fact it unrolls 4, because nb_iterations is not updated when
nb_iterations_upper_bound is reduced from UB analysis.

Can we simply warn when UB analysis reduces nb_iterations_upper_bound and makes
it statically smaller than symbolic nb_iterations?


             reply	other threads:[~2012-05-07 14:07 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-07 14:09 amonakov at gcc dot gnu.org [this message]
2012-05-07 14:21 ` [Bug tree-optimization/53265] " rguenth at gcc dot gnu.org
2013-01-29 17:23 ` jakub at gcc dot gnu.org
2013-01-31 12:13 ` jakub at gcc dot gnu.org
2013-03-11 10:15 ` rguenth at gcc dot gnu.org
2013-03-11 14:14 ` rguenth at gcc dot gnu.org
2013-03-11 14:16 ` rguenth at gcc dot gnu.org
2013-03-11 14:17 ` rguenth at gcc dot gnu.org
2013-03-11 15:49 ` jakub at gcc dot gnu.org
2013-03-11 15:59 ` rguenth at gcc dot gnu.org
2013-03-11 16:16 ` amonakov at gcc dot gnu.org
2013-03-11 16:42 ` manu at gcc dot gnu.org
2013-03-11 16:50 ` manu at gcc dot gnu.org
2013-03-11 17:11 ` jakub at gcc dot gnu.org
2013-03-12 11:25 ` jakub at gcc dot gnu.org
2013-03-12 11:26 ` jakub at gcc dot gnu.org
2013-03-12 11:30 ` jakub at gcc dot gnu.org
2013-03-12 11:48 ` jakub at gcc dot gnu.org
2013-03-12 12:37 ` jakub at gcc dot gnu.org
2013-03-12 12:41 ` jakub at gcc dot gnu.org
2013-03-12 14:15 ` jakub at gcc dot gnu.org
2013-03-12 20:10 ` jakub at gcc dot gnu.org
2013-03-12 21:48 ` jakub at gcc dot gnu.org
2013-03-13 10:36 ` jakub at gcc dot gnu.org
2013-03-13 11:01 ` rguenth at gcc dot gnu.org
2013-03-14  9:14 ` jakub at gcc dot gnu.org
2013-03-14 10:55 ` jakub at gcc dot gnu.org
2013-04-29 23:18 ` ppluzhnikov at google dot com
2013-04-30  6:07 ` jakub at gcc dot gnu.org
2013-04-30  6:46 ` jakub at gcc dot gnu.org
2014-02-16 13:14 ` jackie.rosen at hushmail dot com
2023-06-09 16:52 ` pinskia at gcc dot gnu.org

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=bug-53265-4@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@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).