public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
From: dann@godzilla.ics.uci.edu
To: gcc-gnats@gcc.gnu.org
Subject: optimization/5738: GCSE missed optimization
Date: Wed, 20 Feb 2002 14:16:00 -0000	[thread overview]
Message-ID: <20020220215216.11937.qmail@sources.redhat.com> (raw)


>Number:         5738
>Category:       optimization
>Synopsis:       GCSE missed optimization
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          pessimizes-code
>Submitter-Id:   net
>Arrival-Date:   Wed Feb 20 13:56:01 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     dann@godzilla.ics.uci.edu
>Release:        CVS
>Organization:
>Environment:
sparc-sun-solaris2.8
>Description:
GCSE not working

The following code: 


struct foo 
{
  unsigned short *p;
};


void
func (struct foo *s, unsigned int *coord, _Bool delta)
{
  unsigned short change;

  if (delta) {
    change = *((s)->p++);
    *coord += change;
  } else {
    change = *((s)->p++);
    *coord += change;
    *coord += *((s)->p++) << 8;
  }
}

generates when compiled with gcc from CVS on SPARC (with -O2)

func:
        !#PROLOGUE# 0
        !#PROLOGUE# 1
        andcc   %o2, 0xff, %g0
        mov     %o0, %o5
        be      .LL2
        mov     %o1, %o4
        ld      [%o0], %o0
        ld      [%o1], %o2
        lduh    [%o0], %o3
        sll     %o3, 16, %o1
        srl     %o1, 16, %o1
        add     %o2, %o1, %o2
        add     %o0, 2, %o0
        st      %o0, [%o5]
        b       .LL1
        st      %o2, [%o4]
.LL2:
        ld      [%o0], %o0
        lduh    [%o0], %o3
        ld      [%o1], %o2
        add     %o0, 2, %o0
        lduh    [%o0], %o1
        add     %o2, %o3, %o2
        sll     %o1, 8, %o1
        add     %o2, %o1, %o2
        add     %o0, 2, %o0
        st      %o2, [%o4]
        st      %o0, [%o5]
.LL1:
        retl
        nop

GCSE should be able to realize that
The sequence:
    change = *((s)->p++);
    *coord += change;

occurs on both branches of the conditional and move it before.
>How-To-Repeat:
Compile the code above with -O2 and look at the assembly,
one branch of the conditional should be empty. 

 
>Fix:
n/a
>Release-Note:
>Audit-Trail:
>Unformatted:


             reply	other threads:[~2002-02-20 21:56 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-02-20 14:16 dann [this message]
2002-04-03  2:25 rth
2002-04-03  5:06 Daniel Berlin
2002-04-03  5:36 Daniel Berlin
2002-04-03 10:06 Dan Nicolaescu
2002-04-03 10:55 rth
2002-04-04  8:16 law
2003-03-15  4:46 bangerth

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=20020220215216.11937.qmail@sources.redhat.com \
    --to=dann@godzilla.ics.uci.edu \
    --cc=gcc-gnats@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).