From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id D99BE3858D38; Thu, 28 Dec 2023 04:23:56 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org D99BE3858D38
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1703737436;
	bh=Bj37s5lgfiogASp+t7Zaq0kJnlH+aoF9Grx9W8rUjqI=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=KyA4h1MbfmQHN9S+SPEc84ux2FNPZcPP+q2kwbrAKoBG4odwE8Lo8pT9P3R3m+3CJ
	 Od0bgQUbQEARc6xFHXHsPM+5auywHDE2iaBUzc9xxQx4gMRhTdUquQqngDwyk3Kkyh
	 ElH0uXyKyQGEwavCPYN6EAnwlZX37BezoDEK6bhk=
From: "juzhe.zhong at rivai dot ai" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c/113134] gcc does not version loops with early break
 conditions that don't have side-effects
Date: Thu, 28 Dec 2023 04:23:56 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c
X-Bugzilla-Version: 14.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: juzhe.zhong at rivai dot ai
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-113134-4-hQF3KBdbAD@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-113134-4@http.gcc.gnu.org/bugzilla/>
References: <bug-113134-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D113134

--- Comment #8 from JuzheZhong <juzhe.zhong at rivai dot ai> ---
(In reply to Tamar Christina from comment #7)
> You may be able to use the same approach as
>=20
>   else if (LOOP_VINFO_FULLY_MASKED_P (loop_vinfo))
>=20
> that is, reverse both the mask and the vector and using extract last.
> It's not going to be performance critical so it's more important to be
> correct rather than fast.

I just carefully read this code:

      /* For an inverted control flow with early breaks we want EXTRACT_FIR=
ST
         instead of EXTRACT_LAST.  Emulate by reversing the vector and mask=
. */
      if (restart_loop && LOOP_VINFO_EARLY_BREAKS (loop_vinfo))
        {
          /* First create the permuted mask.  */
          tree perm_mask =3D perm_mask_for_reverse (TREE_TYPE (mask));
          tree perm_dest =3D copy_ssa_name (mask);
          gimple *perm_stmt
                =3D gimple_build_assign (perm_dest, VEC_PERM_EXPR, mask,
                                       mask, perm_mask);
          vect_finish_stmt_generation (loop_vinfo, stmt_info, perm_stmt,
                                       &gsi);
          mask =3D perm_dest;

          /* Then permute the vector contents.  */
          tree perm_elem =3D perm_mask_for_reverse (vectype);
          perm_dest =3D copy_ssa_name (vec_lhs_phi);
          perm_stmt
                =3D gimple_build_assign (perm_dest, VEC_PERM_EXPR, vec_lhs_=
phi,
                                       vec_lhs_phi, perm_elem);
          vect_finish_stmt_generation (loop_vinfo, stmt_info, perm_stmt,
                                       &gsi);
          vec_lhs_phi =3D perm_dest;
        }


Suppose the loop mask is generated by whilelo instruction of ARM SVE.

Suppose we have 8 elements in a single whole vector.

mask =3D whilo (0, res) if res =3D 6, then mask =3D 11111000.
data =3D 12345678

Then if it is early break. You are reversing both data and mask as follows:

new_mask =3D 00011111
new_data =3D 87654321

Then use the EXTRACT_LAST, we will get value =3D 1 for early break.

Am I right ?=