This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: GCC's data dependence analyse may inaccuracy

From: Richard Guenther <richard dot guenther at gmail dot com>
To: Revital1 Eres <ERES at il dot ibm dot com>
Cc: Jianzhang Peng <jianzhangpeng at gmail dot com>, gcc at gcc dot gnu dot org
Date: Mon, 14 Dec 2009 14:25:51 +0100
Subject: Re: GCC's data dependence analyse may inaccuracy
References: <8ff4389f0912131820o3120d0b7g25f515bbbddae523@mail.gmail.com> <OF992AD641.9FCFB058-ONC225768C.00479E14-C225768C.004827AE@il.ibm.com>

On Mon, Dec 14, 2009 at 2:08 PM, Revital1 Eres <ERES@il.ibm.com> wrote:
> Hello,
>
>> I unroll the following code one times in a gimpile pass.
>
> Can you please post the flags you used and the full test?
> I can try to reproduce this.

insn 53 (set (mem/s:SF (reg:SI 234 [ ivtmp.51 ])  (reg:SF 245))
//reg245->a[i]

insn 54 (set (reg:SF 246)
// reg246 <-b[i+1]
             (mem/s:SF (plus:SI (reg:SI 236 [ ivtmp.46 ])
                                 (const_int 4 [0x4]))

well, it's easy.  RTL does not have enough alias information to tell
them apart.  This is mainly because TARGET_MEM_REF and
plain INDIRECT_REFs are not handled by the RTL oracle.

Richard.

> Thanks,
> Revital
>
>

References:
- GCC's data dependence analyse may inaccuracy
  - From: Jianzhang Peng
- Re: GCC's data dependence analyse may inaccuracy
  - From: Revital1 Eres

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]