From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12088 invoked by alias); 24 May 2007 07:51:49 -0000 Received: (qmail 12079 invoked by uid 22791); 24 May 2007 07:51:48 -0000 X-Spam-Check-By: sourceware.org Received: from hu-out-0506.google.com (HELO hu-out-0506.google.com) (72.14.214.235) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 24 May 2007 07:51:46 +0000 Received: by hu-out-0506.google.com with SMTP id 16so24750hue for ; Thu, 24 May 2007 00:51:43 -0700 (PDT) Received: by 10.67.89.5 with SMTP id r5mr1959302ugl.1179993103716; Thu, 24 May 2007 00:51:43 -0700 (PDT) Received: from scientist.local ( [213.140.22.65]) by mx.google.com with ESMTP id y1sm4324337uge.2007.05.24.00.51.41; Thu, 24 May 2007 00:51:42 -0700 (PDT) Message-ID: <4655440F.9010502@gnu.org> Date: Thu, 24 May 2007 07:51:00 -0000 From: Paolo Bonzini User-Agent: Thunderbird 2.0.0.0 (Macintosh/20070326) MIME-Version: 1.0 Newsgroups: gmane.comp.gcc.patches To: Andrew_Pinski@PlayStation.Sony.Com CC: Kenneth Zadeck , Serge Belyshev , "Berlin, Daniel" , "Edelsohn, David" , echristo@apple.com, gcc-patches , Ian Lance Taylor , richard.earnshaw@arm.com, "Park, Seongbae" , Steven Bosscher , "Weigand, Ulrich" Subject: Re: dataflow branch merging plans. References: <46543F49.8060104@naturalbridge.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2007-05/txt/msg01599.txt.bz2 > The instruction at the time we are verifying: > (insn 147 146 148 7 (set (reg:SI 240) > (mult:SI (zero_extend:SI (subreg:HI (reg:SI 234) 2)) > (zero_extend:SI (subreg:HI (reg:SI 238) 2)))) -1 > (expr_list:REG_EQUAL (mult:SI (zero_extend:SI (subreg:HI (reg:SI 234) 2)) > (const_int 43690 [0xaaaa])) > (nil))) This is an RTL sharing problem between insn 146 and insn 147, caused by the umulsi3_highpart pattern of spu.md (but the same applies to smulsi3_highpart), expanded by loop unrolling. I think you have to use a define_insn_and_split the same way you do with mulsi3 vs. _mulsi3. Paolo From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12630 invoked by alias); 24 May 2007 07:52:05 -0000 Received: (qmail 12616 invoked by uid 22791); 24 May 2007 07:52:03 -0000 X-Spam-Check-By: sourceware.org Received: from main.gmane.org (HELO ciao.gmane.org) (80.91.229.2) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 24 May 2007 07:52:01 +0000 Received: from list by ciao.gmane.org with local (Exim 4.43) id 1Hr86x-0002zW-QJ for gcc-patches@gcc.gnu.org; Thu, 24 May 2007 09:51:51 +0200 Received: from 213-140-22-65.fastres.net ([213.140.22.65]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Thu, 24 May 2007 09:51:51 +0200 Received: from bonzini by 213-140-22-65.fastres.net with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Thu, 24 May 2007 09:51:51 +0200 To: gcc-patches@gcc.gnu.org From: Paolo Bonzini Subject: Re: dataflow branch merging plans. Date: Thu, 24 May 2007 08:12:00 -0000 Message-ID: <4655440F.9010502@gnu.org> References: <46543F49.8060104@naturalbridge.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Kenneth Zadeck , Serge Belyshev , "Berlin, Daniel" , "Edelsohn, David" , echristo@apple.com, gcc-patches , Ian Lance Taylor , richard.earnshaw@arm.com, "Park, Seongbae" , Steven Bosscher , "Weigand, Ulrich" User-Agent: Thunderbird 2.0.0.0 (Macintosh/20070326) In-Reply-To: X-IsSubscribed: yes Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2007-05/txt/msg01600.txt.bz2 Message-ID: <20070524081200.eOIrOIzDRbby4J6j4U-yM5kY9OQ-Mv367x8uT4r4hdE@z> > The instruction at the time we are verifying: > (insn 147 146 148 7 (set (reg:SI 240) > (mult:SI (zero_extend:SI (subreg:HI (reg:SI 234) 2)) > (zero_extend:SI (subreg:HI (reg:SI 238) 2)))) -1 > (expr_list:REG_EQUAL (mult:SI (zero_extend:SI (subreg:HI (reg:SI 234) 2)) > (const_int 43690 [0xaaaa])) > (nil))) This is an RTL sharing problem between insn 146 and insn 147, caused by the umulsi3_highpart pattern of spu.md (but the same applies to smulsi3_highpart), expanded by loop unrolling. I think you have to use a define_insn_and_split the same way you do with mulsi3 vs. _mulsi3. Paolo