From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 882 invoked by alias); 26 Apr 2011 22:33:13 -0000 Received: (qmail 873 invoked by uid 22791); 26 Apr 2011 22:33:12 -0000 X-SWARE-Spam-Status: No, hits=-2.3 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW,RFC_ABUSE_POST X-Spam-Check-By: sourceware.org Received: from mail-qy0-f182.google.com (HELO mail-qy0-f182.google.com) (209.85.216.182) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 26 Apr 2011 22:32:57 +0000 Received: by qyk27 with SMTP id 27so609081qyk.20 for ; Tue, 26 Apr 2011 15:32:57 -0700 (PDT) MIME-Version: 1.0 Received: by 10.229.2.135 with SMTP id 7mr1021777qcj.253.1303857176926; Tue, 26 Apr 2011 15:32:56 -0700 (PDT) Received: by 10.229.229.196 with HTTP; Tue, 26 Apr 2011 15:32:56 -0700 (PDT) Date: Tue, 26 Apr 2011 23:00:00 -0000 Message-ID: Subject: PATCH: PR rtl-optimization/48575: RTL vector patterns are limited to 26 elements From: "H.J. Lu" To: GCC Patches Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable 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: 2011-04/txt/msg02078.txt.bz2 On Mon, Apr 4, 2011 at 6:05 PM, H.J. Lu wrote: > On Thu, Mar 31, 2011 at 5:05 AM, Kenneth Zadeck > wrote: >> we hit this limit trying to write the explicit semantics for a >> vec_interleave_evenv32qi. >> >> ;;(define_insn "vec_interleave_evenv32qi" >> ;; =A0[(set (match_operand:V32QI 0 "register_operand" "=3Dr") >> ;; =A0 =A0(vec_select:V32QI >> ;; =A0 =A0 =A0(vec_concat:V64QI >> ;; =A0 =A0 =A0 =A0(match_operand:V32QI 1 "register_operand" "0") >> ;; =A0 =A0 =A0 =A0(match_operand:V32QI 2 "register_operand" "r")) >> ;; =A0 =A0 =A0(parallel [(const_int =A00) (const_int 32) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int =A02) (const_int 34) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int =A04) (const_int 36) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int =A06) (const_int 38) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int =A08) (const_int 40) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 10) (const_int 42) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 12) (const_int 44) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 14) (const_int 46) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 16) (const_int 48) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 18) (const_int 50) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 20) (const_int 52) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 22) (const_int 54) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 24) (const_int 56) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 26) (const_int 58) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 28) (const_int 60) >> ;; =A0 =A0 =A0 =A0 =A0 =A0 (const_int 30) (const_int 62)])))] >> ;; =A0"" >> ;; =A0"rimihv\t%0,%2,8,15,8" >> ;; =A0[(set_attr "type" "rimi")]) >> >> >> kenny >> >> On 03/31/2011 06:16 AM, Mike Stump wrote: >>> >>> On Mar 31, 2011, at 1:41 AM, Richard Guenther wrote: >>>> >>>> On Wed, Mar 30, 2011 at 8:09 PM, H.J. Lu =A0wrot= e: >>>>> >>>>> On Wed, Mar 30, 2011 at 08:02:38AM -0700, H.J. Lu wrote: >>>>>> >>>>>> Hi, >>>>>> >>>>>> Currently, we limit XVECEXP to 26 elements in machine description >>>>>> since we use letters 'a' to 'z' to encode them. =A0I don't see any >>>>>> reason why we can't go beyond 'z'. =A0This patch removes this >>>>>> restriction. >>>>>> Any comments? >>>>>> >>>>> That was wrong. =A0The problem is in vector elements. =A0This patch p= asses >>>>> bootstrap. =A0Any comments? >>>> >>>> Do you really need it? >>> >>> I'm trying to recall if this is the limit Kenny and I hit.... =A0If so, >>> annoying. =A0Kenny could confirm if it was. =A0gcc's general strategy o= f, no >>> fixed N gives gcc a certain flexibility that is very nice to have, on t= hose >>> general grounds, I kinda liked this patch. >> > > Is my patch OK to install? > Here is my patch: http://gcc.gnu.org/ml/gcc-patches/2011-03/msg02105.html OK for trunk? Thanks. --=20 H.J.