From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 9485 invoked by alias); 6 Jun 2011 14:28:52 -0000 Received: (qmail 9477 invoked by uid 22791); 6 Jun 2011 14:28:51 -0000 X-SWARE-Spam-Status: No, hits=-2.4 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-ww0-f41.google.com (HELO mail-ww0-f41.google.com) (74.125.82.41) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Mon, 06 Jun 2011 14:28:38 +0000 Received: by wwi18 with SMTP id 18so1490533wwi.2 for ; Mon, 06 Jun 2011 07:28:37 -0700 (PDT) MIME-Version: 1.0 Received: by 10.227.209.146 with SMTP id gg18mr5127008wbb.71.1307370516926; Mon, 06 Jun 2011 07:28:36 -0700 (PDT) Received: by 10.227.37.152 with HTTP; Mon, 6 Jun 2011 07:28:36 -0700 (PDT) In-Reply-To: References: Date: Mon, 06 Jun 2011 14:28:00 -0000 Message-ID: Subject: Re: [patch] Improve detection of widening multiplication in the vectorizer From: Richard Guenther To: Richard Guenther , Ira Rosen , gcc-patches@gcc.gnu.org, Patch Tracking , richard.sandiford@linaro.org 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: 2011-06/txt/msg00412.txt.bz2 On Mon, Jun 6, 2011 at 3:04 PM, Richard Sandiford wrote: > Richard Guenther writes: >> Thanks. =A0I would hope that we eventually can get rid of the >> pattern recognizer ... at least for SSE there is also always >> a scalar variant instruction for each vectorized one. > > AFAIK, that isn't true for ARM and NEON. =A0E.g. I don't know of a single > instruction that does the scalar equivalent of things like VADDHN > (add values and narrow to high half), VSUBL.U32 (subtract two values > and extend the result), etc. > > FWIW, I think MIPS only has minimum and maximum operations for paired > floats, not for single floats or doubles. =A0I don't have the manuals to > hand to check though. > > It's probably OK for the particular case of widening multiplications. > It sounded like you were making a more general statement though. > If so, I think we should try to avoid assuming that every vectorisable > operation has an equivalent scalar machine instruction. Hmm, too bad ;) Yes, I was suggesting that we assume that. I guess for now we can go with the vectorizer pattern matching enhancement and re-visit re-ordering the passes later (I don't have time right now to look into the reported issue). Thanks, Richard. > Richard >