From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 7241 invoked by alias); 29 Apr 2009 07:46:45 -0000 Received: (qmail 7172 invoked by uid 48); 29 Apr 2009 07:46:30 -0000 Date: Wed, 29 Apr 2009 07:46:00 -0000 Message-ID: <20090429074630.7171.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug target/39942] Nonoptimal code - leaveq; xchg %ax,%ax; retq In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "vvv at ru dot ru" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2009-04/txt/msg02846.txt.bz2 ------- Comment #11 from vvv at ru dot ru 2009-04-29 07:46 ------- (In reply to comment #8) > From config/i386/i386.c: > /* AMD Athlon works faster > when RET is not destination of conditional jump or directly preceded > by other jump instruction. We avoid the penalty by inserting NOP just > before the RET instructions in such cases. */ > static void > ix86_pad_returns (void) > ... But I am using Core 2 Duo. Why we see multibyte nop, not single byte nop? Why if change line u = F(u)*3+1; to u = F(u)*4+1; or u = F(u); number of nops changed? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39942