{"id":2232219,"url":"http://patchwork.ozlabs.org/api/patches/2232219/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/patch/20260503235813.389169-1-herumi@nifty.com/","project":{"id":17,"url":"http://patchwork.ozlabs.org/api/projects/17/?format=json","name":"GNU Compiler Collection","link_name":"gcc","list_id":"gcc-patches.gcc.gnu.org","list_email":"gcc-patches@gcc.gnu.org","web_url":null,"scm_url":null,"webscm_url":null,"list_archive_url":"","list_archive_url_format":"","commit_url_format":""},"msgid":"<20260503235813.389169-1-herumi@nifty.com>","list_archive_url":null,"date":"2026-05-03T23:58:13","name":"[v3,3/3] i386: Add peephole2 to convert highpart mul to mulx","commit_ref":null,"pull_url":null,"state":"new","archived":false,"hash":"eebc4ffc02b163d5e5507e5bfa4fc1c842b2ac9d","submitter":{"id":92964,"url":"http://patchwork.ozlabs.org/api/people/92964/?format=json","name":null,"email":"herumi@nifty.com"},"delegate":null,"mbox":"http://patchwork.ozlabs.org/project/gcc/patch/20260503235813.389169-1-herumi@nifty.com/mbox/","series":[{"id":502595,"url":"http://patchwork.ozlabs.org/api/series/502595/?format=json","web_url":"http://patchwork.ozlabs.org/project/gcc/list/?series=502595","date":"2026-05-03T23:58:13","name":null,"version":3,"mbox":"http://patchwork.ozlabs.org/series/502595/mbox/"}],"comments":"http://patchwork.ozlabs.org/api/patches/2232219/comments/","check":"pending","checks":"http://patchwork.ozlabs.org/api/patches/2232219/checks/","tags":{},"related":[],"headers":{"Return-Path":"","X-Original-To":["incoming@patchwork.ozlabs.org","gcc-patches@gcc.gnu.org"],"Delivered-To":["patchwork-incoming@legolas.ozlabs.org","gcc-patches@gcc.gnu.org"],"Authentication-Results":["legolas.ozlabs.org;\n\tdkim=pass (2048-bit key;\n unprotected) header.d=nifty.com header.i=@nifty.com header.a=rsa-sha256\n header.s=default-1th84yt82rvi header.b=m6wLnjNa;\n\tdkim-atps=neutral","legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org\n (client-ip=2620:52:6:3111::32; helo=vm01.sourceware.org;\n envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org;\n receiver=patchwork.ozlabs.org)","sourceware.org;\n\tdkim=pass (2048-bit key,\n unprotected) header.d=nifty.com header.i=@nifty.com header.a=rsa-sha256\n header.s=default-1th84yt82rvi header.b=m6wLnjNa","sourceware.org;\n dmarc=pass (p=none dis=none) header.from=nifty.com","sourceware.org; spf=pass smtp.mailfrom=nifty.com","server2.sourceware.org;\n arc=none smtp.remote-ip=106.153.226.39"],"Received":["from vm01.sourceware.org (vm01.sourceware.org\n [IPv6:2620:52:6:3111::32])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519 server-signature ECDSA (secp384r1) server-digest SHA384)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4g81wk1nXFz1yJV\n\tfor ; Mon, 04 May 2026 09:59:50 +1000 (AEST)","from vm01.sourceware.org (localhost [127.0.0.1])\n\tby sourceware.org (Postfix) with ESMTP id 3DCA64B99F58\n\tfor ; Sun, 3 May 2026 23:59:48 +0000 (GMT)","from mta-snd-e07.mail.nifty.com (mta-snd-e07.mail.nifty.com\n [106.153.226.39])\n by sourceware.org (Postfix) with ESMTPS id 8B6F94BAE7F3\n for ; Sun, 3 May 2026 23:58:21 +0000 (GMT)","from sapp by mta-snd-e07.mail.nifty.com with ESMTP\n id <20260503235819490.KMTE.14880.sapp@nifty.com>;\n Mon, 4 May 2026 08:58:19 +0900"],"DKIM-Filter":["OpenDKIM Filter v2.11.0 sourceware.org 3DCA64B99F58","OpenDKIM Filter v2.11.0 sourceware.org 8B6F94BAE7F3"],"DMARC-Filter":"OpenDMARC Filter v1.4.2 sourceware.org 8B6F94BAE7F3","ARC-Filter":"OpenARC Filter v1.0.0 sourceware.org 8B6F94BAE7F3","ARC-Seal":"i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1777852702; cv=none;\n b=LC74dq5n3vArOncEz+6spxJtGgnbceFxMBExsr0No/oAmedW8KPCHap+I7vBK3QCzMuUdAZMhJUj8qeVhRPE6iHg9G5Gf0rT6zsHW52ei9b8nvJ6mcx08NSxCq3R63TAbWWCqaNxAxpLcwEVLBJQGbq6ziKU4zwegiCV2djup6I=","ARC-Message-Signature":"i=1; a=rsa-sha256; d=sourceware.org; s=key;\n t=1777852702; c=relaxed/simple;\n bh=DHkX2PVEhZ2bmiRE0yQJu/+TVDMTL9edqfwS+dBzWmc=;\n h=From:To:Subject:Date:Message-ID:MIME-Version:DKIM-Signature;\n b=rr21SK/AKsI7rOig45Mubro1NbVm+1DtnLKDgGQ+UMX50rW6MfQpEBOfGqRgWG6uPH9Dgs3qYN3iZ+6YXsfsTJwnjLYGqni2bdHNr1Oj2PcyNWcWWTiSwkfF9p+jwFlhVLVYeD1LVlYVBBkrsG6ylhxk0+fCE2VOpLHAmxiOl8g=","ARC-Authentication-Results":"i=1; server2.sourceware.org","From":"herumi@nifty.com","To":"gcc-patches@gcc.gnu.org","Cc":"MITSUNARI Shigeo ","Subject":"[PATCH v3 3/3] i386: Add peephole2 to convert highpart mul to mulx","Date":"Mon, 4 May 2026 08:58:13 +0900","Message-ID":"<20260503235813.389169-1-herumi@nifty.com>","X-Mailer":"git-send-email 2.43.0","In-Reply-To":"<20260503053451.48504-4-herumi@nifty.com>","References":"<20260503053451.48504-4-herumi@nifty.com>","MIME-Version":"1.0","Content-Transfer-Encoding":"8bit","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed; d=nifty.com;\n s=default-1th84yt82rvi; t=1777852699;\n bh=MCvJVkm6yLGd0qNF9M3sA7HWxW9kmwLhAoeohDPuCCI=;\n h=From:To:Cc:Subject:Date:In-Reply-To:References;\n b=m6wLnjNa9pmCrhqXoNik+WFNGggPWpuAZ/bKmhteqSvbsL2YJuIWuEiumnkfX1XEcryIZF9b\n 9OlpKWzm5jUfAP/Xl1tRHesWMY1wpfX4xMsDPG57VFSUbtkKKXkOrttHOptnKhfKh25EbVeOHs\n MZzsRNtdWx7ODXDzbLWWNAW/gmE/oN1GhLfLM9S4WXvy5KmBUMVwo02bMOqEtR1AlQEx5cOxhm\n u5bwjPVt9DhkFFSZSq5SeaUL6J6/QIGXEAU2Al8F3v9wg8TqF2FdTQIau3MhjFzkVehwflvwDS\n E9igfsBINmtJ807UyRYI7bK/LrU0ZogSIFfPMb1DLGPw323w==","X-BeenThere":"gcc-patches@gcc.gnu.org","X-Mailman-Version":"2.1.30","Precedence":"list","List-Id":"Gcc-patches mailing list ","List-Unsubscribe":",\n ","List-Archive":"","List-Post":"","List-Help":"","List-Subscribe":",\n ","Errors-To":"gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org"},"content":"From: MITSUNARI Shigeo \n\nWhen the register allocator selects the MUL-based highpart pattern\n(umuldi3_highpart) with the source value already in %rdx, it inserts\na redundant mov to %rax before the mul instruction. Add a peephole2\nthat detects this mov + mul sequence and converts it to a single mulx,\neliminating the extra mov.\n\nThis improves inlined loops that perform multiple unsigned divisions\nby constants. For example, a loop with three div-by-constant\noperations now generates 15 instructions (matching LLVM) instead\nof 18.\n\nBefore (loop body excerpt):\n\tmov\trax, rdx\n\tmul\tr9\n\nAfter:\n\tmulx\trdx, rax, r9\n\ngcc/ChangeLog:\n\n\t* config/i386/i386.md: Add peephole2 to convert\n\tmov + umul_highpart to mulx on BMI2 targets.\n---\n gcc/config/i386/i386.md | 17 +++++++++++++++++\n .../gcc.target/i386/bmi2-mulx-highpart-2.c | 19 +++++++++++++++++++\n 2 files changed, 36 insertions(+)\n create mode 100644 gcc/testsuite/gcc.target/i386/bmi2-mulx-highpart-2.c","diff":"diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md\nindex 472f9d41332..1c394690b04 100644\n--- a/gcc/config/i386/i386.md\n+++ b/gcc/config/i386/i386.md\n@@ -11522,6 +11522,23 @@\n (set_attr \"prefix\" \"vex\")\n (set_attr \"mode\" \"\")])\n \n+;; Convert mov + highpart mul to mulx when the mov source is %rdx.\n+;; mov %rdx, %rax; mulq %src -> mulx %src, %rax, %out\n+(define_peephole2\n+ [(set (match_operand:DWIH 0 \"register_operand\")\n+\t(match_operand:DWIH 1 \"register_operand\"))\n+ (parallel [(set (match_operand:DWIH 2 \"register_operand\")\n+\t\t (umul_highpart:DWIH (match_dup 0)\n+\t\t\t(match_operand:DWIH 3 \"nonimmediate_operand\")))\n+\t (clobber (match_dup 0))\n+\t (clobber (reg:CC FLAGS_REG))])]\n+ \"TARGET_BMI2\n+ && REGNO (operands[1]) == DX_REG\n+ && REGNO (operands[0]) != REGNO (operands[2])\"\n+ [(parallel [(set (match_dup 2)\n+\t\t (umul_highpart:DWIH (match_dup 1) (match_dup 3)))\n+\t (clobber (match_dup 0))])])\n+\n ;; Highpart multiplication patterns\n (define_insn \"mul3_highpart\"\n [(set (match_operand:DWIH 0 \"register_operand\" \"=d\")\ndiff --git a/gcc/testsuite/gcc.target/i386/bmi2-mulx-highpart-2.c b/gcc/testsuite/gcc.target/i386/bmi2-mulx-highpart-2.c\nnew file mode 100644\nindex 00000000000..be56cf15d07\n--- /dev/null\n+++ b/gcc/testsuite/gcc.target/i386/bmi2-mulx-highpart-2.c\n@@ -0,0 +1,19 @@\n+/* { dg-do compile { target { ! ia32 } } } */\n+/* { dg-options \"-O2 -mbmi2\" } */\n+/* { dg-final { check-function-bodies \"**\" \"\" \"\" { target *-*-linux* *-*-gnu* } } } */\n+\n+/*\n+**div7loop:\n+**...\n+**\tmulx\t%rsi, %rax, %rdx\n+**...\n+*/\n+\n+unsigned int\n+div7loop (unsigned int x)\n+{\n+ for (int i = 0; i < 10000; i++) {\n+ x ^= (i ^ x) / 7;\n+ }\n+ return x;\n+}\n","prefixes":["v3","3/3"]}