From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [216.205.24.124]) by sourceware.org (Postfix) with ESMTP id 135833851C06 for ; Tue, 17 Nov 2020 00:05:06 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org 135833851C06 Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-293-9JsTQZzRMOmYe2I8vzEuEQ-1; Mon, 16 Nov 2020 19:05:03 -0500 X-MC-Unique: 9JsTQZzRMOmYe2I8vzEuEQ-1 Received: from smtp.corp.redhat.com (int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id 838845F9CA; Tue, 17 Nov 2020 00:05:02 +0000 (UTC) Received: from localhost.localdomain (ovpn-112-176.phx2.redhat.com [10.3.112.176]) by smtp.corp.redhat.com (Postfix) with ESMTP id 570C951512; Tue, 17 Nov 2020 00:05:02 +0000 (UTC) Subject: Re: [PATCH][AVX512]Lower AVX512 vector compare to AVX version when dest is vector To: Hongtao Liu , GCC Patches References: From: Jeff Law Message-ID: <43d01dc3-4d2e-2b94-d83b-aa6e9225b00f@redhat.com> Date: Mon, 16 Nov 2020 17:05:01 -0700 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.3.1 MIME-Version: 1.0 In-Reply-To: X-Scanned-By: MIMEDefang 2.79 on 10.5.11.11 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Content-Language: en-US X-Spam-Status: No, score=-4.0 required=5.0 tests=BAYES_00, BODY_8BITS, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, NICE_REPLY_A, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H4, RCVD_IN_MSPIKE_WL, SPF_HELO_NONE, SPF_PASS, TXREP autolearn=ham autolearn_force=no version=3.4.2 X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on server2.sourceware.org X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 17 Nov 2020 00:05:07 -0000 On 9/2/20 3:34 AM, Hongtao Liu via Gcc-patches wrote: > Hi: > Add define_peephole2 to eliminate potential redundant conversion > from mask to vector. > Bootstrap is ok, regression test is ok for i386/x86-64 backend. > Ok for trunk? > > gcc/ChangeLog: > PR target/96891 > * config/i386/sse.md (VI_128_256): New mode iterator. > (define_peephole2): Lower avx512 vector compare to avx version > when dest is vector. > > gcc/testsuite/ChangeLog: > > * gcc.target/i386/avx512bw-pr96891-1.c: New test. > * gcc.target/i386/avx512f-pr96891-1.c: New test. > * gcc.target/i386/avx512f-pr96891-2.c: New test. Aren't these the two insns in question: (insn 7 4 8 2 (set (reg:QI 86)         (unspec:QI [                 (reg:V8SF 90)                 (reg:V8SF 89)                 (const_int 2 [0x2])             ] UNSPEC_PCMP)) "j.c":4:14 1911 {avx512vl_cmpv8sf3}      (expr_list:REG_DEAD (reg:V8SF 90)         (expr_list:REG_DEAD (reg:V8SF 89)             (nil)))) (note 8 7 9 2 NOTE_INSN_DELETED) (insn 9 8 14 2 (set (reg:V8SI 82 [ _2 ])         (vec_merge:V8SI (const_vector:V8SI [                     (const_int -1 [0xffffffffffffffff]) repeated x8                 ])             (const_vector:V8SI [                     (const_int 0 [0]) repeated x8                 ])             (reg:QI 86))) "j.c":4:14 2705 {*avx512vl_cvtmask2dv8si}      (expr_list:REG_DEAD (reg:QI 86)         (nil))) Note there's a data dependency between them.  insn 7 feeds insn 9.  When there's a data dependency, combiner patterns are usually the better choice than peepholes.  I think you'd be looking to match something likethis (from the . combine dump): (set (reg:V8SI 82 [ _2 ])     (vec_merge:V8SI (const_vector:V8SI [                 (const_int -1 [0xffffffffffffffff]) repeated x8             ])         (const_vector:V8SI [                 (const_int 0 [0]) repeated x8             ])         (unspec:QI [                 (reg:V8SF 90)                 (reg:V8SF 89)                 (const_int 2 [0x2])             ] UNSPEC_PCMP))) Jeff