From patchwork Mon Nov 30 13:11:10 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Hongtao Liu X-Patchwork-Id: 1408219 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org (client-ip=8.43.85.97; helo=sourceware.org; envelope-from=gcc-patches-bounces@gcc.gnu.org; receiver=) Authentication-Results: ozlabs.org; dmarc=pass (p=none dis=none) header.from=gcc.gnu.org Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; unprotected) header.d=gcc.gnu.org header.i=@gcc.gnu.org header.a=rsa-sha256 header.s=default header.b=QmPooMRA; dkim-atps=neutral Received: from sourceware.org (unknown [8.43.85.97]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 4Cl5DL2Rx1z9sSf for ; Tue, 1 Dec 2020 00:09:36 +1100 (AEDT) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 53FFC383302D; Mon, 30 Nov 2020 13:09:34 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 53FFC383302D DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1606741774; bh=KYcr1y53NPpHQoV/Gx4fbzGzby1vfDwL6SdwR9k//DM=; h=References:In-Reply-To:Date:Subject:To:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To:Cc: From; b=QmPooMRAjpP8zQULI1S7btVDsxIYKAIDQZBAh+zfeO8mdlaw8vHxneDrrzTLtAAQo 3oRDWdBs2HKWXrLlrIz6hNAIYrsT4YzaCDEhvV3kWLIQ85rcbE4GSC4Bk+XHU8kdyg TWFL4GllNg6vy5NUdpMBoTDEhNMyI0/dB5fZGSU4= X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from mail-vs1-xe36.google.com (mail-vs1-xe36.google.com [IPv6:2607:f8b0:4864:20::e36]) by sourceware.org (Postfix) with ESMTPS id 0D9AA386F81D for ; Mon, 30 Nov 2020 13:09:31 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.3.2 sourceware.org 0D9AA386F81D Received: by mail-vs1-xe36.google.com with SMTP id z16so1868362vsp.5 for ; Mon, 30 Nov 2020 05:09:31 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=KYcr1y53NPpHQoV/Gx4fbzGzby1vfDwL6SdwR9k//DM=; b=WgHK+CAnNQMM3ENM+avBD1AmvwpdNPKfkMXYWdkq6CzegJ4OSO7Ruvx4ltFDvuYvJF WnCg8APiX3LJMOb391xQ7OirgHzUtyosyz+1KvDT7v+Wtdb/3xdQaPVnOBBx74IbHZzP y5EYfgbjFZp1TKZgEAemEVk2Lt2MEVs/WFOB1SPtW5WwhBjvksN+OwKB16BCXLNNa2yO QM0BqctlCJmXMAwp0QnZ/Z0UmKrktJ6ZiA6M4kbodBfTLx/zCTH+OFNevzZnMMlvNVY+ ZgVuHsOvuB2LQF8meNgTLBPlhk7Icf+AZr/qjWZ4oYBe+1Hlq1PkFiOKbZv4ahYFlFxc MVXQ== X-Gm-Message-State: AOAM533teQYJ2Z8SKpMmHMLbuxKexnoRw0672TCCq6qpxiv3iUfzlJxk pHwBcdY4j/zAtjVi868kPUXUUVMTMkS5Xq7AReI= X-Google-Smtp-Source: ABdhPJyRqUdKXxcByaWlo4v3NyBlnNHK0cwJBp0A+uCHbMUcNDSZaKFsMgGC9f0TBgOePhBT7xdX4lmEodN8Rqot2rI= X-Received: by 2002:a67:2ac1:: with SMTP id q184mr14476229vsq.57.1606741770506; Mon, 30 Nov 2020 05:09:30 -0800 (PST) MIME-Version: 1.0 References: <20201125103359.GA3788@tucnak> <20201125113211.GC3788@tucnak> In-Reply-To: Date: Mon, 30 Nov 2020 21:11:10 +0800 Message-ID: Subject: [PATCH] i386: Optimize vpsubusw compared to 0 into vpcmpleuw or vpcmpnltuw[PR96906] To: Kirill Yukhin X-Spam-Status: No, score=-10.3 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, FREEMAIL_FROM, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, 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-Patchwork-Original-From: Hongtao Liu via Gcc-patches From: Hongtao Liu Reply-To: Hongtao Liu Cc: Jakub Jelinek , Uros Bizjak , "gcc-patches@gcc.gnu.org" Errors-To: gcc-patches-bounces@gcc.gnu.org Sender: "Gcc-patches" Hi: This patch is quite similar like what jakub did in https://gcc.gnu.org/pipermail/gcc-patches/2020-November/560151.html but for target avx512bw. .i.e. for -mavx512bw -mavx512vl transform code from vpsubusw %xmm1, %xmm0, %xmm0 vpxor %xmm1, %xmm1, %xmm1 vpcmpw $0, %xmm1, %xmm0, %k0 to vpcmpleuw %xmm1, %xmm0, %k0 Bootstrapped/regtested on x86_64-linux is ok. gcc/ChangeLog PR target/96906 * config/i386/sse.md (_ucmp3): Add a new define_split after this insn. gcc/testsuite/ChangeLog * gcc.target/i386/avx512bw-pr96906-1.c: New test. * gcc.target/i386/pr96906-1.c: Add -mno-avx512f. diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md index 4aad462f882..eebc3750584 100644 --- a/gcc/config/i386/sse.md +++ b/gcc/config/i386/sse.md @@ -3006,6 +3006,30 @@ (define_insn "_ucmp3" (set_attr "prefix" "evex") (set_attr "mode" "")]) +;; PR96906 - optimize vpsubusw compared to 0 into vpcmpleuw or vpcmpnltuw. +(define_split + [(set (match_operand: 0 "register_operand") + (unspec: + [(us_minus:VI12_AVX512VL + (match_operand:VI12_AVX512VL 1 "vector_operand") + (match_operand:VI12_AVX512VL 2 "vector_operand")) + (match_operand:VI12_AVX512VL 3 "const0_operand") + (match_operand:SI 4 "const0_operand")] + UNSPEC_PCMP))] + "TARGET_AVX512BW && ix86_binary_operator_ok (US_MINUS, mode, operands)" + [(const_int 0)] + { + /* LE: 2, NLT: 5. */ + rtx cmp_predicate = GEN_INT (2); + if (MEM_P (operands[1])) + { + std::swap (operands[1], operands[2]); + cmp_predicate = GEN_INT (5); + } + emit_insn (gen__ucmp3 (operands[0], operands[1], + operands[2], cmp_predicate)); + DONE; + }) + (define_insn "avx512f_vmcmp3" [(set (match_operand: 0 "register_operand" "=k") (and: diff --git a/gcc/testsuite/gcc.target/i386/avx512bw-pr96906-1.c b/gcc/testsuite/gcc.target/i386/avx512bw-pr96906-1.c new file mode 100644 index 00000000000..ae7ec7abed1 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/avx512bw-pr96906-1.c @@ -0,0 +1,80 @@ +/* PR target/96906 */ +/* { dg-do compile } */ +/* { dg-options "-O2 -mavx512bw -mavx512vl -masm=att" } */ +/* { dg-final { scan-assembler-times {(?n)vpcmpub[ \t]*\$2} 6 } } */ +/* { dg-final { scan-assembler-times {(?n)vpcmpuw[ \t]*\$2} 6 } } */ + + +#include + +__mmask8 +ff1 (__m128i x, __m128i y) +{ + return _mm_cmp_epi16_mask (_mm_subs_epu16 (x, y), _mm_setzero_si128 (), 0); +} + +__mmask8 +ff2 (__m128i x, __m128i y) +{ + return _mm_cmple_epu16_mask (x, y); +} + +__mmask16 +ff3 (__m128i x, __m128i y) +{ + return _mm_cmp_epi8_mask (_mm_subs_epu8 (x, y), _mm_setzero_si128 (), 0); +} + +__mmask16 +ff4 (__m128i x, __m128i y) +{ + return _mm_cmple_epu8_mask (x, y); +} + +__mmask16 +ff5 (__m256i x, __m256i y) +{ + return _mm256_cmp_epi16_mask (_mm256_subs_epu16 (x, y), _mm256_setzero_si256 (), 0); +} + +__mmask16 +ff6 (__m256i x, __m256i y) +{ + return _mm256_cmple_epu16_mask (x, y); +} + +__mmask32 +ff7 (__m256i x, __m256i y) +{ + return _mm256_cmp_epi8_mask (_mm256_subs_epu8 (x, y), _mm256_setzero_si256 (), 0); +} + +__mmask32 +ff8 (__m256i x, __m256i y) +{ + return _mm256_cmple_epu8_mask (x, y); +} + +__mmask32 +ff9 (__m512i x, __m512i y) +{ + return _mm512_cmp_epi16_mask (_mm512_subs_epu16 (x, y), _mm512_setzero_si512 (), 0); +} + +__mmask32 +ff10 (__m512i x, __m512i y) +{ + return _mm512_cmple_epu16_mask (x, y); +} + +__mmask64 +ff11 (__m512i x, __m512i y) +{ + return _mm512_cmp_epi8_mask (_mm512_subs_epu8 (x, y), _mm512_setzero_si512 (), 0); +} + +__mmask64 +ff12 (__m512i x, __m512i y) +{ + return _mm512_cmple_epu8_mask (x, y); +} diff --git a/gcc/testsuite/gcc.target/i386/pr96906-1.c b/gcc/testsuite/gcc.target/i386/pr96906-1.c index 9d836eb2bdd..b1b41bf522d 100644 --- a/gcc/testsuite/gcc.target/i386/pr96906-1.c +++ b/gcc/testsuite/gcc.target/i386/pr96906-1.c @@ -1,6 +1,6 @@ /* PR target/96906 */ /* { dg-do compile } */ -/* { dg-options "-O2 -mavx2" } */ +/* { dg-options "-O2 -mavx2 -mno-avx512f" } */ /* { dg-final { scan-assembler-times "\tvpminub\[^\n\r]*xmm" 2 } } */ /* { dg-final { scan-assembler-times "\tvpminuw\[^\n\r]*xmm" 2 } } */ /* { dg-final { scan-assembler-times "\tvpminub\[^\n\r]*ymm" 2 } } */