From patchwork Thu May 25 09:03:27 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Robin Dapp X-Patchwork-Id: 98906 Return-Path: Delivered-To: ouuuleilei@gmail.com Received: by 2002:a59:994d:0:b0:3d9:f83d:47d9 with SMTP id k13csp218712vqr; Thu, 25 May 2023 02:04:26 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ6Rvc4Li0JGA2uk2pyloXnImOVC//GS78lFZc0jZQa5hm6U0FL3/5Iy9TQ9r4KnYVR76QPi X-Received: by 2002:a17:907:268d:b0:973:797a:d1a8 with SMTP id bn13-20020a170907268d00b00973797ad1a8mr717643ejc.56.1685005466209; Thu, 25 May 2023 02:04:26 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1685005466; cv=none; d=google.com; s=arc-20160816; b=QSmI+1ubMsKO77ippwRi1nG5PYPpKM5wOeN+UNcSpQrxDiJfvJskpel/N8y/7xwCnz uxtKnOjxocF/OJhFQ5wGLy4dggQWsjo4f2xqea0iuTeAw5FmGnyvM1qxH1QMvREzk7RQ ySAqY94dz8fhjBsStafd+RatL8hWBAOislZ3y2WDK4y1CY6aiRDdxX0Q2ke9BxbsYWdM Y7FXi+HpjvTxM2ph1AnP7bj0Pf69bFGx9C3CP96nIiEdJ91CPo5XpC+I8LHBB2u8pN2+ RQbDBOBmLl3pEtjiuQSAX+/FZmfmG0WdUAQvRv5iG1nGlmR9pKj7ZS9kATFlMGRPkHgG d+uA== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:reply-to:from:list-subscribe:list-help:list-post :list-archive:list-unsubscribe:list-id:precedence :content-transfer-encoding:subject:to:content-language:cc:user-agent :mime-version:date:message-id:dmarc-filter:delivered-to :dkim-signature:dkim-filter; bh=OfEcqmsJ8Zmg3A+eqe4ykaQd02RcMVK8bK54pwZJLnE=; b=0ZIW08cvRjkrJpRnqcs6WO9B/uKi3AD9XDXCRNGPuisYryM2ryU4feSRgfyipahQ32 0dBkHLlRsCyz87E9Lhvf/NfRkNeulk3wMDQG90hAbJOffu3tqKLRYL7qDQH+qyVNSM8I t6lDQQTqptqvWds0T9JMFcAwP7BAKCePxeFxSzFI0F2YVtGVbfylNsX9qEaVqt3aBAp/ HXD7eyAinPNBBilyr6k7YuTaId9UVhEjFDFCyn8NrfHCVYlumyttIjb2aHrz54TDOfgR czqm7U//znuJQygfr70v83k7weWtBA92i6RW284Kom0pxWJ2XS6j3tssuvZiRTCSMm/a J+7w== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=ltUy0+hl; spf=pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 2620:52:3:1:0:246e:9693:128c as permitted sender) smtp.mailfrom="gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=gnu.org Received: from sourceware.org (server2.sourceware.org. [2620:52:3:1:0:246e:9693:128c]) by mx.google.com with ESMTPS id b15-20020a170906490f00b00959af644ac4si685416ejq.761.2023.05.25.02.04.25 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 25 May 2023 02:04:26 -0700 (PDT) Received-SPF: pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 2620:52:3:1:0:246e:9693:128c as permitted sender) client-ip=2620:52:3:1:0:246e:9693:128c; Authentication-Results: mx.google.com; dkim=pass header.i=@gcc.gnu.org header.s=default header.b=ltUy0+hl; spf=pass (google.com: domain of gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org designates 2620:52:3:1:0:246e:9693:128c as permitted sender) smtp.mailfrom="gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=gnu.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 8B0153857726 for ; Thu, 25 May 2023 09:04:24 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 8B0153857726 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1685005464; bh=OfEcqmsJ8Zmg3A+eqe4ykaQd02RcMVK8bK54pwZJLnE=; h=Date:Cc:To:Subject:List-Id:List-Unsubscribe:List-Archive: List-Post:List-Help:List-Subscribe:From:Reply-To:From; b=ltUy0+hlcOagpBVKGl5q27CUhC4h8iFfpa6z+NcbHGuixy/LUFYfcUtT6QjRf4pOM df91PRQxjcmTOWBHqunjAsUdOCkGSPDvp135kxVuM/0J9v8gsLFtLymnPDS9jXvGSw EfNYYHVp4H3mB30VHBXamweA71NwFSky9D5g4/bY= X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from mail-ej1-x62c.google.com (mail-ej1-x62c.google.com [IPv6:2a00:1450:4864:20::62c]) by sourceware.org (Postfix) with ESMTPS id 1F30A3857726 for ; Thu, 25 May 2023 09:03:31 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 1F30A3857726 Received: by mail-ej1-x62c.google.com with SMTP id a640c23a62f3a-96f7bf29550so54910766b.3 for ; Thu, 25 May 2023 02:03:31 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1685005409; x=1687597409; h=content-transfer-encoding:subject:from:to:content-language:cc :user-agent:mime-version:date:message-id:x-gm-message-state:from:to :cc:subject:date:message-id:reply-to; bh=OfEcqmsJ8Zmg3A+eqe4ykaQd02RcMVK8bK54pwZJLnE=; b=lOOMkhkKegWJzC/v06wATJaSMpt68xCyKicod1PZnIlU2QMyl2RG4Lbs52gTQ4gH4M SCnA1xEC67qbljSMOcTkLlPstEGGK1daEIrF4GOKOD5rq4VwwozhoH5GK55m6FwLpPcY hN3ScaTZZpAXxty6Ncgc7EBqMBsZZ+E+DjwmpCo2peBPeOOqJVZKPHFCqDbjql3feDsa Rq4uisK0ylpa9z/T9gcZEtNoQZNLusASuoKTOfCooBYwuUNCSnK6okxH9ZfT2y5j5+YT KJBp0xkZ0F+MF+m7NurtaZWh1J3uYT1QHtB4/eLg1mD/ekChf9IWnPn/VZttklUSAFz2 Ykaw== X-Gm-Message-State: AC+VfDx8r+vWu9SddxPN7kCtJxspp7xhBxyYxkHV8HLF9ff+Es2m1b0r ga0tMWdEqfDdOaGouQsl89OIKyXL/Wk= X-Received: by 2002:a17:907:3da4:b0:968:892b:1903 with SMTP id he36-20020a1709073da400b00968892b1903mr917440ejc.5.1685005408775; Thu, 25 May 2023 02:03:28 -0700 (PDT) Received: from [192.168.1.23] (ip-046-005-130-086.um12.pools.vodafone-ip.de. [46.5.130.86]) by smtp.gmail.com with ESMTPSA id p5-20020a05640210c500b0050bca43ff55sm329891edu.68.2023.05.25.02.03.27 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Thu, 25 May 2023 02:03:28 -0700 (PDT) Message-ID: <1a1cc437-12bd-b3e8-5fe4-7edb41625753@gmail.com> Date: Thu, 25 May 2023 11:03:27 +0200 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.10.0 Cc: rdapp.gcc@gmail.com Content-Language: en-US To: gcc-patches , Kito Cheng , palmer , "juzhe.zhong@rivai.ai" , jeffreyalaw Subject: [PATCH] RISC-V: Add autovec sign/zero extension and truncation. X-Spam-Status: No, score=-9.7 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, FREEMAIL_FROM, GIT_PATCH_0, KAM_ASCII_DIVIDERS, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) 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: Robin Dapp via Gcc-patches From: Robin Dapp Reply-To: Robin Dapp Errors-To: gcc-patches-bounces+ouuuleilei=gmail.com@gcc.gnu.org Sender: "Gcc-patches" X-getmail-retrieved-from-mailbox: =?utf-8?q?INBOX?= X-GMAIL-THRID: =?utf-8?q?1766856291833462459?= X-GMAIL-MSGID: =?utf-8?q?1766856291833462459?= Hi, this patch implements the autovec expanders for sign and zero extension patterns as well as the accompanying truncations. In order to use them additional mode_attr iterators as well as vectorizer hooks are required. Using these hooks we can e.g. vectorize with VNx4QImode as base mode and extend VNx4SI to VNx4DI. They are still going to be expanded in the future. vf4 and vf8 truncations are emulated by truncating two and three times respectively. The patch also adds tests and changes some expectations for already existing ones. Combine does not yet handle binary operations of two widened operands as we are missing the necessary split/rewrite patterns. These will be added at a later time. Co-authored-by: Juzhe Zhong riscv.exp testsuite is unchanged. zero-scratch-regs-3.c seems to FAIL in vcondu but that already happens on trunk. Regards Robin gcc/ChangeLog: * config/riscv/autovec.md (2): New expander. (2): Dito. (2): Dito. (trunc2): Dito. (trunc2): Dito. (trunc2): Dito. * config/riscv/riscv-protos.h (riscv_v_ext_mode_p): Declare. (vectorize_related_mode): Define. (autovectorize_vector_modes): Define. * config/riscv/riscv-v.cc (vectorize_related_mode): Implement hook. (autovectorize_vector_modes): Implement hook. * config/riscv/riscv.cc (riscv_v_ext_tuple_mode_p): Export. (riscv_autovectorize_vector_modes): Implement target hook. (riscv_vectorize_related_mode): Implement target hook. (TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES): Define. (TARGET_VECTORIZE_RELATED_MODE): Define. * config/riscv/vector-iterators.md: Add lowercase versions of mode_attr iterators. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/binop/shift-rv32gcv.c: Adjust expectation. * gcc.target/riscv/rvv/autovec/binop/shift-rv64gcv.c: Dito. * gcc.target/riscv/rvv/autovec/binop/vdiv-run.c: Dito. * gcc.target/riscv/rvv/autovec/binop/vdiv-rv32gcv.c: Dito. * gcc.target/riscv/rvv/autovec/binop/vdiv-rv64gcv.c: Dito. * gcc.target/riscv/rvv/autovec/binop/vdiv-template.h: Dito. * gcc.target/riscv/rvv/autovec/binop/vrem-rv32gcv.c: Dito. * gcc.target/riscv/rvv/autovec/binop/vrem-rv64gcv.c: Dito. * gcc.target/riscv/rvv/autovec/zve32f_zvl128b-2.c: Dito. * gcc.target/riscv/rvv/autovec/zve32x_zvl128b-2.c: Dito. * gcc.target/riscv/rvv/autovec/zve64d-2.c: Dito. * gcc.target/riscv/rvv/autovec/zve64f-2.c: Dito. * gcc.target/riscv/rvv/autovec/zve64x-2.c: Dito. * gcc.target/riscv/rvv/rvv.exp: Add new conversion tests. * gcc.target/riscv/rvv/vsetvl/avl_single-38.c: Do not vectorize. * gcc.target/riscv/rvv/vsetvl/avl_single-47.c: Dito. * gcc.target/riscv/rvv/vsetvl/avl_single-48.c: Dito. * gcc.target/riscv/rvv/vsetvl/avl_single-49.c: Dito. * gcc.target/riscv/rvv/vsetvl/imm_switch-8.c: Dito. * gcc.target/riscv/rvv/autovec/conversions/vncvt-run.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vncvt-rv32gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vncvt-rv64gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vncvt-template.h: New test. * gcc.target/riscv/rvv/autovec/conversions/vsext-run.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vsext-rv32gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vsext-rv64gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vsext-template.h: New test. * gcc.target/riscv/rvv/autovec/conversions/vzext-run.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vzext-rv32gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vzext-rv64gcv.c: New test. * gcc.target/riscv/rvv/autovec/conversions/vzext-template.h: New test. --- gcc/config/riscv/autovec.md | 104 ++++++++++++++++++ gcc/config/riscv/riscv-protos.h | 5 + gcc/config/riscv/riscv-v.cc | 83 ++++++++++++++ gcc/config/riscv/riscv.cc | 31 +++++- gcc/config/riscv/vector-iterators.md | 33 +++++- .../riscv/rvv/autovec/binop/shift-rv32gcv.c | 1 - .../riscv/rvv/autovec/binop/shift-rv64gcv.c | 5 +- .../riscv/rvv/autovec/binop/vdiv-run.c | 4 +- .../riscv/rvv/autovec/binop/vdiv-rv32gcv.c | 7 +- .../riscv/rvv/autovec/binop/vdiv-rv64gcv.c | 7 +- .../riscv/rvv/autovec/binop/vdiv-template.h | 5 +- .../riscv/rvv/autovec/binop/vrem-rv32gcv.c | 7 +- .../riscv/rvv/autovec/binop/vrem-rv64gcv.c | 7 +- .../riscv/rvv/autovec/conversions/vncvt-run.c | 35 ++++++ .../rvv/autovec/conversions/vncvt-rv32gcv.c | 6 + .../rvv/autovec/conversions/vncvt-rv64gcv.c | 6 + .../rvv/autovec/conversions/vncvt-template.h | 19 ++++ .../riscv/rvv/autovec/conversions/vsext-run.c | 35 ++++++ .../rvv/autovec/conversions/vsext-rv32gcv.c | 8 ++ .../rvv/autovec/conversions/vsext-rv64gcv.c | 8 ++ .../rvv/autovec/conversions/vsext-template.h | 19 ++++ .../riscv/rvv/autovec/conversions/vzext-run.c | 35 ++++++ .../rvv/autovec/conversions/vzext-rv32gcv.c | 8 ++ .../rvv/autovec/conversions/vzext-rv64gcv.c | 8 ++ .../rvv/autovec/conversions/vzext-template.h | 19 ++++ .../riscv/rvv/autovec/zve32f_zvl128b-2.c | 2 +- .../riscv/rvv/autovec/zve32x_zvl128b-2.c | 2 +- .../gcc.target/riscv/rvv/autovec/zve64d-2.c | 2 +- .../gcc.target/riscv/rvv/autovec/zve64f-2.c | 2 +- .../gcc.target/riscv/rvv/autovec/zve64x-2.c | 2 +- gcc/testsuite/gcc.target/riscv/rvv/rvv.exp | 2 + .../riscv/rvv/vsetvl/avl_single-38.c | 2 +- .../riscv/rvv/vsetvl/avl_single-47.c | 2 +- .../riscv/rvv/vsetvl/avl_single-48.c | 2 +- .../riscv/rvv/vsetvl/avl_single-49.c | 2 +- .../riscv/rvv/vsetvl/imm_switch-8.c | 2 +- 36 files changed, 491 insertions(+), 36 deletions(-) create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-run.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv32gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv64gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-template.h create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-run.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv32gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv64gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-template.h create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-run.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv32gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv64gcv.c create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-template.h diff --git a/gcc/config/riscv/autovec.md b/gcc/config/riscv/autovec.md index 1c03468c2f4..c2f2d5083db 100644 --- a/gcc/config/riscv/autovec.md +++ b/gcc/config/riscv/autovec.md @@ -373,3 +373,107 @@ (define_expand "vcondu" DONE; } ) + +;; ------------------------------------------------------------------------- +;; ---- [INT] Sign and zero extension +;; ------------------------------------------------------------------------- +;; Includes: +;; - vzext.vf[2|4|8] +;; - vsext.vf[2|4|8] +;; ------------------------------------------------------------------------- + +(define_expand "2" + [(set (match_operand:VWEXTI 0 "register_operand") + (any_extend:VWEXTI + (match_operand: 1 "register_operand")))] + "TARGET_VECTOR" +{ + insn_code icode = code_for_pred_vf2 (, mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, operands); + DONE; +}) + +(define_expand "2" + [(set (match_operand:VQEXTI 0 "register_operand") + (any_extend:VQEXTI + (match_operand: 1 "register_operand")))] + "TARGET_VECTOR" +{ + insn_code icode = code_for_pred_vf4 (, mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, operands); + DONE; +}) + +(define_expand "2" + [(set (match_operand:VOEXTI 0 "register_operand") + (any_extend:VOEXTI + (match_operand: 1 "register_operand")))] + "TARGET_VECTOR" +{ + insn_code icode = code_for_pred_vf8 (, mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, operands); + DONE; +}) + +;; ------------------------------------------------------------------------- +;; ---- [INT] Truncation +;; ------------------------------------------------------------------------- +;; - vncvt.x.x.w +;; ------------------------------------------------------------------------- +(define_expand "trunc2" + [(set (match_operand: 0 "register_operand") + (truncate: + (match_operand:VWEXTI 1 "register_operand")))] + "TARGET_VECTOR" +{ + insn_code icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, operands); + DONE; +}) + +;; ------------------------------------------------------------------------- +;; Truncation to a mode whose inner mode size is a quarter of mode's. +;; We emulate this with two consecutive vncvts. +;; ------------------------------------------------------------------------- +(define_expand "trunc2" + [(set (match_operand: 0 "register_operand") + (truncate: + (match_operand:VQEXTI 1 "register_operand")))] + "TARGET_VECTOR" +{ + rtx half = gen_reg_rtx (mode); + rtx opshalf[] = {half, operands[1]}; + insn_code icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, opshalf); + + rtx ops[] = {operands[0], half}; + icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, ops); + DONE; +}) + +;; ------------------------------------------------------------------------- +;; Truncation to a mode whose inner mode size is an eigth of mode's. +;; We emulate this with three consecutive vncvts. +;; ------------------------------------------------------------------------- +(define_expand "trunc2" + [(set (match_operand: 0 "register_operand") + (truncate: + (match_operand:VOEXTI 1 "register_operand")))] + "TARGET_VECTOR" +{ + rtx half = gen_reg_rtx (mode); + rtx opshalf[] = {half, operands[1]}; + insn_code icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, opshalf); + + rtx quarter = gen_reg_rtx (mode); + rtx opsquarter[] = {quarter, half}; + icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, opsquarter); + + rtx ops[] = {operands[0], quarter}; + icode = code_for_pred_trunc (mode); + riscv_vector::emit_vlmax_insn (icode, riscv_vector::RVV_UNOP, ops); + DONE; +}) diff --git a/gcc/config/riscv/riscv-protos.h b/gcc/config/riscv/riscv-protos.h index 5f78fd579bb..6219cff1354 100644 --- a/gcc/config/riscv/riscv-protos.h +++ b/gcc/config/riscv/riscv-protos.h @@ -81,6 +81,7 @@ extern void riscv_reinit (void); extern poly_uint64 riscv_regmode_natural_size (machine_mode); extern bool riscv_v_ext_vector_mode_p (machine_mode); extern bool riscv_v_ext_tuple_mode_p (machine_mode); +extern bool riscv_v_ext_mode_p (machine_mode mode); extern bool riscv_shamt_matches_mask_p (int, HOST_WIDE_INT); extern void riscv_subword_address (rtx, rtx *, rtx *, rtx *, rtx *); extern void riscv_lshift_subword (machine_mode, rtx, rtx, rtx *); @@ -253,6 +254,10 @@ enum frm_field_enum FRM_RMM = 0b100, DYN = 0b111 }; + +opt_machine_mode vectorize_related_mode (machine_mode, scalar_mode, + poly_uint64); +unsigned int autovectorize_vector_modes (vec *, bool); } /* We classify builtin types into two classes: diff --git a/gcc/config/riscv/riscv-v.cc b/gcc/config/riscv/riscv-v.cc index f71ad9e46a1..18b39093495 100644 --- a/gcc/config/riscv/riscv-v.cc +++ b/gcc/config/riscv/riscv-v.cc @@ -1396,9 +1396,92 @@ get_cmp_insn_code (rtx_code code, machine_mode mode) default: gcc_unreachable (); } + return icode; } +/* This hook gives the vectorizer more vector mode options. We want it to not + only try modes with the maximum number of units a full vector can hold but + for example also half the number of units for a smaller elements size. + Such vectors can be promoted to a full vector of widened elements + (still with the same number of elements, essentially vectorizing at a + fixed number of units rather than a fixed number of bytes). */ +unsigned int +autovectorize_vector_modes (vector_modes *modes, bool) +{ + if (TARGET_VECTOR) + { + /* TODO: We will support RVV VLS auto-vectorization mode in the future. */ + poly_uint64 full_size + = BYTES_PER_RISCV_VECTOR * ((int) riscv_autovec_lmul); + + /* Start with a VNxYYQImode where YY is the number of units that + fit a whole vector. + Then try YY = nunits / 2, nunits / 4 and nunits / 8 which + is guided by the extensions we have available (vf2, vf4 and vf8). + + - full_size: Try using full vectors for all element types. + - full_size / 2: + Try using 16-bit containers for 8-bit elements and full vectors + for wider elements. + - full_size / 4: + Try using 32-bit containers for 8-bit and 16-bit elements and + full vectors for wider elements. + - full_size / 8: + Try using 64-bit containers for all element types. */ + static const int rvv_factors[] = {1, 2, 4, 8}; + for (unsigned int i = 0; i < sizeof (rvv_factors) / sizeof (int); i++) + { + poly_uint64 units; + machine_mode mode; + if (can_div_trunc_p (full_size, rvv_factors[i], &units) + && get_vector_mode (QImode, units).exists (&mode)) + modes->safe_push (mode); + } + } + return 0; +} + +/* If the given VECTOR_MODE is an RVV mode, first get the largest number + of units that fit into a full vector at the given ELEMENT_MODE. + We will have the vectorizer call us with a successively decreasing + number of units (as specified in autovectorize_vector_modes). + The starting mode is always the one specified by preferred_simd_mode. */ +opt_machine_mode +vectorize_related_mode (machine_mode vector_mode, scalar_mode element_mode, + poly_uint64 nunits) +{ + /* TODO: We will support RVV VLS auto-vectorization mode in the future. */ + poly_uint64 min_units; + if (riscv_v_ext_mode_p (vector_mode) + && multiple_p (BYTES_PER_RISCV_VECTOR * ((int) riscv_autovec_lmul), + GET_MODE_SIZE (element_mode), &min_units)) + { + machine_mode rvv_mode; + if (maybe_ne (nunits, 0U)) + { + /* If we were given a number of units NUNITS, try to find an + RVV vector mode of inner mode ELEMENT_MODE with the same + number of units. */ + if (multiple_p (min_units, nunits) + && get_vector_mode (element_mode, nunits).exists (&rvv_mode)) + return rvv_mode; + } + else + { + /* Look for a vector mode with the same number of units as the + VECTOR_MODE we were given. We keep track of the minimum + number of units so far which determines the smallest necessary + but largest possible, suitable mode for vectorization. */ + min_units = ordered_min (min_units, GET_MODE_SIZE (vector_mode)); + if (get_vector_mode (element_mode, min_units).exists (&rvv_mode)) + return rvv_mode; + } + } + + return default_vectorize_related_mode (vector_mode, element_mode, nunits); +} + /* Expand an RVV comparison. */ void diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc index 23040557881..3dfccd3a8c3 100644 --- a/gcc/config/riscv/riscv.cc +++ b/gcc/config/riscv/riscv.cc @@ -1031,7 +1031,7 @@ riscv_v_ext_tuple_mode_p (machine_mode mode) /* Return true if it is either RVV vector mode or RVV tuple mode. */ -static bool +bool riscv_v_ext_mode_p (machine_mode mode) { return riscv_v_ext_vector_mode_p (mode) || riscv_v_ext_tuple_mode_p (mode); @@ -7616,6 +7616,28 @@ riscv_mode_priority (int, int n) return n; } +/* Implement TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES. */ +unsigned int +riscv_autovectorize_vector_modes (vector_modes *modes, bool all) +{ + if (TARGET_VECTOR) + return riscv_vector::autovectorize_vector_modes (modes, all); + + return default_autovectorize_vector_modes (modes, all); +} + +/* Implement TARGET_VECTORIZE_RELATED_MODE. */ +opt_machine_mode +riscv_vectorize_related_mode (machine_mode vector_mode, scalar_mode element_mode, + poly_uint64 nunits) +{ + if (TARGET_VECTOR) + return riscv_vector::vectorize_related_mode (vector_mode, element_mode, + nunits); + return default_vectorize_related_mode (vector_mode, element_mode, nunits); +} + + /* Initialize the GCC target structure. */ #undef TARGET_ASM_ALIGNED_HI_OP #define TARGET_ASM_ALIGNED_HI_OP "\t.half\t" @@ -7907,6 +7929,13 @@ riscv_mode_priority (int, int n) #undef TARGET_MODE_PRIORITY #define TARGET_MODE_PRIORITY riscv_mode_priority +#undef TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES +#define TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES \ + riscv_autovectorize_vector_modes + +#undef TARGET_VECTORIZE_RELATED_MODE +#define TARGET_VECTORIZE_RELATED_MODE riscv_vectorize_related_mode + struct gcc_target targetm = TARGET_INITIALIZER; #include "gt-riscv.h" diff --git a/gcc/config/riscv/vector-iterators.md b/gcc/config/riscv/vector-iterators.md index 3096ac5be3c..70fb5b80b1b 100644 --- a/gcc/config/riscv/vector-iterators.md +++ b/gcc/config/riscv/vector-iterators.md @@ -1118,8 +1118,10 @@ (define_mode_attr V_DOUBLE_TRUNC [ (VNx16HI "VNx16QI") (VNx32HI "VNx32QI") (VNx64HI "VNx64QI") (VNx1SI "VNx1HI") (VNx2SI "VNx2HI") (VNx4SI "VNx4HI") (VNx8SI "VNx8HI") (VNx16SI "VNx16HI") (VNx32SI "VNx32HI") - (VNx1DI "VNx1SI") (VNx2DI "VNx2SI") (VNx4DI "VNx4SI") (VNx8DI "VNx8SI") (VNx16DI "VNx16SI") - (VNx1DF "VNx1SF") (VNx2DF "VNx2SF") (VNx4DF "VNx4SF") (VNx8DF "VNx8SF") (VNx16DF "VNx16SF") + (VNx1DI "VNx1SI") (VNx2DI "VNx2SI") (VNx4DI "VNx4SI") (VNx8DI "VNx8SI") + (VNx16DI "VNx16SI") + (VNx1DF "VNx1SF") (VNx2DF "VNx2SF") (VNx4DF "VNx4SF") (VNx8DF "VNx8SF") + (VNx16DF "VNx16SF") ]) (define_mode_attr V_QUAD_TRUNC [ @@ -1130,7 +1132,32 @@ (define_mode_attr V_QUAD_TRUNC [ ]) (define_mode_attr V_OCT_TRUNC [ - (VNx1DI "VNx1QI") (VNx2DI "VNx2QI") (VNx4DI "VNx4QI") (VNx8DI "VNx8QI") (VNx16DI "VNx16QI") + (VNx1DI "VNx1QI") (VNx2DI "VNx2QI") (VNx4DI "VNx4QI") (VNx8DI "VNx8QI") + (VNx16DI "VNx16QI") +]) + +; Again in lower case. +(define_mode_attr v_double_trunc [ + (VNx1HI "vnx1qi") (VNx2HI "vnx2qi") (VNx4HI "vnx4qi") (VNx8HI "vnx8qi") + (VNx16HI "vnx16qi") (VNx32HI "vnx32qi") (VNx64HI "vnx64qi") + (VNx1SI "vnx1hi") (VNx2SI "vnx2hi") (VNx4SI "vnx4hi") (VNx8SI "vnx8hi") + (VNx16SI "vnx16hi") (VNx32SI "vnx32hi") + (VNx1DI "vnx1si") (VNx2DI "vnx2si") (VNx4DI "vnx4si") (VNx8DI "vnx8si") + (VNx16DI "vnx16si") + (VNx1DF "vnx1sf") (VNx2DF "vnx2sf") (VNx4DF "vnx4sf") (VNx8DF "vnx8sf") + (VNx16DF "vnx16sf") +]) + +(define_mode_attr v_quad_trunc [ + (VNx1SI "vnx1qi") (VNx2SI "vnx2qi") (VNx4SI "vnx4qi") (VNx8SI "vnx8qi") + (VNx16SI "vnx16qi") (VNx32SI "vnx32qi") + (VNx1DI "vnx1hi") (VNx2DI "vnx2hi") (VNx4DI "vnx4hi") (VNx8DI "vnx8hi") + (VNx16DI "vnx16hi") +]) + +(define_mode_attr v_oct_trunc [ + (VNx1DI "vnx1qi") (VNx2DI "vnx2qi") (VNx4DI "vnx4qi") (VNx8DI "vnx8qi") + (VNx16DI "vnx16qi") ]) (define_mode_attr VINDEX_DOUBLE_TRUNC [ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv32gcv.c index d98100b3276..557a7c82531 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv32gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv32gcv.c @@ -10,4 +10,3 @@ /* { dg-final { scan-assembler {\tvsll\.vv} } } */ /* { dg-final { scan-assembler {\tvsrl\.vv} } } */ /* { dg-final { scan-assembler {\tvsra\.vv} } } */ - diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv64gcv.c index d9109fd8774..01a9cb21efc 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv64gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/shift-rv64gcv.c @@ -3,9 +3,6 @@ #include "shift-template.h" -/* TODO: For int16_t and uint16_t we need widening/promotion patterns. - Therefore, expect only 4 vsll.vv instead of 6 for now. */ - -/* { dg-final { scan-assembler-times {\tvsll\.vv} 4 } } */ +/* { dg-final { scan-assembler-times {\tvsll\.vv} 6 } } */ /* { dg-final { scan-assembler-times {\tvsrl\.vv} 3 } } */ /* { dg-final { scan-assembler-times {\tvsra\.vv} 3 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-run.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-run.c index aa9a3c55abe..5de339172fc 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-run.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-run.c @@ -15,7 +15,7 @@ a##TYPE[i] = VAL * 3; \ b##TYPE[i] = VAL; \ } \ - vadd_##TYPE (a##TYPE, a##TYPE, b##TYPE, SZ); \ + vdiv_##TYPE (a##TYPE, a##TYPE, b##TYPE, SZ); \ for (int i = 0; i < SZ; i++) \ assert (a##TYPE[i] == 3); @@ -23,7 +23,7 @@ TYPE as##TYPE[SZ]; \ for (int i = 0; i < SZ; i++) \ as##TYPE[i] = VAL * 5; \ - vadds_##TYPE (as##TYPE, as##TYPE, VAL, SZ); \ + vdivs_##TYPE (as##TYPE, as##TYPE, VAL, SZ); \ for (int i = 0; i < SZ; i++) \ assert (as##TYPE[i] == 5); diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv32gcv.c index 9759401b9ef..1dce9dd562e 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv32gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv32gcv.c @@ -3,7 +3,8 @@ #include "vdiv-template.h" -/* TODO: Implement vector type promotion. We should have 6 vdiv.vv here. */ +/* Currently we use an epilogue loop which also contains vdivs. Therefore we + expect 10 vdiv[u]s instead of 6. */ -/* { dg-final { scan-assembler-times {\tvdiv\.vv} 4 } } */ -/* { dg-final { scan-assembler-times {\tvdivu\.vv} 6 } } */ +/* { dg-final { scan-assembler-times {\tvdiv\.vv} 10 } } */ +/* { dg-final { scan-assembler-times {\tvdivu\.vv} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv64gcv.c index 7d9b75ae0b1..16a18c466e0 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv64gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-rv64gcv.c @@ -3,7 +3,8 @@ #include "vdiv-template.h" -/* TODO: Implement vector type promotion. We should have 6 vdiv.vv here. */ +/* Currently we use an epilogue loop which also contains vdivs. Therefore we + expect 10 vdiv[u]s instead of 6. */ -/* { dg-final { scan-assembler-times {\tvdiv\.vv} 4 } } */ -/* { dg-final { scan-assembler-times {\tvdivu\.vv} 6 } } */ +/* { dg-final { scan-assembler-times {\tvdiv\.vv} 10 } } */ +/* { dg-final { scan-assembler-times {\tvdivu\.vv} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-template.h b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-template.h index 12a1de32874..f8d3bfde4ed 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-template.h +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vdiv-template.h @@ -2,7 +2,7 @@ #define TEST_TYPE(TYPE) \ __attribute__((noipa)) \ - void vadd_##TYPE (TYPE *dst, TYPE *a, TYPE *b, int n) \ + void vdiv_##TYPE (TYPE *dst, TYPE *a, TYPE *b, int n) \ { \ for (int i = 0; i < n; i++) \ dst[i] = a[i] / b[i]; \ @@ -10,13 +10,12 @@ #define TEST2_TYPE(TYPE) \ __attribute__((noipa)) \ - void vadds_##TYPE (TYPE *dst, TYPE *a, TYPE b, int n) \ + void vdivs_##TYPE (TYPE *dst, TYPE *a, TYPE b, int n) \ { \ for (int i = 0; i < n; i++) \ dst[i] = a[i] / b; \ } -/* *int8_t not autovec currently. */ #define TEST_ALL() \ TEST_TYPE(int16_t) \ TEST_TYPE(uint16_t) \ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv32gcv.c index 28cba510a93..df99f5019fb 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv32gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv32gcv.c @@ -3,7 +3,8 @@ #include "vrem-template.h" -/* TODO: Implement vector type promotion. We should have 6 vrem.vv here. */ +/* Currently we use an epilogue loop which also contains vrems. Therefore we + expect 10 vrem[u]s instead of 6. */ -/* { dg-final { scan-assembler-times {\tvrem\.vv} 5 } } */ -/* { dg-final { scan-assembler-times {\tvremu\.vv} 6 } } */ +/* { dg-final { scan-assembler-times {\tvrem\.vv} 10 } } */ +/* { dg-final { scan-assembler-times {\tvremu\.vv} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv64gcv.c index 5b6961d1f63..3cff13a47e4 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv64gcv.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/binop/vrem-rv64gcv.c @@ -3,7 +3,8 @@ #include "vrem-template.h" -/* TODO: Implement vector type promotion. We should have 6 vrem.vv here. */ +/* Currently we use an epilogue loop which also contains vrems. Therefore we + expect 10 vrem[u]s instead of 6. */ -/* { dg-final { scan-assembler-times {\tvrem\.vv} 5 } } */ -/* { dg-final { scan-assembler-times {\tvremu\.vv} 6 } } */ +/* { dg-final { scan-assembler-times {\tvrem\.vv} 10 } } */ +/* { dg-final { scan-assembler-times {\tvremu\.vv} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-run.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-run.c new file mode 100644 index 00000000000..f55d2dfce7f --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-run.c @@ -0,0 +1,35 @@ +/* { dg-do run { target { riscv_vector } } } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vncvt-template.h" + +#include + +#define SZ 256 + +#define RUN(TYPE1,TYPE2) \ + TYPE1 src##TYPE1##TYPE2[SZ]; \ + TYPE2 dst##TYPE1##TYPE2[SZ]; \ + for (int i = 0; i < SZ; i++) \ + { \ + src##TYPE1##TYPE2[i] = i; \ + dst##TYPE1##TYPE2[i] = -1; \ + } \ + vncvt_##TYPE1##TYPE2 (dst##TYPE1##TYPE2, \ + src##TYPE1##TYPE2, SZ); \ + for (int i = 0; i < SZ; i++) \ + assert (dst##TYPE1##TYPE2[i] == i); + + +#define RUN_ALL() \ + RUN(uint16_t, uint8_t) \ + RUN(uint32_t, uint8_t) \ + RUN(uint64_t, uint8_t) \ + RUN(uint32_t, uint16_t) \ + RUN(uint64_t, uint16_t) \ + RUN(uint64_t, uint32_t) \ + +int main () +{ + RUN_ALL() +} diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv32gcv.c new file mode 100644 index 00000000000..2b5aa0051cf --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv32gcv.c @@ -0,0 +1,6 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv -mabi=ilp32d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vncvt-template.h" + +/* { dg-final { scan-assembler-times {\tvncvt.x.x.w} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv64gcv.c new file mode 100644 index 00000000000..29349b33da6 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-rv64gcv.c @@ -0,0 +1,6 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv -mabi=lp64d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vncvt-template.h" + +/* { dg-final { scan-assembler-times {\tvncvt.x.x.w} 10 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-template.h b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-template.h new file mode 100644 index 00000000000..6b19ff12abe --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vncvt-template.h @@ -0,0 +1,19 @@ +#include + +#define TEST(TYPE1, TYPE2) \ + __attribute__((noipa)) \ + void vncvt_##TYPE1##TYPE2 (TYPE2 *dst, TYPE1 *a, int n) \ + { \ + for (int i = 0; i < n; i++) \ + dst[i] = (TYPE1)a[i]; \ + } + +#define TEST_ALL() \ + TEST(uint16_t, uint8_t) \ + TEST(uint32_t, uint8_t) \ + TEST(uint32_t, uint16_t) \ + TEST(uint64_t, uint8_t) \ + TEST(uint64_t, uint16_t) \ + TEST(uint64_t, uint32_t) \ + +TEST_ALL() diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-run.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-run.c new file mode 100644 index 00000000000..d5f0190957a --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-run.c @@ -0,0 +1,35 @@ +/* { dg-do run { target { riscv_vector } } } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vsext-template.h" + +#include + +#define SZ 256 + +#define RUN(TYPE1,TYPE2) \ + TYPE1 src##TYPE1##TYPE2[SZ]; \ + TYPE2 dst##TYPE1##TYPE2[SZ]; \ + for (int i = 0; i < SZ; i++) \ + { \ + src##TYPE1##TYPE2[i] = i - 128; \ + dst##TYPE1##TYPE2[i] = 0; \ + } \ + vsext_##TYPE1##TYPE2 (dst##TYPE1##TYPE2, \ + src##TYPE1##TYPE2, SZ); \ + for (int i = 0; i < SZ; i++) \ + assert (dst##TYPE1##TYPE2[i] == i - 128); + + +#define RUN_ALL() \ + RUN(int8_t, int16_t) \ + RUN(int8_t, int32_t) \ + RUN(int8_t, int64_t) \ + RUN(int16_t, int32_t) \ + RUN(int16_t, int64_t) \ + RUN(int32_t, int64_t) \ + +int main () +{ + RUN_ALL() +} diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv32gcv.c new file mode 100644 index 00000000000..538216ab9c3 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv32gcv.c @@ -0,0 +1,8 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv -mabi=ilp32d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vsext-template.h" + +/* { dg-final { scan-assembler-times {\tvsext\.vf2} 3 } } */ +/* { dg-final { scan-assembler-times {\tvsext\.vf4} 2 } } */ +/* { dg-final { scan-assembler-times {\tvsext\.vf8} 1 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv64gcv.c new file mode 100644 index 00000000000..29348cc67e5 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-rv64gcv.c @@ -0,0 +1,8 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv -mabi=lp64d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vsext-template.h" + +/* { dg-final { scan-assembler-times {\tvsext\.vf2} 3 } } */ +/* { dg-final { scan-assembler-times {\tvsext\.vf4} 2 } } */ +/* { dg-final { scan-assembler-times {\tvsext\.vf8} 1 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-template.h b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-template.h new file mode 100644 index 00000000000..c2f5fc92c99 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vsext-template.h @@ -0,0 +1,19 @@ +#include + +#define TEST(TYPE1, TYPE2) \ + __attribute__((noipa)) \ + void vsext_##TYPE1##TYPE2 (TYPE2 *dst, TYPE1 *a, int n) \ + { \ + for (int i = 0; i < n; i++) \ + dst[i] = (TYPE1)a[i]; \ + } + +#define TEST_ALL() \ + TEST(int8_t, int16_t) \ + TEST(int8_t, int32_t) \ + TEST(int8_t, int64_t) \ + TEST(int16_t, int32_t) \ + TEST(int16_t, int64_t) \ + TEST(int32_t, int64_t) \ + +TEST_ALL() diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-run.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-run.c new file mode 100644 index 00000000000..9d1c259f592 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-run.c @@ -0,0 +1,35 @@ +/* { dg-do run { target { riscv_vector } } } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vzext-template.h" + +#include + +#define SZ 256 + +#define RUN(TYPE1,TYPE2) \ + TYPE1 src##TYPE1##TYPE2[SZ]; \ + TYPE2 dst##TYPE1##TYPE2[SZ]; \ + for (int i = 0; i < SZ; i++) \ + { \ + src##TYPE1##TYPE2[i] = i; \ + dst##TYPE1##TYPE2[i] = -1; \ + } \ + vzext_##TYPE1##TYPE2 (dst##TYPE1##TYPE2, \ + src##TYPE1##TYPE2, SZ); \ + for (int i = 0; i < SZ; i++) \ + assert (dst##TYPE1##TYPE2[i] == i); + + +#define RUN_ALL() \ + RUN(uint8_t, uint16_t) \ + RUN(uint8_t, uint32_t) \ + RUN(uint8_t, uint64_t) \ + RUN(uint16_t, uint32_t) \ + RUN(uint16_t, uint64_t) \ + RUN(uint32_t, uint64_t) \ + +int main () +{ + RUN_ALL() +} diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv32gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv32gcv.c new file mode 100644 index 00000000000..3e92843a5c2 --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv32gcv.c @@ -0,0 +1,8 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv32gcv -mabi=ilp32d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vzext-template.h" + +/* { dg-final { scan-assembler-times {\tvzext\.vf2} 3 } } */ +/* { dg-final { scan-assembler-times {\tvzext\.vf4} 2 } } */ +/* { dg-final { scan-assembler-times {\tvzext\.vf8} 1 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv64gcv.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv64gcv.c new file mode 100644 index 00000000000..cee0012d58c --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-rv64gcv.c @@ -0,0 +1,8 @@ +/* { dg-do compile } */ +/* { dg-additional-options "-std=c99 -fno-vect-cost-model -march=rv64gcv -mabi=lp64d --param=riscv-autovec-preference=fixed-vlmax" } */ + +#include "vzext-template.h" + +/* { dg-final { scan-assembler-times {\tvzext\.vf2} 3 } } */ +/* { dg-final { scan-assembler-times {\tvzext\.vf4} 2 } } */ +/* { dg-final { scan-assembler-times {\tvzext\.vf8} 1 } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-template.h b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-template.h new file mode 100644 index 00000000000..847905b690c --- /dev/null +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/conversions/vzext-template.h @@ -0,0 +1,19 @@ +#include + +#define TEST(TYPE1, TYPE2) \ + __attribute__((noipa)) \ + void vzext_##TYPE1##TYPE2 (TYPE2 *dst, TYPE1 *a, int n) \ + { \ + for (int i = 0; i < n; i++) \ + dst[i] = (TYPE1)a[i]; \ + } + +#define TEST_ALL() \ + TEST(uint8_t, uint16_t) \ + TEST(uint8_t, uint32_t) \ + TEST(uint8_t, uint64_t) \ + TEST(uint16_t, uint32_t) \ + TEST(uint16_t, uint64_t) \ + TEST(uint32_t, uint64_t) \ + +TEST_ALL() diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32f_zvl128b-2.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32f_zvl128b-2.c index c53975e9b01..7f499befa82 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32f_zvl128b-2.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32f_zvl128b-2.c @@ -3,4 +3,4 @@ #include "template-1.h" -/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 4 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 5 "vect" } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32x_zvl128b-2.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32x_zvl128b-2.c index 80a4796fd38..2de09a29f02 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32x_zvl128b-2.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve32x_zvl128b-2.c @@ -3,4 +3,4 @@ #include "template-1.h" -/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 2 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 3 "vect" } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64d-2.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64d-2.c index 5e38b41a5c3..95d54d7b281 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64d-2.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64d-2.c @@ -3,4 +3,4 @@ #include "template-1.h" -/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 3 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 6 "vect" } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64f-2.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64f-2.c index ee37282f1f8..f9f44a94902 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64f-2.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64f-2.c @@ -3,4 +3,4 @@ #include "template-1.h" -/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 3 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 5 "vect" } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64x-2.c b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64x-2.c index 6a64a1a1fdf..12703a7e036 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64x-2.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/autovec/zve64x-2.c @@ -3,4 +3,4 @@ #include "template-1.h" -/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 2 "vect" } } */ +/* { dg-final { scan-tree-dump-times "vectorized 1 loops in function" 3 "vect" } } */ diff --git a/gcc/testsuite/gcc.target/riscv/rvv/rvv.exp b/gcc/testsuite/gcc.target/riscv/rvv/rvv.exp index 9809a421fc8..9466c2032e3 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/rvv.exp +++ b/gcc/testsuite/gcc.target/riscv/rvv/rvv.exp @@ -65,6 +65,8 @@ foreach op $AUTOVEC_TEST_OPTS { "" "$op" dg-runtest [lsort [glob -nocomplain $srcdir/$subdir/autovec/cmp/*.\[cS\]]] \ "" "$op" + dg-runtest [lsort [glob -nocomplain $srcdir/$subdir/autovec/conversions/*.\[cS\]]] \ + "" "$op" } # VLS-VLMAX tests diff --git a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-38.c b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-38.c index 34f1cd43a21..8606b10268d 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-38.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-38.c @@ -1,5 +1,5 @@ /* { dg-do compile } */ -/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2" } */ +/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ #include "riscv_vector.h" diff --git a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-47.c b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-47.c index 935e1b10630..11403203aa9 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-47.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-47.c @@ -1,5 +1,5 @@ /* { dg-do compile } */ -/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2" } */ +/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ #include "riscv_vector.h" diff --git a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-48.c b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-48.c index a1f4b70d4b4..79af2ef450a 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-48.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-48.c @@ -1,5 +1,5 @@ /* { dg-do compile } */ -/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2" } */ +/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ #include "riscv_vector.h" diff --git a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-49.c b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-49.c index 4fa68627cc0..77fe05b68a7 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-49.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/avl_single-49.c @@ -1,5 +1,5 @@ /* { dg-do compile } */ -/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2" } */ +/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ #include "riscv_vector.h" diff --git a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/imm_switch-8.c b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/imm_switch-8.c index 4a1ef3ce5e9..f8568a8c898 100644 --- a/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/imm_switch-8.c +++ b/gcc/testsuite/gcc.target/riscv/rvv/vsetvl/imm_switch-8.c @@ -1,5 +1,5 @@ /* { dg-do compile } */ -/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2" } */ +/* { dg-options "-march=rv32gcv -mabi=ilp32 -fno-schedule-insns -fno-schedule-insns2 -fno-tree-vectorize" } */ #include "riscv_vector.h"