From patchwork Fri Sep 9 15:48:53 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: =?utf-8?q?R=C3=A9mi_Denis-Courmont?= X-Patchwork-Id: 37807 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a05:6a20:139a:b0:8f:1db5:eae2 with SMTP id w26csp998748pzh; Fri, 9 Sep 2022 08:51:35 -0700 (PDT) X-Google-Smtp-Source: AA6agR4WVBkmEXfclaLmlEIfYWMXmNK84TINbsw+M5Bi3mlKDUuuhra2MixnY2+yFA/mc1sxfWpz X-Received: by 2002:a05:6402:1a4f:b0:44e:f731:f7d5 with SMTP id bf15-20020a0564021a4f00b0044ef731f7d5mr11831220edb.357.1662738694887; Fri, 09 Sep 2022 08:51:34 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1662738694; cv=none; d=google.com; s=arc-20160816; b=hsWKHktWdHIhW+gPzq9zgHGZV1dWnR8KwYSfYkDwuBjR9/ykVbW8Lu2Dm1d2b4kV+/ 8FWEFTfrASQlcsgMbWChihw2YklTLRE60cpkOprwAMf3ZjcLlp1KCq5akCgQ0dRfWdav 9I9Cyzs/Phqjd1iU2+36oa5F7IcTVtPChu3WaFmTDzK4oHbafd0TnxK1O+2mgQa1s8cO Kl0vy0/nm/cDV0bJ7y5zaKwxH6LMUGkij8eg3jriwNmNFrASwaFp1aQFXpKiA4KEKBUm s9tmso/h3JcOqJxlKvwD5vHwO2H92ORvGSVL05GrQACNwpnqmlNmoQ2Xcpv/LzGuKa4R Oj9w== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:content-transfer-encoding:reply-to:list-subscribe :list-help:list-post:list-archive:list-unsubscribe:list-id :precedence:subject:mime-version:message-id:date:to:from :delivered-to; bh=jWABRTLogK9lpFxfzusEOrTC2Ei/RwvEISAqs5Qx+nA=; b=Z/uwvNG3BIO7SfVucjC7SNo4zKY4AzF9ePW1hlDxvMyqnXet4Tkv3sLnkwnSXMJrCh v60smch2zYnpkdmrd9uPYsiACy4olMHyucS7ncF+QE34mdhPITBtmkOdfi4XEUlX4Ysr yYqtY6jDDnx/ksLeRzKJzsKmhkv7vyV/iLtstoU2qlHzfAJSDYEspBo7W3YVyxlmbAUp oy13wThjMNPY6DmCL2EzWIyVEeuNI4ohR8Ixdvw/IN32XWeNwGlSbdxeO5hJPlUDVtbw 9lZOPUbkQkipwg/Fd61n6mHfCFZGCU4JYOZic/WxPUNzsCkmDbFOMdKd58dDkgD8fXPV /l9Q== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) smtp.mailfrom=ffmpeg-devel-bounces@ffmpeg.org Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id i18-20020a17090671d200b0076ed46e4445si622393ejk.810.2022.09.09.08.51.34; Fri, 09 Sep 2022 08:51:34 -0700 (PDT) Received-SPF: pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) client-ip=79.124.17.100; Authentication-Results: mx.google.com; spf=pass (google.com: domain of ffmpeg-devel-bounces@ffmpeg.org designates 79.124.17.100 as permitted sender) smtp.mailfrom=ffmpeg-devel-bounces@ffmpeg.org Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 19AA068BBA2; Fri, 9 Sep 2022 18:49:23 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from ursule.remlab.net (vps-a2bccee9.vps.ovh.net [51.75.19.47]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 6177E68BB4B for ; Fri, 9 Sep 2022 18:49:05 +0300 (EEST) Received: from basile.remlab.net (localhost [IPv6:::1]) by ursule.remlab.net (Postfix) with ESMTP id 4BD8BC00B8 for ; Fri, 9 Sep 2022 18:49:01 +0300 (EEST) From: remi@remlab.net To: ffmpeg-devel@ffmpeg.org Date: Fri, 9 Sep 2022 18:48:53 +0300 Message-Id: <20220909154859.68954-12-remi@remlab.net> X-Mailer: git-send-email 2.37.2 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 12/18] lavu/riscv: float vector multiply-accumulate with RVV X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" X-TUID: QbisStHOFzel From: Rémi Denis-Courmont --- libavutil/riscv/float_dsp_init.c | 6 +++++ libavutil/riscv/float_dsp_rvv.S | 38 ++++++++++++++++++++++++++++++++ 2 files changed, 44 insertions(+) diff --git a/libavutil/riscv/float_dsp_init.c b/libavutil/riscv/float_dsp_init.c index 49a4c95a0b..b63da72acd 100644 --- a/libavutil/riscv/float_dsp_init.c +++ b/libavutil/riscv/float_dsp_init.c @@ -24,11 +24,15 @@ void ff_vector_fmul_rvv(float *dst, const float *src0, const float *src1, int len); +void ff_vector_fmac_scalar_rvv(float *dst, const float *src, float mul, + int len); void ff_vector_fmul_scalar_rvv(float *dst, const float *src, float mul, int len); void ff_vector_dmul_rvv(double *dst, const double *src0, const double *src1, int len); +void ff_vector_dmac_scalar_rvv(double *dst, const double *src, double mul, + int len); void ff_vector_dmul_scalar_rvv(double *dst, const double *src, double mul, int len); @@ -39,10 +43,12 @@ av_cold void ff_float_dsp_init_riscv(AVFloatDSPContext *fdsp) if (flags & AV_CPU_FLAG_ZVE32F) { fdsp->vector_fmul = ff_vector_fmul_rvv; + fdsp->vector_fmac_scalar = ff_vector_fmac_scalar_rvv; fdsp->vector_fmul_scalar = ff_vector_fmul_scalar_rvv; if (flags & AV_CPU_FLAG_ZVE64D) { fdsp->vector_dmul = ff_vector_dmul_rvv; + fdsp->vector_dmac_scalar = ff_vector_dmac_scalar_rvv; fdsp->vector_dmul_scalar = ff_vector_dmul_scalar_rvv; } } diff --git a/libavutil/riscv/float_dsp_rvv.S b/libavutil/riscv/float_dsp_rvv.S index 65c3a77b01..5a7d92abd6 100644 --- a/libavutil/riscv/float_dsp_rvv.S +++ b/libavutil/riscv/float_dsp_rvv.S @@ -36,6 +36,25 @@ func ff_vector_fmul_rvv, zve32f ret endfunc +// (a0) += (a1) * fa0 [0..a2-1] +func ff_vector_fmac_scalar_rvv, zve32f +NOHWF fmv.w.x fa0, a2 +NOHWF mv a2, a3 + +1: vsetvli t0, a2, e32, m8, ta, ma + slli t1, t0, 2 + vle32.v v24, (a1) + add a1, a1, t1 + vle32.v v16, (a0) + vfmacc.vf v16, fa0, v24 + sub a2, a2, t0 + vse32.v v16, (a0) + add a0, a0, t1 + bnez a2, 1b + + ret +endfunc + // (a0) = (a1) * fa0 [0..a2-1] func ff_vector_fmul_scalar_rvv, zve32f NOHWF fmv.w.x fa0, a2 @@ -71,6 +90,25 @@ func ff_vector_dmul_rvv, zve64d ret endfunc +// (a0) += (a1) * fa0 [0..a2-1] +func ff_vector_dmac_scalar_rvv, zve64d +NOHWD fmv.d.x fa0, a2 +NOHWD mv a2, a3 + +1: vsetvli t0, a2, e64, m8, ta, ma + slli t1, t0, 3 + vle64.v v24, (a1) + add a1, a1, t1 + vle64.v v16, (a0) + vfmacc.vf v16, fa0, v24 + sub a2, a2, t0 + vse64.v v16, (a0) + add a0, a0, t1 + bnez a2, 1b + + ret +endfunc + // (a0) = (a1) * fa0 [0..a2-1] func ff_vector_dmul_scalar_rvv, zve64d NOHWD fmv.d.x fa0, a2