From patchwork Sat Sep 30 17:45:54 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: =?utf-8?q?R=C3=A9mi_Denis-Courmont?= X-Patchwork-Id: 44053 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a05:6a20:1204:b0:15d:8365:d4b8 with SMTP id v4csp507486pzf; Sat, 30 Sep 2023 10:46:06 -0700 (PDT) X-Google-Smtp-Source: AGHT+IG+7WIx9zgGU+XcavUVX2FmMI4zMANLuYlrwT1wILT8zPb+L69FaCRPMegvPPPbNPEl4F8W X-Received: by 2002:aa7:cf92:0:b0:533:795d:8c48 with SMTP id z18-20020aa7cf92000000b00533795d8c48mr6788394edx.10.1696095966341; Sat, 30 Sep 2023 10:46:06 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1696095966; cv=none; d=google.com; s=arc-20160816; b=wb2mZtrK62d1cPLo5akkhzZIkl533NxUUcZr4nqMWrgWS+FYhV/le3VbLO+zpyE2xe a64j4+yhX4ffJdCgpBSQVI6e1+QJv/A8fy+Tc2182s+UwjQeTe7Jpq+sGPxL10K8KLwX GFOtDzqZSUQOi+CqaCfQgLkL62ZtNW2ydh3WKXVmGfnIzTZd37UUTCTcBtunHyUQfrH2 jKvWLxTh3sdt+vxK93bgkE7WIEKXnI2aP9NrOnC2H5AFPYGEcS9Q1ELc3AYAdlVCPH28 9jcFMdu+v9IQg0YZ34bORL1VqAcqoFGpyZm0kT172myigAdT+PAZvOhu8LyCtbCOvUvL BrbA== 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=Dp6z18QavXCHITfDzbRTieNLomsJkXGtKo3ktGr380w=; fh=YOA8vD9MJZuwZ71F/05pj6KdCjf6jQRmzLS+CATXUQk=; b=o476Io/fa1T3ajN+hLTGzhZMozQiWUa2VXRo/8nDwO3RQu1TBnAL2QAmb+BdF01uWh XeAOqkK8ueA+UMkY5NJ+O/ChVcSrYZM5TVwhdZ2J0WN/3MZEUuKzpmhqBDHRdPbI8o0L GWHROYDR8v1764z+jNe2qyE+xEqJqZEFa8yp27kQZ4R2LGtavcmLSEVRwOV7VGPtb9pw VTgZon+2hwPwKA61wi3RbcPn1cox4jkDpLHpBrfAV/ERUed/8DEgC43hFKoYnMryLXG3 U/rNJgik/SYxXxIF9JzWNwjNQv4jw+Wx2Ieb/qsP3JYU5OQlf7CkvgGMHHZdBrDNEEzC oSDQ== 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 q2-20020aa7da82000000b0053424a6c4efsi10015881eds.210.2023.09.30.10.46.05; Sat, 30 Sep 2023 10:46:06 -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 C6A2E68CBE8; Sat, 30 Sep 2023 20:46:01 +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 760A268CC0D for ; Sat, 30 Sep 2023 20:45:55 +0300 (EEST) Received: from basile.remlab.net (localhost [IPv6:::1]) by ursule.remlab.net (Postfix) with ESMTP id 7A292C006C for ; Sat, 30 Sep 2023 20:45:54 +0300 (EEST) From: =?utf-8?q?R=C3=A9mi_Denis-Courmont?= To: ffmpeg-devel@ffmpeg.org Date: Sat, 30 Sep 2023 20:45:54 +0300 Message-ID: <20230930174554.8650-1-remi@remlab.net> X-Mailer: git-send-email 2.42.0 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH] lavu/float_dsp: avoid reg-stride in R-V V reverse_fmul 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: sXDUwA4QChqs This revectors the inner loop to reverse vectors element in vectors, thus eliminating the negative register stride. Note that RVV does not have a vector reverse instruction, so this uses a gather. --- libavutil/riscv/float_dsp_rvv.S | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/libavutil/riscv/float_dsp_rvv.S b/libavutil/riscv/float_dsp_rvv.S index 7cfc890bc2..7e9e84d526 100644 --- a/libavutil/riscv/float_dsp_rvv.S +++ b/libavutil/riscv/float_dsp_rvv.S @@ -125,20 +125,25 @@ func ff_vector_fmul_add_rvv, zve32f ret endfunc +// TODO factor vrsub, separate last iteration? // (a0) = (a1) * reverse(a2) [0..a3-1] func ff_vector_fmul_reverse_rvv, zve32f + vsetvli t0, zero, e16, m4, ta, ma sh2add a2, a3, a2 - li t2, -4 // byte stride - addi a2, a2, -4 + vid.v v0 + vadd.vi v0, v0, 1 1: - vsetvli t0, a3, e32, m8, ta, ma + vsetvli t0, a3, e16, m4, ta, ma slli t1, t0, 2 - vle32.v v16, (a1) + vrsub.vx v4, v0, t0 // v4[i] = [VL-1, VL-2... 1, 0] + sub a2, a2, t1 + vsetvli zero, zero, e32, m8, ta, ma + vle32.v v8, (a2) sub a3, a3, t0 - vlse32.v v24, (a2), t2 + vle32.v v16, (a1) add a1, a1, t1 + vrgatherei16.vv v24, v8, v4 // v24 = reverse(v8) vfmul.vv v16, v16, v24 - sub a2, a2, t1 vse32.v v16, (a0) add a0, a0, t1 bnez a3, 1b