From patchwork Wed May 15 03:55:49 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: uk7b@foxmail.com X-Patchwork-Id: 48893 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a05:6a21:3a48:b0:1af:fc2d:ff5a with SMTP id zu8csp1353746pzb; Tue, 14 May 2024 20:56:15 -0700 (PDT) X-Forwarded-Encrypted: i=2; AJvYcCUEuJLUrEsVVUXF09iX9P3g7+5r0+9ZVKioh6BM3t/5PnY31eY8bBtV7zY2Y3Qh57urA2amiIr9CpA0pmPWEld9jZoe5N26UrPuIQ== X-Google-Smtp-Source: AGHT+IF4VBPwaqGPZ7j68XaucqirAI0jgjaCvEGsQpEg10fNQIofNmXqk6uXIRIKXBtpEs+ZG+PM X-Received: by 2002:a05:6402:8d0:b0:571:bff5:b793 with SMTP id 4fb4d7f45d1cf-5734d450172mr9788568a12.0.1715745375558; Tue, 14 May 2024 20:56:15 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1715745375; cv=none; d=google.com; s=arc-20160816; b=Iq/gfmX/FNkrvBfqB6k1CTIo37We2qlYnRUGYMpMq/eiXYHgAV8yFrSZXn8/eb/J63 bqGDMPAso2cSi672tTd/66wMLYQtk9ou+W5X1ael9BK7JC9wMd64VIUWpEulipP32ofU Cn6JW/goOj31du9atpFM8ri93Sjx65KwDo5pjNsyDWLAYaOYFCUczm2zYNVZZ5n4Ho2z JVxh/RJ73lUDNZoShoOTA1AKZqHblhKxBGLDbN8LS2FuM7ZlPJCdOiPOxCu6fPB6KqyC jmme8mkU216EW0pewvbVP55VLlkE6ncJoUexCCRiHJIweEl1KB6/94bSQvtwxonJWeFZ 2oEg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:content-transfer-encoding:cc:reply-to :list-subscribe:list-help:list-post:list-archive:list-unsubscribe :list-id:precedence:subject:mime-version:references:in-reply-to:date :to:from:message-id:dkim-signature:delivered-to; bh=at8jcy6NcvA6Z6e63Nq8DqdoJ+/sRI8cfbJ4ABC8NLA=; fh=D0bFwGkf4X22/D/bfeDVrXKIx7S6kcXsNzy10j8ORbQ=; b=rLew7byyIvBUSj93nghySngkpbwwlsDjwaVJ+DpuqvqNHnOxzSZmSLnj/kcKoplahK MLOwnP1DVOnW/k0IDrKhlMd7WXJrGPFDBjROIYOi9gc++XfQkirQyuDTvgyAhJj55e1t cWSb8QERA67wpM0RPOxWAtn1EUe98u+pFFhsdlSLg1fpg+VE1UihjQeysGZSxFup79io Cb16+sxbSxZOR4VvpCuJQplERc+joyVwgXXtC5idViSTzRDIdCFG+sJybORDCpHFTi/t /qQAMUtkalNcRCgff4foJ0Q2lo7a68hAik3k0xKvEYwpfh4tkw32SZfSn+2cE6Jnl1TF 6jbw==; dara=google.com ARC-Authentication-Results: i=1; mx.google.com; dkim=neutral (body hash did not verify) header.i=@foxmail.com header.s=s201512 header.b=jFwHi9LH; 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; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=foxmail.com Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id 4fb4d7f45d1cf-5733c2d5396si6960131a12.298.2024.05.14.20.56.15; Tue, 14 May 2024 20:56:15 -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; dkim=neutral (body hash did not verify) header.i=@foxmail.com header.s=s201512 header.b=jFwHi9LH; 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; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=foxmail.com Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 36A2068D359; Wed, 15 May 2024 06:56:11 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from out203-205-251-59.mail.qq.com (out203-205-251-59.mail.qq.com [203.205.251.59]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 5583168D359 for ; Wed, 15 May 2024 06:56:02 +0300 (EEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=foxmail.com; s=s201512; t=1715745352; bh=4LX50/I/fa0OKkZfgpMyrZ+SrDnj/XfO0KEWq0oCMJ0=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=jFwHi9LHgkyxwdyzDA64L3DlQr4QfUahgnmdOjj8PiEiE7RhrZfEvGIEwfau5ZWb3 oAhRlnmBe22H59fsTQHy4Bu6d8io0eyNhCv+GCGBfy/L8pwHmQFYaVsU2rG3SzoYyJ WXypc5aa/fkh86M77bGBEL9ciEiadlNTTgSFvcu4= Received: from localhost.localdomain ([42.56.223.122]) by newxmesmtplogicsvrszb16-1.qq.com (NewEsmtp) with SMTP id DF303EAF; Wed, 15 May 2024 11:55:51 +0800 X-QQ-mid: xmsmtpt1715745351tg2egpj72 Message-ID: X-QQ-XMAILINFO: MQspGcEBiT+Rk5ZQDWbscFQ6xtL6nCGfWpBXeTs3R5y02hnXP+akOlg7JxMnAW JmUMSluubM7I+o3N0fPRE+TAkfFC3pAr9DXZA0w6Dawl6zxAIyXaSDGKsBjhYONYkJafXvChixdf t5cyi+Lo2IKnplqbQ9UpYmfk25swuSbVygkk763PfHoSlHgEoKsMtWjDZiJkyi6M45tCV46KKVoE ZMAK9vbC1NY6vO4XNXAsTq3XtYJnzy40JWGt7xw+Zkt3o+LJZ2xrs36i/Kdf+UIKXGrZYl92B/F7 CSD1KqbxoCv9HAvsFOBxLnOKgcp06tXqNWODRrDFsrbeFjYvxWvbtSWJIN4IPkEfiW+OOX+GEAvG GwRliwVBjVJkeGMMBcnoa07WjdsXl/wLSd8Gdj1f5tw3632718BYknphfUN285IjcaafwuzUaELA Me/I1CcLn2nA+TzmU5nVjEJV41IPXxjSTwAjePUEnQ43hhCvwSXtg4VqPwU0nyI3XjiR+HqeGCha pa5hzgZmxxaHGlpyWF7lwwL/l0EzcARliPAnMImanQJ/MHma5K1LAYyMSdB13XYLCs+2yfjkJQV3 U1D4A0YtNuzHC/jp54qqXUQOgrI0Sqf75Bd6b3EsD/C/Gz/j+DB1h/qpPKYSboO3vHmnsV3TFY46 a5dn45ljk5fAdOHKyESi4zXq/C/cC++11yyr2WGkpANYy5XeUj4lmewslAnaTfnUURdgrD6ikV+a JE0nvZQmrhitUxSexlVJQL5XGNN521Grk90HH8xzgAOYrmt7X3LxTeWhn5lEffGAEyDeHyBfHzAR 4W87VIE6fvJFggqndTpMR3wzYTGXmx+SSJZuOquNePJoy+1qEDP6twa8csixIevXkWKi8l+MPUTs ySs4SJHRfug6Blj2hEhp+F/5RJ5ME+7iQkjexmsYKTSlzMb3Z9caRGYD7WfL8YGw== X-QQ-XMRINFO: NyFYKkN4Ny6FSmKK/uo/jdU= From: uk7b@foxmail.com To: ffmpeg-devel@ffmpeg.org Date: Wed, 15 May 2024 11:55:49 +0800 X-OQ-MSGID: <20240515035549.612197-1-uk7b@foxmail.com> X-Mailer: git-send-email 2.45.1 In-Reply-To: References: MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 4/9] lavc/vp9dsp: R-V V ipred tm 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 Cc: sunyuechi Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" X-TUID: qadQKMaYsxEB From: sunyuechi C908: vp9_tm_4x4_8bpp_c: 116.5 vp9_tm_4x4_8bpp_rvv_i32: 43.5 vp9_tm_8x8_8bpp_c: 416.2 vp9_tm_8x8_8bpp_rvv_i32: 86.0 vp9_tm_16x16_8bpp_c: 1665.5 vp9_tm_16x16_8bpp_rvv_i32: 187.2 vp9_tm_32x32_8bpp_c: 6974.2 vp9_tm_32x32_8bpp_rvv_i32: 625.7 --- libavcodec/riscv/vp9_intra_rvv.S | 118 +++++++++++++++++++++++++++++++ libavcodec/riscv/vp9dsp.h | 8 +++ libavcodec/riscv/vp9dsp_init.c | 4 ++ 3 files changed, 130 insertions(+) diff --git a/libavcodec/riscv/vp9_intra_rvv.S b/libavcodec/riscv/vp9_intra_rvv.S index ca156d65cd..280c497687 100644 --- a/libavcodec/riscv/vp9_intra_rvv.S +++ b/libavcodec/riscv/vp9_intra_rvv.S @@ -173,3 +173,121 @@ func ff_h_8x8_rvv, zve32x ret endfunc + +.macro tm_sum4 dst1, dst2, dst3, dst4, top, n1 + lbu t1, \n1(a2) + lbu t2, (\n1-1)(a2) + lbu t3, (\n1-2)(a2) + lbu t4, (\n1-3)(a2) + sub t1, t1, a4 + sub t2, t2, a4 + sub t3, t3, a4 + sub t4, t4, a4 + vadd.vx \dst1, \top, t1 + vadd.vx \dst2, \top, t2 + vadd.vx \dst3, \top, t3 + vadd.vx \dst4, \top, t4 +.endm + +func ff_tm_32x32_rvv, zve32x + lbu a4, -1(a3) + li t5, 32 + + .irp offset 31, 23, 15, 7 + vsetvli zero, t5, e16, m4, ta, ma + vle8.v v8, (a3) + vzext.vf2 v28, v8 + + tm_sum4 v0, v4, v8, v12, v28, \offset + tm_sum4 v16, v20, v24, v28, v28, (\offset-4) + + .irp n 0, 4, 8, 12, 16, 20, 24, 28 + vmax.vx v\n, v\n, zero + .endr + + vsetvli zero, zero, e8, m2, ta, ma + .irp n 0, 4, 8, 12, 16, 20, 24, 28 + vnclipu.wi v\n, v\n, 0 + vse8.v v\n, (a0) + add a0, a0, a1 + .endr + .endr + + ret +endfunc + +func ff_tm_16x16_rvv, zve32x + vsetivli zero, 16, e16, m2, ta, ma + vle8.v v8, (a3) + vzext.vf2 v30, v8 + lbu a4, -1(a3) + + tm_sum4 v0, v2, v4, v6, v30, 15 + tm_sum4 v8, v10, v12, v14, v30, 11 + tm_sum4 v16, v18, v20, v22, v30, 7 + tm_sum4 v24, v26, v28, v30, v30, 3 + + .irp n 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30 + vmax.vx v\n, v\n, zero + .endr + + vsetvli zero, zero, e8, m1, ta, ma + .irp n 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28 + vnclipu.wi v\n, v\n, 0 + vse8.v v\n, (a0) + add a0, a0, a1 + .endr + vnclipu.wi v30, v30, 0 + vse8.v v30, (a0) + + ret +endfunc + +func ff_tm_8x8_rvv, zve32x + vsetivli zero, 8, e16, m1, ta, ma + vle8.v v8, (a3) + vzext.vf2 v28, v8 + lbu a4, -1(a3) + + tm_sum4 v16, v17, v18, v19, v28, 7 + tm_sum4 v20, v21, v22, v23, v28, 3 + + .irp n 16, 17, 18, 19, 20, 21, 22, 23 + vmax.vx v\n, v\n, zero + .endr + + vsetvli zero, zero, e8, mf2, ta, ma + .irp n 16, 17, 18, 19, 20, 21, 22 + vnclipu.wi v\n, v\n, 0 + vse8.v v\n, (a0) + add a0, a0, a1 + .endr + vnclipu.wi v24, v23, 0 + vse8.v v24, (a0) + + ret +endfunc + +func ff_tm_4x4_rvv, zve32x + vsetivli zero, 4, e16, mf2, ta, ma + vle8.v v8, (a3) + vzext.vf2 v28, v8 + lbu a4, -1(a3) + + tm_sum4 v16, v17, v18, v19, v28, 3 + + .irp n 16, 17, 18, 19 + vmax.vx v\n, v\n, zero + .endr + + vsetvli zero, zero, e8, mf4, ta, ma + .irp n 16, 17, 18 + vnclipu.wi v\n, v\n, 0 + vse8.v v\n, (a0) + add a0, a0, a1 + .endr + vnclipu.wi v24, v19, 0 + vse8.v v24, (a0) + + ret +endfunc diff --git a/libavcodec/riscv/vp9dsp.h b/libavcodec/riscv/vp9dsp.h index 0ad961c7e0..79330b4968 100644 --- a/libavcodec/riscv/vp9dsp.h +++ b/libavcodec/riscv/vp9dsp.h @@ -72,6 +72,14 @@ void ff_h_16x16_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, const uint8_t *a); void ff_h_8x8_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, const uint8_t *a); +void ff_tm_32x32_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, + const uint8_t *a); +void ff_tm_16x16_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, + const uint8_t *a); +void ff_tm_8x8_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, + const uint8_t *a); +void ff_tm_4x4_rvv(uint8_t *dst, ptrdiff_t stride, const uint8_t *l, + const uint8_t *a); #define VP9_8TAP_RISCV_RVV_FUNC(SIZE, type, type_idx) \ void ff_put_8tap_##type##_##SIZE##h_rvv(uint8_t *dst, ptrdiff_t dststride, \ diff --git a/libavcodec/riscv/vp9dsp_init.c b/libavcodec/riscv/vp9dsp_init.c index eab3e9cb0a..184fadbaf7 100644 --- a/libavcodec/riscv/vp9dsp_init.c +++ b/libavcodec/riscv/vp9dsp_init.c @@ -89,6 +89,10 @@ static av_cold void vp9dsp_intrapred_init_riscv(VP9DSPContext *dsp, int bpp) dsp->intra_pred[TX_32X32][HOR_PRED] = ff_h_32x32_rvv; dsp->intra_pred[TX_16X16][HOR_PRED] = ff_h_16x16_rvv; dsp->intra_pred[TX_8X8][HOR_PRED] = ff_h_8x8_rvv; + dsp->intra_pred[TX_32X32][TM_VP8_PRED] = ff_tm_32x32_rvv; + dsp->intra_pred[TX_16X16][TM_VP8_PRED] = ff_tm_16x16_rvv; + dsp->intra_pred[TX_8X8][TM_VP8_PRED] = ff_tm_8x8_rvv; + dsp->intra_pred[TX_4X4][TM_VP8_PRED] = ff_tm_4x4_rvv; } #endif #endif