From patchwork Wed May 29 17:15:37 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: 49356 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a59:8f0d:0:b0:460:55fa:d5ed with SMTP id i13csp805217vqu; Wed, 29 May 2024 10:16:36 -0700 (PDT) X-Forwarded-Encrypted: i=2; AJvYcCUDGYFLvSXqcYZD2MZctNptk0LIKzHm6iWjycmlAk69T/gfgOxXqf2/yp8U9QE43MHWD1tTuMW340lb+jrNtF/iMqJZI6Xf1aSfaQ== X-Google-Smtp-Source: AGHT+IGZJgtZdafl3jp+o2J5qar5Sg4sALtNS1StEAXHWy9k0llgaKqdNpvTHRooUVaAUfz6TjIQ X-Received: by 2002:a17:906:71b:b0:a64:e464:7123 with SMTP id a640c23a62f3a-a64e46471efmr153577266b.7.1717002996006; Wed, 29 May 2024 10:16:36 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1717002995; cv=none; d=google.com; s=arc-20160816; b=o5oILqJlA2+Pbp5pwG0jlDCpUA+/+mye5a0q+t4XeajXQnNKahhB0QKMzC99l21bTw xAP8nH9nSAa+/A5JchTsI7NjIQpjdjdlWUhFzvHn59Hg1yPrn5yUbs9WONEAWxHenzBS zpOqrNbjGWnda2NnVfgajK9c8WANcKsAjKFVfK6DcI++8STAB+v51OGlyZcUT8XZg7qM s1Sm9FAAtgDv4D29MgrWQDXNiyuOk/TFJM7ooxe4i3MCuS7ZiOfFjYq/h4OYrjes5SgH iP8rH9hJqiuDVCNb/YX2DyjGqvsLe39MIXgxSUvDqA/iYJBrBLXwiImfemMhZo5NN4ev w3kQ== 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=+Wqv9iK6nYNYRhSfO0Qz/2Ag0JmZwJKo83rnqe0Tp4E=; fh=D0bFwGkf4X22/D/bfeDVrXKIx7S6kcXsNzy10j8ORbQ=; b=y9e0kii1Z3pRtwBsCpNiWVvANzz8axLksuDkRVs0Yd5zJ25G/Ly5VIhBrKsoDnBB6w TFPCR1p76z/kgovYkGd10amIwlZDnDJ0q6+bGEO9qbLdLHPesT1poR02jumzE0Y80uOl G5xhHqP0Hs33IRjhoSdEWp7llYlRnSLMp/3qHRawi/KuKig+hvVk9vpBO7ifgp/NWvrl Z4Uxi+4J8WULcuLE2YhD0qXSyyvUrA0KSAYJmoF4MSWAchljNS0ESVaosVtnm6wqhQYA CSAb4vMlr1ooVJ8UZOnO0ovPECo4Z+WWsvF7yKPTuZHuBh00V2tDz9UGCJItkQCNpJt2 B96w==; 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=purKfssv; 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 a640c23a62f3a-a626cd90567si629710866b.702.2024.05.29.10.16.35; Wed, 29 May 2024 10:16:35 -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=purKfssv; 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 DDA1E68D52E; Wed, 29 May 2024 20:16:31 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from out203-205-221-205.mail.qq.com (out203-205-221-205.mail.qq.com [203.205.221.205]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id DB22D68D4E9 for ; Wed, 29 May 2024 20:16:21 +0300 (EEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=foxmail.com; s=s201512; t=1717002973; bh=5z1z8I8xyi8e3T714I2WJSm27td3cuW7k0lN2iTDHqM=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=purKfssvU4dNWw/FbQf0QS0y8LaIfruomCQyyo87LQLwFjN/9tmtm15DJ5nFdoYsQ gE/x3QBNDHfhISl2NnuTUdY41Se1qAxWplsOqD+mkJhQgcZvhOLozMQMrBdruNrSsD ad/YQXzTV1XZjzlzcRieoHArnSFIAlOJ1C9o+i6k= Received: from localhost.localdomain ([42.86.114.86]) by newxmesmtplogicsvrsza10-0.qq.com (NewEsmtp) with SMTP id 40A88E50; Thu, 30 May 2024 01:16:10 +0800 X-QQ-mid: xmsmtpt1717002972t5e4ppc43 Message-ID: X-QQ-XMAILINFO: M0PjjqbLT90wCYSuEpFX9CHxclNhSIqHEM+vVnGCqRnoRveOZ3uvIjuDKPaGdc Y6NYFNcV2cO3RzmmM6s9FVzQM7of+XwkIDjb9vZpON2Dqo2K0sYUAi7hS1LM2EWRM3lBAAjmRtd6 NOVBDGz1cDHKwAd9w/li04/DonJE3r9jWA2g1nsTkHqGRInZ+nXrk3KbRV71aaFS28kKW4cDtji0 1zfwFpcxteBg7UMF7Ukwnj5Yx5QJt2oZLOlP6dOGt5sSesHAqHTaRHE1iTmMfQamglqxVddQe6pE /wOXUNn2hpkPi5MSBEwUB8VUBUbvNTgwpx31lHeWHhntL0AlJFYYmJdrrTEzlqRbnsM43Po1ysqI 3cDM+jTxHXWjXGSdu3oA31m8KEg4hbJ1DoccNwNV97k0qDFqY7o9aJ1m3S4HtFqF3RYd4ZO5y70/ oZGFH4bwnUrbHIlDy0B2fsmbiHfg10gwdUAOy8ecuGu3euUhPKC2GmxloIB+5G/OCJVEDqO4RMxJ KgEz12wd4+SReuSAaAICM6nQ8jDL7Gkd/DmClADQhG4kyUNQ5r5xiWpmPAel+FaFg8Hvz+IUp353 Qdn9jDUP2alGar1ayf/Y5s+cwgFRRER7WLAESCTDBUQCiq5aUkKMquPHGTMZLsm8BdpVzTnMnYRW thBWs7vhRlCoN5qHoYwPqltIJfdESIu0wDT/D3LSqMS92sNpG5cKJhzlnnwuzCUxmZE/XNTr3x/u 4W0piLv/rk898BTvkEwbTet+P6eqiBWN/J6A59l8ZN1e1kflMv7OExgEDE9iDJYSFtZzFpeHJVwP Ysel1P49zjHLhrLtSwyk+Nv0VALXZlbnoplFrJlekQlPlayE8cgIMjXMgKnA+lnW86mMxrfHaZHP IdOQXjPZmKqwHXJI8MX15y1TwlKs7LRg== X-QQ-XMRINFO: Nq+8W0+stu50PRdwbJxPCL0= From: uk7b@foxmail.com To: ffmpeg-devel@ffmpeg.org Date: Thu, 30 May 2024 01:15:37 +0800 X-OQ-MSGID: <20240529171540.911099-2-uk7b@foxmail.com> X-Mailer: git-send-email 2.45.1 In-Reply-To: <20240529171540.911099-1-uk7b@foxmail.com> References: <20240529171540.911099-1-uk7b@foxmail.com> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH v3 2/5] lavc/vp9dsp: R-V V mc bilin h v 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: 7nTkiSA1iipN From: sunyuechi C908: vp9_avg_bilin_4h_8bpp_c: 5.2 vp9_avg_bilin_4h_8bpp_rvv_i64: 2.2 vp9_avg_bilin_4v_8bpp_c: 5.5 vp9_avg_bilin_4v_8bpp_rvv_i64: 2.2 vp9_avg_bilin_8h_8bpp_c: 20.0 vp9_avg_bilin_8h_8bpp_rvv_i64: 4.5 vp9_avg_bilin_8v_8bpp_c: 21.0 vp9_avg_bilin_8v_8bpp_rvv_i64: 4.2 vp9_avg_bilin_16h_8bpp_c: 78.2 vp9_avg_bilin_16h_8bpp_rvv_i64: 9.0 vp9_avg_bilin_16v_8bpp_c: 82.0 vp9_avg_bilin_16v_8bpp_rvv_i64: 9.0 vp9_avg_bilin_32h_8bpp_c: 325.5 vp9_avg_bilin_32h_8bpp_rvv_i64: 26.2 vp9_avg_bilin_32v_8bpp_c: 326.2 vp9_avg_bilin_32v_8bpp_rvv_i64: 26.2 vp9_avg_bilin_64h_8bpp_c: 1265.7 vp9_avg_bilin_64h_8bpp_rvv_i64: 91.5 vp9_avg_bilin_64v_8bpp_c: 1317.0 vp9_avg_bilin_64v_8bpp_rvv_i64: 91.2 vp9_put_bilin_4h_8bpp_c: 4.5 vp9_put_bilin_4h_8bpp_rvv_i64: 1.7 vp9_put_bilin_4v_8bpp_c: 4.7 vp9_put_bilin_4v_8bpp_rvv_i64: 1.7 vp9_put_bilin_8h_8bpp_c: 17.0 vp9_put_bilin_8h_8bpp_rvv_i64: 3.5 vp9_put_bilin_8v_8bpp_c: 18.0 vp9_put_bilin_8v_8bpp_rvv_i64: 3.5 vp9_put_bilin_16h_8bpp_c: 65.2 vp9_put_bilin_16h_8bpp_rvv_i64: 7.5 vp9_put_bilin_16v_8bpp_c: 85.7 vp9_put_bilin_16v_8bpp_rvv_i64: 7.5 vp9_put_bilin_32h_8bpp_c: 257.5 vp9_put_bilin_32h_8bpp_rvv_i64: 23.5 vp9_put_bilin_32v_8bpp_c: 274.5 vp9_put_bilin_32v_8bpp_rvv_i64: 23.5 vp9_put_bilin_64h_8bpp_c: 1040.5 vp9_put_bilin_64h_8bpp_rvv_i64: 82.5 vp9_put_bilin_64v_8bpp_c: 1108.7 vp9_put_bilin_64v_8bpp_rvv_i64: 82.2 --- libavcodec/riscv/vp9_mc_rvv.S | 60 ++++++++++++++++++++++++++++++++++ libavcodec/riscv/vp9dsp.h | 12 +++---- libavcodec/riscv/vp9dsp_init.c | 21 ++++++++++++ 3 files changed, 87 insertions(+), 6 deletions(-) diff --git a/libavcodec/riscv/vp9_mc_rvv.S b/libavcodec/riscv/vp9_mc_rvv.S index 7cb38ec94a..9611aba0ed 100644 --- a/libavcodec/riscv/vp9_mc_rvv.S +++ b/libavcodec/riscv/vp9_mc_rvv.S @@ -53,6 +53,66 @@ func ff_vp9_avg\len\()_rvv, zve32x endfunc .endm +.macro bilin_load dst, op, type, mn +.ifc \type,v + add t5, a2, a3 +.else + addi t5, a2, 1 +.endif + vle8.v v8, (a2) + vle8.v v0, (t5) + vwmulu.vx v16, v0, \mn + vwmaccsu.vx v16, t1, v8 + vwadd.wx v16, v16, t4 + vnsra.wi v16, v16, 4 + vadd.vv \dst, v16, v8 +.ifc \op,avg + vle8.v v16, (a0) + vaaddu.vv \dst, \dst, v16 +.endif +.endm + +.macro bilin_h_v op, type, mn +func ff_\op\()_vp9_bilin_4\type\()_rvv, zve32x + vsetvlstatic8 4, t0, 64 +.Lbilin_\type\op: +.ifc \op,avg + csrwi vxrm, 0 +.endif + li t4, 8 + neg t1, \mn +1: + addi a4, a4, -1 + bilin_load v0, \op, \type, \mn + vse8.v v0, (a0) + add a2, a2, a3 + add a0, a0, a1 + bnez a4, 1b + + ret +endfunc +.endm + .irp len, 64, 32, 16, 8, 4 copy_avg \len .endr + +bilin_h_v put, h, a5 +bilin_h_v avg, h, a5 +bilin_h_v put, v, a6 +bilin_h_v avg, v, a6 + +.macro func_bilin_h_v len, op, type +func ff_\op\()_vp9_bilin_\len\()\type\()_rvv, zve32x + vsetvlstatic8 \len, t0, 64 + j .Lbilin_\type\()\op +endfunc +.endm + +.irp len, 64, 32, 16, 8 + .irp op, put, avg + .irp type, h, v + func_bilin_h_v \len, \op, \type + .endr + .endr +.endr diff --git a/libavcodec/riscv/vp9dsp.h b/libavcodec/riscv/vp9dsp.h index ff8431591c..8fb326dae0 100644 --- a/libavcodec/riscv/vp9dsp.h +++ b/libavcodec/riscv/vp9dsp.h @@ -113,27 +113,27 @@ void ff_avg_8tap_##type##_##SIZE##hv_rvv(uint8_t *dst, ptrdiff_t dststride, \ int h, int mx, int my); #define VP9_BILINEAR_RISCV_RVV_FUNC(SIZE) \ -void ff_put_bilin_##SIZE##h_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_put_vp9_bilin_##SIZE##h_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); \ \ -void ff_put_bilin_##SIZE##v_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_put_vp9_bilin_##SIZE##v_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); \ \ -void ff_put_bilin_##SIZE##hv_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_put_vp9_bilin_##SIZE##hv_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); \ \ -void ff_avg_bilin_##SIZE##h_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_avg_vp9_bilin_##SIZE##h_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); \ \ -void ff_avg_bilin_##SIZE##v_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_avg_vp9_bilin_##SIZE##v_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); \ \ -void ff_avg_bilin_##SIZE##hv_rvv(uint8_t *dst, ptrdiff_t dststride, \ +void ff_avg_vp9_bilin_##SIZE##hv_rvv(uint8_t *dst, ptrdiff_t dststride, \ const uint8_t *src, ptrdiff_t srcstride, \ int h, int mx, int my); diff --git a/libavcodec/riscv/vp9dsp_init.c b/libavcodec/riscv/vp9dsp_init.c index 454dcd963f..9606d8545f 100644 --- a/libavcodec/riscv/vp9dsp_init.c +++ b/libavcodec/riscv/vp9dsp_init.c @@ -63,6 +63,27 @@ static av_cold void vp9dsp_mc_init_riscv(VP9DSPContext *dsp, int bpp) init_fpel(3, 8); init_fpel(4, 4); + dsp->mc[0][FILTER_BILINEAR ][0][0][1] = ff_put_vp9_bilin_64v_rvv; + dsp->mc[0][FILTER_BILINEAR ][0][1][0] = ff_put_vp9_bilin_64h_rvv; + dsp->mc[0][FILTER_BILINEAR ][1][0][1] = ff_avg_vp9_bilin_64v_rvv; + dsp->mc[0][FILTER_BILINEAR ][1][1][0] = ff_avg_vp9_bilin_64h_rvv; + dsp->mc[1][FILTER_BILINEAR ][0][0][1] = ff_put_vp9_bilin_32v_rvv; + dsp->mc[1][FILTER_BILINEAR ][0][1][0] = ff_put_vp9_bilin_32h_rvv; + dsp->mc[1][FILTER_BILINEAR ][1][0][1] = ff_avg_vp9_bilin_32v_rvv; + dsp->mc[1][FILTER_BILINEAR ][1][1][0] = ff_avg_vp9_bilin_32h_rvv; + dsp->mc[2][FILTER_BILINEAR ][0][0][1] = ff_put_vp9_bilin_16v_rvv; + dsp->mc[2][FILTER_BILINEAR ][0][1][0] = ff_put_vp9_bilin_16h_rvv; + dsp->mc[2][FILTER_BILINEAR ][1][0][1] = ff_avg_vp9_bilin_16v_rvv; + dsp->mc[2][FILTER_BILINEAR ][1][1][0] = ff_avg_vp9_bilin_16h_rvv; + dsp->mc[3][FILTER_BILINEAR ][0][0][1] = ff_put_vp9_bilin_8v_rvv; + dsp->mc[3][FILTER_BILINEAR ][0][1][0] = ff_put_vp9_bilin_8h_rvv; + dsp->mc[3][FILTER_BILINEAR ][1][0][1] = ff_avg_vp9_bilin_8v_rvv; + dsp->mc[3][FILTER_BILINEAR ][1][1][0] = ff_avg_vp9_bilin_8h_rvv; + dsp->mc[4][FILTER_BILINEAR ][0][0][1] = ff_put_vp9_bilin_4v_rvv; + dsp->mc[4][FILTER_BILINEAR ][0][1][0] = ff_put_vp9_bilin_4h_rvv; + dsp->mc[4][FILTER_BILINEAR ][1][0][1] = ff_avg_vp9_bilin_4v_rvv; + dsp->mc[4][FILTER_BILINEAR ][1][1][0] = ff_avg_vp9_bilin_4h_rvv; + #undef init_fpel } #endif