From patchwork Wed Feb 23 08:57:32 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Wu, Jianhua" X-Patchwork-Id: 34467 Delivered-To: ffmpegpatchwork2@gmail.com Received: by 2002:a05:6838:d078:0:0:0:0 with SMTP id x24csp632982nkx; Wed, 23 Feb 2022 00:58:19 -0800 (PST) X-Google-Smtp-Source: ABdhPJxAn4aSp+eP9/RlIq4RzhV18LdqnRA+jexDTd7+NaRae87mRI8m90k3+5zvhtNCHJYkWsMr X-Received: by 2002:a50:934b:0:b0:410:befb:cfd0 with SMTP id n11-20020a50934b000000b00410befbcfd0mr29850762eda.27.1645606699203; Wed, 23 Feb 2022 00:58:19 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1645606699; cv=none; d=google.com; s=arc-20160816; b=v5tJtd8Pn/8+XmQivMCVGm9E7l9f1PqnlM283G+JYTBpmKCPvt8Sb4qv5ef6pKlvvI 54UXHLyIj9XLzfJ/5gQOq+n9Q0v7s16CFYoDoKCG6xhgMR3cazyYMQNtle3dWbeA/ngJ vh9DkpPogP2tafmkMFqXkU7QNowjtexpYFnkviwXvxowu5lEj2ftrHqd1ctG5/6czMN9 CIylvPPBDFAaxX1ZDXrKG7ty81OCl8m10LR+4pkpWTokYeruETXSCFbmSMvuV4QAIA15 WCcfn5alebueVoix5p9yOYUUl8GcO8dDin+eNgn/oALrqgVhCQiaj218mzLeHGP+3bhb 66xg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:content-transfer-encoding:mime-version:cc:reply-to :list-subscribe:list-help:list-post:list-archive:list-unsubscribe :list-id:precedence:subject:references:in-reply-to:message-id:date :to:from:dkim-signature:delivered-to; bh=9TW38HCORQBdX1sEM6W/Nktdt0aboEBq1qod6o3H1IM=; b=09GpsAW7jr8Hsy4P42tHJH8S6EpOgANFu8rRP91XzYKSHp59zX0Eg4a0IG7b9/z2ik jlisekyNM+NRPlcQao+n+pjv/1QJ3iE1ttnLTCG+DHCL2M3onlavrIQg3peiMtz0VYhG MdooXkEnK+2YU9qBO+hfMxIrdAlJQ78+bPqYk8dovl/+LmsLuwxHffRv7DgCqkfwuU3A DVAOJDzGaSElL2fceQI+z0iPubLa3UW+97FlbU3w/ogvLTQGwo3tlyPgDW4CkHJ1m9ym zyCy/hr8A7yQ7HZxrnT3AWHGV2ZZks5HdF2yDCMI6Llb1T+9RsSwwjlx7A4VNMKsfjrc pLNQ== ARC-Authentication-Results: i=1; mx.google.com; dkim=neutral (body hash did not verify) header.i=@intel.com header.s=Intel header.b=lsYEOUXf; 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 c7si13465000ejd.257.2022.02.23.00.58.18; Wed, 23 Feb 2022 00:58:19 -0800 (PST) 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=@intel.com header.s=Intel header.b=lsYEOUXf; 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 9C38168B297; Wed, 23 Feb 2022 10:57:58 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mga18.intel.com (mga18.intel.com [134.134.136.126]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id BCA5468B287 for ; Wed, 23 Feb 2022 10:57:50 +0200 (EET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1645606676; x=1677142676; h=from:to:cc:subject:date:message-id:in-reply-to: references; bh=Zq6E5AuwXTSWX6Mk+7odHID3aC/zjXwTDRloAZm1P08=; b=lsYEOUXfI/V/kMMfjN/5ZgJdyzT54IkYvjnJFgQK6wyy+PKbsBcVvsva eaEXSDZW5Wm43LYc1kMITegFEGo2+Mv7a2I/uGRrjf208kjzvMeW8a8+H FlMjgR9yOH/9JVRanmP6rnAL3IzDOF402Z4oIdHoLNbJe1hA/EWTeSbJd M6B7ZZL4WRmLf90Hj2X9LpCwh5mFyK5VyaPsvHXRCnBhcnFpNQS0rfeFn UjDnPwwoJPP6umoWPAjUn1x0+Dzz3zkb9GuSQpKqF5iAgefoM1QvleHyV V1tpTCv+PRINiI0DjX4pRNqdHBJ8GTtIOdbh0ErZkgyW7hKAJyh3h/1Ao A==; X-IronPort-AV: E=McAfee;i="6200,9189,10266"; a="235427230" X-IronPort-AV: E=Sophos;i="5.88,390,1635231600"; d="scan'208";a="235427230" Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by orsmga106.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 23 Feb 2022 00:57:47 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.88,390,1635231600"; d="scan'208";a="628005794" Received: from otc-skl-e5-server.sh.intel.com ([10.239.43.106]) by FMSMGA003.fm.intel.com with ESMTP; 23 Feb 2022 00:57:46 -0800 From: jianhua.wu-at-intel.com@ffmpeg.org To: ffmpeg-devel@ffmpeg.org Date: Wed, 23 Feb 2022 16:57:32 +0800 Message-Id: <20220223085735.70854-3-jianhua.wu@intel.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20220223085735.70854-1-jianhua.wu@intel.com> References: <20220223085735.70854-1-jianhua.wu@intel.com> Subject: [FFmpeg-devel] [PATCH 3/6] avcodec/x86/hevc_mc: add qpel_h16_8_avx512icl 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: Wu Jianhua MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" X-TUID: AfTsKTe5wxJT From: Wu Jianhua ff_hevc_put_hevc_qpel_h16_8_sse4 3290870 ff_hevc_put_hevc_qpel_h16_8_avx512icl 1730033 Signed-off-by: Wu Jianhua --- libavcodec/x86/hevc_mc.asm | 39 ++++++++++++++++++++++++++++++----- libavcodec/x86/hevcdsp.h | 1 + libavcodec/x86/hevcdsp_init.c | 1 + 3 files changed, 36 insertions(+), 5 deletions(-) diff --git a/libavcodec/x86/hevc_mc.asm b/libavcodec/x86/hevc_mc.asm index 026b6b48ee..8c128f5202 100644 --- a/libavcodec/x86/hevc_mc.asm +++ b/libavcodec/x86/hevc_mc.asm @@ -89,6 +89,7 @@ QPEL_TABLE 10, 8, w, avx2 QPEL_TABLE 8, 1, b, avx512icl_h QPEL_TABLE 8, 1, d, avx512icl_v +QPEL_TABLE 16, 1, b, avx512icl_h pb_qpel_shuffle_index: db 0, 1, 2, 3 db 1, 2, 3, 4 @@ -98,6 +99,14 @@ pb_qpel_shuffle_index: db 0, 1, 2, 3 db 5, 6, 7, 8 db 6, 7, 8, 9 db 7, 8, 9, 10 + db 8, 9, 10, 11 + db 9, 10, 11, 12 + db 10, 11, 12, 13 + db 11, 12, 13, 14 + db 12, 13, 14, 15 + db 13, 14, 15, 16 + db 14, 15, 16, 17 + db 15, 16, 17, 18 db 4, 5, 6, 7 db 5, 6, 7, 8 db 6, 7, 8, 9 @@ -106,6 +115,14 @@ pb_qpel_shuffle_index: db 0, 1, 2, 3 db 9, 10, 11, 12 db 10, 11, 12, 13 db 11, 12, 13, 14 + db 12, 13, 14, 15 + db 13, 14, 15, 16 + db 14, 15, 16, 17 + db 15, 16, 17, 18 + db 16, 17, 18, 19 + db 17, 18, 19, 20 + db 18, 19, 20, 21 + db 19, 20, 21, 22 SECTION .text @@ -1712,7 +1729,7 @@ HEVC_PUT_HEVC_QPEL_HV 16, 10 %macro QPEL_LOAD_SHUF 2 movu m%1, [pb_qpel_shuffle_index + 0] - movu m%2, [pb_qpel_shuffle_index + 32] + movu m%2, [pb_qpel_shuffle_index + 64] %endmacro ; required: m0-m5 @@ -1720,7 +1737,11 @@ HEVC_PUT_HEVC_QPEL_HV 16, 10 ; %2: name for src %macro QPEL_H_LOAD_COMPUTE 2 pxor m%1, m%1 - movu xm4, [%2q - 3] +%if mmsize == 64 + movu ym4, [%2] +%else + movu xm4, [%2] +%endif vpermb m5, m2, m4 vpermb m4, m3, m4 vpdpbusd m%1, m5, m0 @@ -1732,9 +1753,14 @@ cglobal hevc_put_hevc_qpel_h%1_%2, 5, 6, 8, dst, src, srcstride, height, mx, tmp QPEL_FILTER_H %1, mx, 0, 1, tmp QPEL_LOAD_SHUF 2, 3 .loop: - QPEL_H_LOAD_COMPUTE 6, src + QPEL_H_LOAD_COMPUTE 6, srcq - 3 +%if %1 == 8 vpmovdw xm6, m6 movu [dstq], xm6 +%else + vpmovdw ym6, m6 + movu [dstq], ym6 +%endif LOOP_END dst, src, srcstride RET %endmacro @@ -1764,12 +1790,12 @@ cglobal hevc_put_hevc_qpel_hv%1_%2, 6, 7, 8, dst, src, srcstride, height, mx, my %assign %%j %%j+1 %endrep %rep %%extra - QPEL_H_LOAD_COMPUTE %%i, src + QPEL_H_LOAD_COMPUTE %%i, srcq - 3 add srcq, srcstrideq %assign %%i %%i+1 %endrep .loop: - QPEL_H_LOAD_COMPUTE %%i, src + QPEL_H_LOAD_COMPUTE %%i, srcq - 3 vpmulld m22, m14, m6 vpmulld m23, m15, m7 vpmulld m24, m16, m8 @@ -1807,5 +1833,8 @@ INIT_YMM avx512icl HEVC_PUT_HEVC_QPEL_AVX512ICL 8, 8 HEVC_PUT_HEVC_QPEL_HV_AVX512ICL 8, 8 +INIT_ZMM avx512icl +HEVC_PUT_HEVC_QPEL_AVX512ICL 16, 8 + %endif %endif diff --git a/libavcodec/x86/hevcdsp.h b/libavcodec/x86/hevcdsp.h index 5a495d2563..6e3fc01ad0 100644 --- a/libavcodec/x86/hevcdsp.h +++ b/libavcodec/x86/hevcdsp.h @@ -234,6 +234,7 @@ WEIGHTING_PROTOTYPES(10, sse4); WEIGHTING_PROTOTYPES(12, sse4); void ff_hevc_put_hevc_qpel_h8_8_avx512icl(int16_t *dst, uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width); +void ff_hevc_put_hevc_qpel_h16_8_avx512icl(int16_t *dst, uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width); void ff_hevc_put_hevc_qpel_hv8_8_avx512icl(int16_t *dst, uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width); /////////////////////////////////////////////////////////////////////////////// diff --git a/libavcodec/x86/hevcdsp_init.c b/libavcodec/x86/hevcdsp_init.c index 0341835944..4023faa654 100644 --- a/libavcodec/x86/hevcdsp_init.c +++ b/libavcodec/x86/hevcdsp_init.c @@ -880,6 +880,7 @@ void ff_hevc_dsp_init_x86(HEVCDSPContext *c, const int bit_depth) } if (EXTERNAL_AVX512ICL(cpu_flags)) { c->put_hevc_qpel[3][0][1] = ff_hevc_put_hevc_qpel_h8_8_avx512icl; + c->put_hevc_qpel[5][0][1] = ff_hevc_put_hevc_qpel_h16_8_avx512icl; c->put_hevc_qpel[3][1][1] = ff_hevc_put_hevc_qpel_hv8_8_avx512icl; } } else if (bit_depth == 10) {