From patchwork Sat Apr 15 01:46:16 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: James Darnley X-Patchwork-Id: 3419 Delivered-To: ffmpegpatchwork@gmail.com Received: by 10.103.3.129 with SMTP id 123csp449585vsd; Fri, 14 Apr 2017 18:53:18 -0700 (PDT) X-Received: by 10.223.170.129 with SMTP id h1mr9015981wrc.140.1492221198770; Fri, 14 Apr 2017 18:53:18 -0700 (PDT) Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id k203si824414wma.165.2017.04.14.18.53.18; Fri, 14 Apr 2017 18:53:18 -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=@ob-encoder-com.20150623.gappssmtp.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 A9AF5689216; Sat, 15 Apr 2017 04:53:08 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-wr0-f182.google.com (mail-wr0-f182.google.com [209.85.128.182]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 6F8A2680A0A for ; Sat, 15 Apr 2017 04:53:06 +0300 (EEST) Received: by mail-wr0-f182.google.com with SMTP id z109so57730308wrb.1 for ; Fri, 14 Apr 2017 18:53:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=ob-encoder-com.20150623.gappssmtp.com; s=20150623; h=sender:from:to:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=X9HyTGjB8zdKzD+1Mmjv4gkH66ER8lMfbwZ4Sd1o2o0=; b=woYngs0y1/iS/MFhKby4YkGKJmVrg9HgbSYF1LtWxrZoBfywUcUzQlm1+4JwJfSWov PX6qxPALfMQfm8tirKJaP8AOx7QeaiUz6NBf/C26877LnaWMiXCIuodnkhoMtYtASAIX wASyp/QB+aNGun9J0Qtx/loZ+eUHVsYmPM1wrzaflpcn4Chg1yGEBFwWGEcmAVZ5lo7P E41g1V3KZ/a1oqJ7lmvLP1/vnN9qoSdb35oBiNdZNJxl5BRSZY72d7JmazaRzJ0zvGR5 ap85McGeyYNYmVxoOfns1BHqaI8W004J3IHeCHIlF5iS8EWJAWXgNcsbfenetnsx57FT 0kJg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:sender:from:to:subject:date:message-id :in-reply-to:references:mime-version:content-transfer-encoding; bh=X9HyTGjB8zdKzD+1Mmjv4gkH66ER8lMfbwZ4Sd1o2o0=; b=cGICEeyOJ8MQmUC4DQnbDb/byq2wLK08nVmkXNMFbTHPe223c/uQbUn+1SoJvRp7Ae bJ1eODEKH7o23pi6tvk0LoBk6dj8ID0DavyRP8h3v7rCBJ8vn3OKKHTf1oyP4vb8I1ZE xPc6qMpROjTyEGPv3Zvpn3870zDBJa6t6uxzcBE+P5OOxaeHUPqc5GTftI1hxmAS9zRb 3phuliX/QLOSBRRY6nJqsBZaNGChdVpABtWS8qoychlF43ZOzqeDfAl6MAddNsoxMUj8 4L4zy/NzFva7f3jvo/dRjQnoIAw+E9+bCrKQ11qIUG0vcaR65tXU98YTPD/8wzaRK/ZF xS0g== X-Gm-Message-State: AN3rC/56xJsI2+gWBn8RxEWOcNnQyBjpBduuey7HBmUkI7O2JROyOmDB DsFi26g+x2+2lcbG X-Received: by 10.223.172.228 with SMTP id o91mr9080022wrc.122.1492220810286; Fri, 14 Apr 2017 18:46:50 -0700 (PDT) Received: from localhost.localdomain (d51a44418.access.telenet.be. [81.164.68.24]) by smtp.gmail.com with ESMTPSA id 3sm4459091wrv.33.2017.04.14.18.46.49 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 14 Apr 2017 18:46:49 -0700 (PDT) From: James Darnley To: FFmpeg development discussions and patches Date: Sat, 15 Apr 2017 03:46:16 +0200 Message-Id: <20170415014618.1592-5-jdarnley@obe.tv> X-Mailer: git-send-email 2.8.3 In-Reply-To: <20170415014618.1592-1-jdarnley@obe.tv> References: <20170415014618.1592-1-jdarnley@obe.tv> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 4/6] avcodec/h264: add avx 8-bit h264_idct_add X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 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" Haswell: - 1.11x faster (522±0.4 vs. 469±1.8 decicycles) compared with mmxext Skylake-U: - 1.21x faster (671±5.5 vs. 555±1.4 decicycles) compared with mmxext --- libavcodec/x86/h264_idct.asm | 33 ++++++++++++++++++++++++++++++++- libavcodec/x86/h264dsp_init.c | 3 +++ 2 files changed, 35 insertions(+), 1 deletion(-) diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm index bc4dce4..24fb4d2 100644 --- a/libavcodec/x86/h264_idct.asm +++ b/libavcodec/x86/h264_idct.asm @@ -65,7 +65,15 @@ SECTION .text IDCT4_1D w, 0, 1, 2, 3, 4, 5 mova m6, [pw_32] - TRANSPOSE4x4W 0, 1, 2, 3, 4 + %if mmsize == 8 + TRANSPOSE4x4W 0, 1, 2, 3, 4 + %else + punpcklwd m0, m1 + punpcklwd m2, m3 + SBUTTERFLY dq, 0, 2, 4 + MOVHL m1, m0 + MOVHL m3, m2 + %endif paddw m0, m6 IDCT4_1D w, 0, 1, 2, 3, 4, 5 pxor m7, m7 @@ -1131,3 +1139,26 @@ INIT_MMX mmx IDCT_DC_DEQUANT 0 INIT_MMX sse2 IDCT_DC_DEQUANT 7 + +INIT_XMM avx + +; %unmacro STORE_DIFFx2 8 ; remove macro from x86util.asm but yasm doesn't have this yet +%macro STORE_DIFFx2 8 ; add1, add2, reg1, reg2, zero, shift, source, stride + movd %3, [%7] + movd %4, [%7+%8] + psraw %1, %6 + psraw %2, %6 + punpcklbw %3, %5 + punpcklbw %4, %5 + paddw %3, %1 + paddw %4, %2 + packuswb %3, %5 + packuswb %4, %5 + movd [%7], %3 + movd [%7+%8], %4 +%endmacro + +cglobal h264_idct_add_8, 3, 3, 8, dst_, block_, stride_ + movsxdifnidn stride_q, stride_d + IDCT4_ADD dst_q, block_q, stride_q +RET diff --git a/libavcodec/x86/h264dsp_init.c b/libavcodec/x86/h264dsp_init.c index 0643b37..8ba085f 100644 --- a/libavcodec/x86/h264dsp_init.c +++ b/libavcodec/x86/h264dsp_init.c @@ -32,6 +32,7 @@ void ff_h264_idct ## NUM ## _add_ ## DEPTH ## _ ## OPT(uint8_t *dst, \ int stride); IDCT_ADD_FUNC(, 8, mmx) +IDCT_ADD_FUNC(, 8, avx) IDCT_ADD_FUNC(, 10, sse2) IDCT_ADD_FUNC(_dc, 8, mmxext) IDCT_ADD_FUNC(_dc, 10, mmxext) @@ -337,6 +338,8 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_8_avx; c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_avx; } + + c->h264_idct_add = ff_h264_idct_add_8_avx; } } else if (bit_depth == 10) { if (EXTERNAL_MMXEXT(cpu_flags)) {