From patchwork Thu Dec 1 16:57:47 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: James Darnley X-Patchwork-Id: 1644 Delivered-To: ffmpegpatchwork@gmail.com Received: by 10.103.65.86 with SMTP id o83csp1050070vsa; Thu, 1 Dec 2016 09:12:25 -0800 (PST) X-Received: by 10.28.41.70 with SMTP id p67mr33774425wmp.4.1480612345439; Thu, 01 Dec 2016 09:12:25 -0800 (PST) Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id y130si12835066wmc.29.2016.12.01.09.12.25; Thu, 01 Dec 2016 09:12:25 -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=@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 42FFD689F5D; Thu, 1 Dec 2016 19:12:15 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-wm0-f68.google.com (mail-wm0-f68.google.com [74.125.82.68]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 8583D689F2C for ; Thu, 1 Dec 2016 19:12:08 +0200 (EET) Received: by mail-wm0-f68.google.com with SMTP id m203so35146972wma.3 for ; Thu, 01 Dec 2016 09:12:16 -0800 (PST) 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; bh=gjPCHAHExeuP6JY0icorUOJpwfOwhEIt+hMKlasGCyg=; b=WjoqYUIA036b0+giC3PEP7EXNdczex0EIgzxc2riF1qsT65pEtrr0Mx4K4tz8oEyUp J1sadpGQVE5YzgbLQoXrzvOXs+C08TyrJUnuF8crt2/ojJt+wz3/uZOYLbMP8OivjXr7 Aay/nfU36gUsAdAgWB8tHQf5ocSC3/16MEdeVJLRB8vTG4WojD6WbXmfWwi41SUJbclZ IVVjphwSgjec1aeZaAD/QXFLfh+SK/K4HIvk4o9l/zh0NejOSNxSjs7bH9CerqikjcES fnVfOHhwe6itj17bAEbKWMUZI8YtJIRbIy3Hjy69ph2YG4zOXDq5h5JA0AtO9XKun0dm 2bDw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:sender:from:to:subject:date:message-id :in-reply-to:references; bh=gjPCHAHExeuP6JY0icorUOJpwfOwhEIt+hMKlasGCyg=; b=lu5kcb/mXcGCqFPJC0Vad0uYKg0rRc8d/2Jlb1ar05SiKqHoDQGaYyRYFTEQn4wAL/ iVFmSMgL81Xmjf7aY4oEl15KFEUOzeCm2mA0TSm1OQs2HGM/GFmjxycCo/+utibVuDyY xsr1ff9teJS9gqdf/+CL/KMOpDgr1ETim+IAyv0URY8GnEaBsUJf7tBKKnpcIEPb0Tw8 aLuNidWXwzL8KXAYwwY1SV9OTMLVZJ4anGHF9ygHEtAmZZLspQbpx/DJs/Gk9rq0RaNx CFzTY5X3oCy5lqVjbGMHRr7rGqV1I3yJIttTR4GBurKz5wgt4KPHEZowaPFopGHFIMfd P6Dw== X-Gm-Message-State: AKaTC00mTnfRL134QWX+faQAVKWjDs9nx1fhP06fFCOyeStjjgHVYaX8zm6vLL4NUsf6SA== X-Received: by 10.28.107.21 with SMTP id g21mr35273905wmc.131.1480611592653; Thu, 01 Dec 2016 08:59:52 -0800 (PST) Received: from Ifrit.systemlords.lan (d51a44418.access.telenet.be. [81.164.68.24]) by smtp.gmail.com with ESMTPSA id kp5sm1206172wjb.8.2016.12.01.08.59.52 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Thu, 01 Dec 2016 08:59:52 -0800 (PST) From: James Darnley To: FFmpeg development discussions and patches Date: Thu, 1 Dec 2016 17:57:47 +0100 Message-Id: <20161201165749.10368-5-jdarnley@obe.tv> X-Mailer: git-send-email 2.10.2 In-Reply-To: <20161201165749.10368-1-jdarnley@obe.tv> References: <20161201165749.10368-1-jdarnley@obe.tv> Subject: [FFmpeg-devel] [PATCH 4/6] avcodec/h264: mmx2, sse2, avx 10-bit 4:2:2 h chroma deblock/loop filter 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 MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Yorkfield: - mmx2: 2.54x faster (500 vs. 197 cycles) - sse2: 3.82x faster (500 vs. 131 cycles) Skylake: - mmx2: 1.80x faster (317 vs. 176 cycles) - sse2: 2.81x faster (317 vs. 113 cycles) - avx: 2.85x faster (317 vs. 111 cycles) --- libavcodec/x86/h264_deblock_10bit.asm | 39 +++++++++++++++++++++++++++++++++++ libavcodec/x86/h264dsp_init.c | 6 ++++++ 2 files changed, 45 insertions(+) diff --git a/libavcodec/x86/h264_deblock_10bit.asm b/libavcodec/x86/h264_deblock_10bit.asm index e91b1c6..4c1dc6a 100644 --- a/libavcodec/x86/h264_deblock_10bit.asm +++ b/libavcodec/x86/h264_deblock_10bit.asm @@ -1032,6 +1032,45 @@ cglobal deblock_h_chroma_10, 5, 7, 8, 2*mmsize, pix_, stride_, alpha_, beta_, tc %endif RET +;----------------------------------------------------------------------------- +; void ff_deblock_h_chroma422_10(uint16_t *pix, int stride, int alpha, int beta, +; int8_t *tc0) +;----------------------------------------------------------------------------- +cglobal deblock_h_chroma422_10, 5, 7, 8, 3*mmsize, pix_, stride_, alpha_, beta_, tc0_ + shl alpha_d, 2 + shl beta_d, 2 + + movd m0, [tc0_q] + punpcklbw m0, m0 + psraw m0, 6 + movq [rsp], m0 + + mov r5, pix_q + lea r6, [3*stride_q] + add r5, r6 + + mov r4, -8 + .loop: + + CHROMA_H_LOAD r5, r6, [rsp + 1*mmsize], [rsp + 2*mmsize] + LOAD_AB m4, m5, alpha_d, beta_d + LOAD_MASK m0, m1, m2, m3, m4, m5, m7, m6, m4 + pxor m4, m4 + movd m6, [rsp + r4 + 8] + punpcklwd m6, m6 + punpcklwd m6, m6 + psubw m6, [pw_3] + pmaxsw m6, m4 + pand m7, m6 + DEBLOCK_P0_Q0 m1, m2, m0, m3, m7, m5, m6 + CHROMA_H_STORE r5, r6, [rsp + 1*mmsize], [rsp + 2*mmsize] + + lea pix_q, [pix_q + (mmsize/2)*stride_q] + lea r5, [r5 + (mmsize/2)*stride_q] + add r4, (mmsize/4) + jl .loop +RET + %endmacro %if ARCH_X86_64 == 0 diff --git a/libavcodec/x86/h264dsp_init.c b/libavcodec/x86/h264dsp_init.c index ab270da..7b3d17f 100644 --- a/libavcodec/x86/h264dsp_init.c +++ b/libavcodec/x86/h264dsp_init.c @@ -315,6 +315,8 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_mmxext; if (chroma_format_idc <= 1) { c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_mmxext; + } else { + c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_mmxext; } c->h264_v_loop_filter_luma = ff_deblock_v_luma_10_mmxext; c->h264_h_loop_filter_luma = ff_deblock_h_luma_10_mmxext; @@ -351,6 +353,8 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_sse2; if (chroma_format_idc <= 1) { c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_sse2; + } else { + c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_sse2; } #if HAVE_ALIGNED_STACK c->h264_v_loop_filter_luma = ff_deblock_v_luma_10_sse2; @@ -389,6 +393,8 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_avx; if (chroma_format_idc <= 1) { c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_avx; + } else { + c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_avx; } #if HAVE_ALIGNED_STACK c->h264_v_loop_filter_luma = ff_deblock_v_luma_10_avx;