From patchwork Fri Mar 17 13:18:45 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: James Darnley X-Patchwork-Id: 2989 Delivered-To: ffmpegpatchwork@gmail.com Received: by 10.103.50.79 with SMTP id y76csp261824vsy; Fri, 17 Mar 2017 06:27:40 -0700 (PDT) X-Received: by 10.28.97.194 with SMTP id v185mr2909118wmb.117.1489757260843; Fri, 17 Mar 2017 06:27:40 -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 f126si3229225wmg.167.2017.03.17.06.27.40; Fri, 17 Mar 2017 06:27:40 -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 E74D6688334; Fri, 17 Mar 2017 15:27:18 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-wr0-f193.google.com (mail-wr0-f193.google.com [209.85.128.193]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 28FBE6882EE for ; Fri, 17 Mar 2017 15:27:13 +0200 (EET) Received: by mail-wr0-f193.google.com with SMTP id u108so9599738wrb.2 for ; Fri, 17 Mar 2017 06:27:30 -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=QuibkF/HB9xdUVWOj459x9triPPkjp88mlhjWqyfvrQ=; b=d268yz0A3QDVA5rYCnKEm9RxIUlCmXPY8Sa0h+DW+fBDMXOfOexDGi7Uufif/LlbIs 9mii41kpzwE632Xaewyf5tAfaPPrksPdVYyPH9qaKrJmt6irhE/Dy29Jw97ljjdgNmI/ I9UPnft+/VCIIkgqrLDpr0tgtznjDhMlXLBenopCN5g5oe0vW0JUhapnmstxGMQqQIvu 9WLGlgqJIT7r8FpocG7aEQITfrIflB1kKqbpRigM4BLyCwCJVnpH1tQQng/ZxBUDjkUP jiwnsbPCwF+07WBLJ760NBHgF6DrXKi04DruTHTL/oQr9jpw/3TzzQzb6WgFBdNJfiau 9LUg== 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=QuibkF/HB9xdUVWOj459x9triPPkjp88mlhjWqyfvrQ=; b=hJZ87lJCpHqTtps2vpUObAf4zk4a2z1XvfzU4W9ou/KTam/gXOUba/iaMbFaImF2a6 Gsf3NbEvaUv8U4t54KK7v+bC+ET9ginAQvbaGekE6wTGjsxC5zJF68i35Wn/22W7kXOK Ya7KPnxFdbMwJYWNJgXo1juKVfWHaDt7On5JNrX9TKxG3gUiDGs+u7vEb2A+vf8sbgxA 6nad9Yay+u5syjEWaUhh+TycdoOjNXECnuMuBks9H7qnrymsA7QsCiQJrnBKbt+rxMxL AKk/7tiPaiYbXpcGZtuEiH2KzDVVbkYaVTvKh+NaI7UjwgKIR4wCxMJVAVmhhaWBAs+m cSuA== X-Gm-Message-State: AFeK/H22qI1f89XDEAnL3NvgHJwLEuAarLrE0BaGWJAL29lwRSkU+8cy6pJqrnUm7stt0A== X-Received: by 10.223.166.5 with SMTP id k5mr12355160wrc.134.1489756772027; Fri, 17 Mar 2017 06:19:32 -0700 (PDT) Received: from localhost.localdomain (d51a44418.access.telenet.be. [81.164.68.24]) by smtp.gmail.com with ESMTPSA id e72sm2676411wma.5.2017.03.17.06.19.31 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 17 Mar 2017 06:19:31 -0700 (PDT) From: James Darnley To: FFmpeg development discussions and patches Date: Fri, 17 Mar 2017 14:18:45 +0100 Message-Id: <20170317131845.7760-11-jdarnley@obe.tv> X-Mailer: git-send-email 2.8.3 In-Reply-To: <20170317131845.7760-1-jdarnley@obe.tv> References: <20170317131845.7760-1-jdarnley@obe.tv> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 10/10] h264_idct_add16intra 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" Broken FATE 1.02x faster (1580±4.8 vs. 1555±3.9 decicycles) compared with sse2 --- libavcodec/x86/h264_idct.asm | 43 +++++++++++++++++++++++++++++++++++++++++-- libavcodec/x86/h264dsp_init.c | 2 ++ 2 files changed, 43 insertions(+), 2 deletions(-) diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm index 1515ea5..16998dc 100644 --- a/libavcodec/x86/h264_idct.asm +++ b/libavcodec/x86/h264_idct.asm @@ -895,7 +895,7 @@ REP_RET %else add r0, r0m %endif - call h264_add8x4_idct_sse2 + call h264_add8x4_idct_ %+ cpuname jmp %%skip %%trydc: movsx r0, word [r2 ] @@ -907,13 +907,15 @@ REP_RET %else add r0, r0m %endif - call h264_idct_dc_add8_mmxext + call h264_idct_dc_add8_ %+ cpuname %%skip: %if %1 < 7 add r2, 64 %endif %endmacro +%define h264_idct_dc_add8_sse2 h264_idct_dc_add8_mmxext + ; void ff_h264_idct_add16intra_8_sse2(uint8_t *dst, const int *block_offset, ; int16_t *block, int stride, ; const uint8_t nnzc[6 * 8]) @@ -1193,6 +1195,27 @@ ret packuswb m1, m1 %endmacro +ALIGN 16 +; r0 = uint8_t *dst, r2 = int16_t *block, r3 = int stride, r6=clobbered +; FIXME: I produce incorrect output +h264_idct_dc_add8_avx: + movsxdifnidn r3, r3d + movd m0, [r2 ] ; 0 0 X D + mov word [r2+ 0], 0 + punpcklwd m0, [r2+32] ; x X d D + mov word [r2+32], 0 + paddsw m0, [pw_32] + psraw m0, 6 + punpcklwd m0, m0 ; d d D D + pxor m1, m1 ; 0 0 0 0 + psubw m1, m0 ; -d-d-D-D + packuswb m0, m1 ; -d-d-D-D d d D D + pshuflw m1, m0, q3322 ; -d-d-d-d-D-D-D-D + punpcklwd m0, m0 ; d d d d D D D D + lea r6, [r3*3] + DC_ADD_MMXEXT_OP movq, r0, r3, r6 +ret + cglobal h264_idct_add_8, 3, 3, 8, dst_, block_, stride_ movsxdifnidn stride_q, stride_d IDCT4_ADD dst_q, block_q, stride_q @@ -1238,6 +1261,22 @@ cglobal h264_idct_add16_8, 5, 5 + ARCH_X86_64, 8, dst_, block_offset_, block_, s add16_sse2_cycle 7, 0x26 RET +; FIXME: I produce incorrect output +cglobal h264_idct_add16intra_8, 5, 7 + ARCH_X86_64, 8, dst_, block_offset_, block_, stride_, nnzc_ + movsxdifnidn stride_q, stride_d + %if ARCH_X86_64 + mov r7, r0 + %endif + add16intra_sse2_cycle 0, 0xc + add16intra_sse2_cycle 1, 0x14 + add16intra_sse2_cycle 2, 0xe + add16intra_sse2_cycle 3, 0x16 + add16intra_sse2_cycle 4, 0x1c + add16intra_sse2_cycle 5, 0x24 + add16intra_sse2_cycle 6, 0x1e + add16intra_sse2_cycle 7, 0x26 +RET + ; dst, block_offset, block, stride, nnzc, counter, coeff, dst2, picreg ; 0 1 2 3 4 5 6 7 8 cglobal h264_idct8_add4_8, 5, 8 + npicregs, 10, dst_, block_offset_, block_, stride_, nnzc_, counter_, coeff_, dst2_, picreg diff --git a/libavcodec/x86/h264dsp_init.c b/libavcodec/x86/h264dsp_init.c index 4050276..e09566d 100644 --- a/libavcodec/x86/h264dsp_init.c +++ b/libavcodec/x86/h264dsp_init.c @@ -68,6 +68,7 @@ IDCT_ADD_REP_FUNC(, 16, 10, sse2) IDCT_ADD_REP_FUNC(, 16intra, 8, mmx) IDCT_ADD_REP_FUNC(, 16intra, 8, mmxext) IDCT_ADD_REP_FUNC(, 16intra, 8, sse2) +IDCT_ADD_REP_FUNC(, 16intra, 8, avx) IDCT_ADD_REP_FUNC(, 16intra, 10, sse2) IDCT_ADD_REP_FUNC(, 16, 10, avx) IDCT_ADD_REP_FUNC(, 16intra, 10, avx) @@ -350,6 +351,7 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, c->h264_idct8_dc_add = ff_h264_idct8_dc_add_8_avx; c->h264_idct_add16 = ff_h264_idct_add16_8_avx; c->h264_idct8_add4 = ff_h264_idct8_add4_8_avx; + c->h264_idct_add16intra = ff_h264_idct_add16intra_8_avx; } } else if (bit_depth == 10) { if (EXTERNAL_MMXEXT(cpu_flags)) {