From patchwork Thu May 25 15:50:52 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: =?utf-8?b?Q2zDqW1lbnQgQsWTc2No?= X-Patchwork-Id: 3738 Delivered-To: ffmpegpatchwork@gmail.com Received: by 10.103.10.2 with SMTP id 2csp860634vsk; Thu, 25 May 2017 08:51:50 -0700 (PDT) X-Received: by 10.28.111.7 with SMTP id k7mr11090473wmc.50.1495727510706; Thu, 25 May 2017 08:51:50 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1495727510; cv=none; d=google.com; s=arc-20160816; b=rVPXzVi2im5DAZUyqYrsNSYSye9d4Do2hMA3YrOlEE1CAst5KbQnJzJXsyg+DQ8jgf PNUgsNN9KBDJi7ym1AXLTJ+u6rAkDzvVWJsRLqrDqx5erB1MXkasmj3C4A9NgMlO2vwE qo6M4Xp/EXdB8ckg/Bde5DHndcjhBkr0yipZTb/hCqcBqiYCYN6w4wnFbjffaGSo4wW+ KkS3j/rLu+DbL+96GVqjG5uCH3HUH1Zqii23+U0Uju0V5LExStZW+Gx95YHtp/HZgT/Y sScpDI1n7XwfbYObV+BppEo1wQ7dyFRNOy9ByX2PQs9+NF4wGG9+mWCc2KJKZdkKVd+8 TWBA== 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:message-id:date:to:from:delivered-to :arc-authentication-results; bh=pvh7v432lmGsag+eBhxPYnsEPS1d/t8TLUl632mZqDQ=; b=yWb1BXQKFBQDcEltppBBTmqVtA3MVr6Oc0kUbzuc4Oz+xi2UQoRE2nzKMRy/HeeRgY YxgdWMdIpNa4uo4+YlesbXrPaLA5NchngAiqULYU+SSI/EKGqPuyqSNwfuGAGxWHIaLD hS6JoSlWQc9Qf6uXpwLka0E6Vddz5Rbd117lWkm2IaIVXqc/c4jU0VISpvn70qY1T2SQ inI+AdkTLxe9hm9C4FEfnO0XTnJUOUqzZV1n4ua5nLVsVN2Gh/EQ7W2WdInYNuyitZbT 2T6E7I8EniJRyQRM7vbQ7o0+zlgmnDSP7+1GrFenxGQy3cF+nc0EsZtoT33VyjpCnbe9 F9sA== ARC-Authentication-Results: i=1; mx.google.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 Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org. [79.124.17.100]) by mx.google.com with ESMTP id g191si8895713wmd.66.2017.05.25.08.51.49; Thu, 25 May 2017 08:51:50 -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; 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 E99AF689C03; Thu, 25 May 2017 18:51:43 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from golem.pkh.me (LStLambert-657-1-117-164.w92-154.abo.wanadoo.fr [92.154.28.164]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 285CD689B94 for ; Thu, 25 May 2017 18:51:38 +0300 (EEST) Received: from localhost (golem.pkh.me [local]) by golem.pkh.me (OpenSMTPD) with ESMTPA id 12f41014; Thu, 25 May 2017 15:51:40 +0000 (UTC) From: =?UTF-8?q?Cl=C3=A9ment=20B=C5=93sch?= To: ffmpeg-devel@ffmpeg.org Date: Thu, 25 May 2017 17:50:52 +0200 Message-Id: <20170525155052.3585-1-u@pkh.me> X-Mailer: git-send-email 2.13.0 Subject: [FFmpeg-devel] [PATCH] lavc/aarch64: add a few SIMD function for AAC PS 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 Cc: =?UTF-8?q?Cl=C3=A9ment=20B=C5=93sch?= MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" --- This is still not benchmarked (written and verified with qemu). I typically wrote an alternative implementation for stereo_interpolate[0] which needs to be compared with the current one: function ff_ps_stereo_interpolate_neon, export=1 ld1 {v0.4S}, [x2] ld1 {v1.4S}, [x3] 1: ld1 {v2.2S}, [x0] ld1 {v3.2S}, [x1] fadd v0.4S, v0.4S, v1.4S fmul v4.2S, v2.2S, v0.S[0] fmul v5.2S, v2.2S, v0.S[1] fmla v4.2S, v3.2S, v0.S[2] fmla v5.2S, v3.2S, v0.S[3] st1 {v4.2S}, [x0], #8 st1 {v5.2S}, [x1], #8 subs w4, w4, #1 b.gt 1b ret endfunc I don't know which is faster. For now, the current version follows the logic I used in stereo_interpolate[1] (the ipdopd one). It's doing less mult operations, but more shuffling. A 3rd alternative would be possible if it was possible to assume len % 2 was always true (allowing overreading and overwriting by one more entry basically). Currently, this is not the case. Speaking of ipdopd, the factors table and the ext may be clumsy. --- libavcodec/aacpsdsp.h | 1 + libavcodec/aacpsdsp_template.c | 2 + libavcodec/aarch64/Makefile | 2 + libavcodec/aarch64/aacpsdsp_init_aarch64.c | 44 +++++++++++++ libavcodec/aarch64/aacpsdsp_neon.S | 101 +++++++++++++++++++++++++++++ 5 files changed, 150 insertions(+) create mode 100644 libavcodec/aarch64/aacpsdsp_init_aarch64.c create mode 100644 libavcodec/aarch64/aacpsdsp_neon.S diff --git a/libavcodec/aacpsdsp.h b/libavcodec/aacpsdsp.h index ad9bbb81bd..3714a08052 100644 --- a/libavcodec/aacpsdsp.h +++ b/libavcodec/aacpsdsp.h @@ -51,6 +51,7 @@ typedef struct PSDSPContext { void AAC_RENAME(ff_psdsp_init)(PSDSPContext *s); void ff_psdsp_init_arm(PSDSPContext *s); +void ff_psdsp_init_aarch64(PSDSPContext *s); void ff_psdsp_init_mips(PSDSPContext *s); void ff_psdsp_init_x86(PSDSPContext *s); diff --git a/libavcodec/aacpsdsp_template.c b/libavcodec/aacpsdsp_template.c index 3049ce8b79..158160b246 100644 --- a/libavcodec/aacpsdsp_template.c +++ b/libavcodec/aacpsdsp_template.c @@ -222,6 +222,8 @@ av_cold void AAC_RENAME(ff_psdsp_init)(PSDSPContext *s) #if !USE_FIXED if (ARCH_ARM) ff_psdsp_init_arm(s); + if (ARCH_AARCH64) + ff_psdsp_init_aarch64(s); if (ARCH_MIPS) ff_psdsp_init_mips(s); if (ARCH_X86) diff --git a/libavcodec/aarch64/Makefile b/libavcodec/aarch64/Makefile index 104bc67802..d440b1b18a 100644 --- a/libavcodec/aarch64/Makefile +++ b/libavcodec/aarch64/Makefile @@ -11,6 +11,7 @@ OBJS-$(CONFIG_NEON_CLOBBER_TEST) += aarch64/neontest.o OBJS-$(CONFIG_VIDEODSP) += aarch64/videodsp_init.o # decoders/encoders +OBJS-$(CONFIG_AAC_DECODER) += aarch64/aacpsdsp_init_aarch64.o OBJS-$(CONFIG_DCA_DECODER) += aarch64/synth_filter_init.o OBJS-$(CONFIG_RV40_DECODER) += aarch64/rv40dsp_init_aarch64.o OBJS-$(CONFIG_VC1DSP) += aarch64/vc1dsp_init_aarch64.o @@ -42,6 +43,7 @@ NEON-OBJS-$(CONFIG_MDCT) += aarch64/mdct_neon.o NEON-OBJS-$(CONFIG_MPEGAUDIODSP) += aarch64/mpegaudiodsp_neon.o # decoders/encoders +NEON-OBJS-$(CONFIG_AAC_DECODER) += aarch64/aacpsdsp_neon.o NEON-OBJS-$(CONFIG_DCA_DECODER) += aarch64/synth_filter_neon.o NEON-OBJS-$(CONFIG_VORBIS_DECODER) += aarch64/vorbisdsp_neon.o NEON-OBJS-$(CONFIG_VP9_DECODER) += aarch64/vp9itxfm_16bpp_neon.o \ diff --git a/libavcodec/aarch64/aacpsdsp_init_aarch64.c b/libavcodec/aarch64/aacpsdsp_init_aarch64.c new file mode 100644 index 0000000000..9c9e5db851 --- /dev/null +++ b/libavcodec/aarch64/aacpsdsp_init_aarch64.c @@ -0,0 +1,44 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" + +#include "libavutil/aarch64/cpu.h" +#include "libavcodec/aacpsdsp.h" + +void ff_ps_add_squares_neon(float *dst, const float (*src)[2], int n); +void ff_ps_mul_pair_single_neon(float (*dst)[2], float (*src0)[2], + float *src1, int n); +void ff_ps_stereo_interpolate_neon(float (*l)[2], float (*r)[2], + float h[2][4], float h_step[2][4], + int len); +void ff_ps_stereo_interpolate_ipdopd_neon(float (*l)[2], float (*r)[2], + float h[2][4], float h_step[2][4], + int len); + +av_cold void ff_psdsp_init_aarch64(PSDSPContext *s) +{ + int cpu_flags = av_get_cpu_flags(); + + if (have_neon(cpu_flags)) { + s->add_squares = ff_ps_add_squares_neon; + s->mul_pair_single = ff_ps_mul_pair_single_neon; + s->stereo_interpolate[0] = ff_ps_stereo_interpolate_neon; + s->stereo_interpolate[1] = ff_ps_stereo_interpolate_ipdopd_neon; + } +} diff --git a/libavcodec/aarch64/aacpsdsp_neon.S b/libavcodec/aarch64/aacpsdsp_neon.S new file mode 100644 index 0000000000..82d641b730 --- /dev/null +++ b/libavcodec/aarch64/aacpsdsp_neon.S @@ -0,0 +1,101 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "libavutil/aarch64/asm.S" + +function ff_ps_add_squares_neon, export=1 +1: + ld1 {v0.4S,v1.4S}, [x1], #32 + fmul v0.4S, v0.4S, v0.4S + fmul v1.4S, v1.4S, v1.4S + faddp v2.4S, v0.4S, v1.4S + ld1 {v3.4S}, [x0] + fadd v3.4S, v3.4S, v2.4S + st1 {v3.4S}, [x0], #16 + subs w2, w2, #4 + b.gt 1b + ret +endfunc + +function ff_ps_mul_pair_single_neon, export=1 +1: + ld1 {v0.4S,v1.4S}, [x1], #32 + ld1 {v2.4S}, [x2], #16 + zip1 v3.4S, v2.4S, v2.4S + zip2 v4.4S, v2.4S, v2.4S + fmul v0.4S, v0.4S, v3.4S + fmul v1.4S, v1.4S, v4.4S + st1 {v0.4S,v1.4S}, [x0], #32 + subs w3, w3, #4 + b.gt 1b + ret +endfunc + +function ff_ps_stereo_interpolate_neon, export=1 + ld1 {v0.4S}, [x2] + ld1 {v1.4S}, [x3] +1: + ld1 {v2.2S}, [x0] + ld1 {v3.2S}, [x1] + dup v2.2D, v2.D[0] + dup v3.2D, v3.D[0] + fadd v0.4S, v0.4S, v1.4S + zip1 v4.4S, v0.4S, v0.4S + zip2 v5.4S, v0.4S, v0.4S + fmul v2.4S, v2.4S, v4.4S + fmla v2.4S, v3.4S, v5.4S + st1 {v2.D}[0], [x0], #8 + st1 {v2.D}[1], [x1], #8 + subs w4, w4, #1 + b.gt 1b + ret +endfunc + +const ipdopd_factors, align=4 + .float -1.0, 1.0, -1.0, 1.0 +endconst + +function ff_ps_stereo_interpolate_ipdopd_neon, export=1 + movrel x5, ipdopd_factors + ld1 {v20.4S}, [x5] + ld1 {v0.4S,v1.4S}, [x2] + ld1 {v6.4S,v7.4S}, [x3] +1: + ld1 {v2.2S}, [x0] + ld1 {v3.2S}, [x1] + dup v2.2D, v2.D[0] + dup v3.2D, v3.D[0] + fadd v0.4S, v0.4S, v6.4S + fadd v1.4S, v1.4S, v7.4S + zip1 v16.4S, v0.4S, v0.4S + zip2 v17.4S, v0.4S, v0.4S + zip1 v18.4S, v1.4S, v1.4S + zip2 v19.4S, v1.4S, v1.4S + fmul v4.4S, v2.4S, v16.4S + fmla v4.4S, v3.4S, v17.4S + ext v2.16B, v2.16B, v2.16B, #4 + ext v3.16B, v3.16B, v3.16B, #4 + fmul v5.4S, v2.4S, v18.4S + fmla v5.4S, v3.4S, v19.4S + fmla v4.4S, v5.4S, v20.4S + st1 {v4.D}[0], [x0], #8 + st1 {v4.D}[1], [x1], #8 + subs w4, w4, #1 + b.gt 1b + ret +endfunc