From patchwork Tue Nov 27 13:26:00 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Lauri Kasanen X-Patchwork-Id: 11184 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id 630E444DD15 for ; Tue, 27 Nov 2018 15:23:45 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id D3AAC68A449; Tue, 27 Nov 2018 15:23:45 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mout.gmx.net (mout.gmx.net [212.227.15.19]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 9AB5268A449 for ; Tue, 27 Nov 2018 15:23:38 +0200 (EET) Received: from Valinor ([84.250.81.169]) by mail.gmx.com (mrgmx001 [212.227.17.184]) with ESMTPSA (Nemesis) id 0LjIit-1fpYQa0DxH-00dXkA for ; Tue, 27 Nov 2018 14:23:43 +0100 Date: Tue, 27 Nov 2018 15:26:00 +0200 From: Lauri Kasanen To: ffmpeg-devel@ffmpeg.org Message-Id: <20181127152600.f425f141da744e2bbd9149d1@gmx.com> X-Mailer: Sylpheed 3.5.0 (GTK+ 2.18.6; x86_64-unknown-linux-gnu) Mime-Version: 1.0 X-Provags-ID: V03:K1:6dH8br3iDPnpHBej1OyFDePmTvVEf0vPN29qHvf1IcISi8RDW3d narMQnT1zK2eiBdgpy/WvKdRlDrd1S8KYvTr3kzdej3zxCqX6KLMhLatvVc5VemkbeC2Ibt pugIk9LA1g0UgXtNrtGmLhDGLGhgHkutlyysHh9m782n+YLx8iAIhvqdAi66OJ/Hv0MyZuP 54WaTAMsrQmXw0qdPA7nQ== X-Spam-Flag: NO X-UI-Out-Filterresults: notjunk:1; V03:K0:I9rZRRGWCuQ=:Z12lJVlMQKeKLzcSuSSO96 4qI/Uwq4SxkwCTZB05GMMwwbUqRGM3l3eF260uOzF89SfxdcuDqpNoT0fV2oeNtEq2dUOS/4x 3NiV/1/MA7DWM/k4LTm2uJai5dtzKBiMzfwmxieVoIHRgCn/wDvwoYb60D+arLS77/OvWyVyt G33cpqof0mAWTNrvFcZCTQWooT77PTbOn2JRUBXJKvfkSfk+yFPiqbsjvtyt5N8PU9CceMu0d 57752bGP3Vi6pBrnIR7X1iYR+hXW30b43VmbRbzxrwaOkl9N3qS9pLumhxJIXdBsnVQyQnDnT v6rhXua5h11RJ4PdZ3stw3/KelpRWkbqL3r+DJigdRMxxi51NlsYe/jRF62SIpnuyBO4lc9LP AqeOk+SbnirQdaDlorh9H63qw/7d+0RsZRiJzkYaoxNeslmP8ZqXH+jXll2Zpa0ebwURd2+3q UDPZzry1jQLJ8zmWVY3RoIekO7yfGYkp6m60gTFVexEoPy7wYJUqrg8DCpSGufbXhjueZecXU q91QMqCek13uIgkp0XJbx94fqgrYcEOvxrXuBOcNqIxRqnfBbfqaljfae+Uae3w3nKQNAzH4L /jVQk6ypdZcelpf4hs70whoCsxEiSywsCECrK2yKOiVZSqfm+Bup69KVLO0+2LMN6KoFmWG46 unE6sqZ3HAeT/7FpCUW1AD/N1BY+T524NoXc9kKwdeI7FDe04Wy02tS3BU0uEBN+D7VFxqbXf 257M9xnxSv3zjbHnwdqcUmukNqfrW5BvA5+4rpkD7fPQv+HXDn+5//GuV66gOSENIYtI3hJI5 tCCoSIlfEL2LCmijtO9oLSmmmNFzPKg+cjftwpW/YktKOduR3iIRrhYGlgP6sHyPRTqPqtgBX uAe8gr7L34qiyhBEtB6NY9r8HMRNWRLtkM4X1i1hivXMfE1ykZBewhFuI65GKK Subject: [FFmpeg-devel] [PATCH] swscale/output: VSX-optimize nbps yuv2plane1 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" ./ffmpeg_g -f rawvideo -pix_fmt rgb24 -s hd1080 -i /dev/zero -pix_fmt yuv420p9le \ -f null -vframes 100 -v error -nostats - Speedups: yuv2plane1_9BE_vsx 11.2042 yuv2plane1_9LE_vsx 11.156 yuv2plane1_10BE_vsx 9.89428 yuv2plane1_10LE_vsx 10.3637 yuv2plane1_12BE_vsx 9.71923 yuv2plane1_12LE_vsx 11.0404 yuv2plane1_14BE_vsx 10.1763 yuv2plane1_14LE_vsx 11.2728 Fate passes, each format tested with an image to video conversion. Depends on "swscale/ppc: Move VSX-using code to its own file". Only tested on LE. Signed-off-by: Lauri Kasanen --- libswscale/ppc/swscale_vsx.c | 83 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 83 insertions(+) diff --git a/libswscale/ppc/swscale_vsx.c b/libswscale/ppc/swscale_vsx.c index 853b587..6462c11 100644 --- a/libswscale/ppc/swscale_vsx.c +++ b/libswscale/ppc/swscale_vsx.c @@ -131,6 +131,75 @@ static void yuv2plane1_8_vsx(const int16_t *src, uint8_t *dest, int dstW, yuv2plane1_8_u(src, dest, dstW, dither, offset, i); } +#if !HAVE_BIGENDIAN + +#define output_pixel(pos, val) \ + if (big_endian) { \ + AV_WB16(pos, av_clip_uintp2(val >> shift, output_bits)); \ + } else { \ + AV_WL16(pos, av_clip_uintp2(val >> shift, output_bits)); \ + } + +static void yuv2plane1_nbps_u(const int16_t *src, uint16_t *dest, int dstW, + int big_endian, int output_bits, int start) +{ + int i; + int shift = 15 - output_bits; + + for (i = start; i < dstW; i++) { + int val = src[i] + (1 << (shift - 1)); + output_pixel(&dest[i], val); + } +} + +static void yuv2plane1_nbps_vsx(const int16_t *src, uint16_t *dest, int dstW, + int big_endian, int output_bits) +{ + const int dst_u = -(uintptr_t)dest & 7; + const int shift = 15 - output_bits; + const int add = (1 << (shift - 1)); + const int clip = (1 << output_bits) - 1; + const vector uint16_t vadd = (vector uint16_t) {add, add, add, add, add, add, add, add}; + const vector uint16_t vswap = (vector uint16_t) vec_splat_u16(big_endian ? 8 : 0); + const vector uint16_t vshift = (vector uint16_t) vec_splat_u16(shift); + const vector uint16_t vlargest = (vector uint16_t) {clip, clip, clip, clip, clip, clip, clip, clip}; + vector uint16_t v; + int i; + + yuv2plane1_nbps_u(src, dest, dst_u, big_endian, output_bits, 0); + + for (i = dst_u; i < dstW - 7; i += 8) { + v = vec_vsx_ld(0, (const uint16_t *) &src[i]); + v = vec_add(v, vadd); + v = vec_sr(v, vshift); + v = vec_min(v, vlargest); + v = vec_rl(v, vswap); + vec_st(v, 0, &dest[i]); + } + + yuv2plane1_nbps_u(src, dest, dstW, big_endian, output_bits, i); +} + +#define yuv2NBPS(bits, BE_LE, is_be, template_size, typeX_t) \ +static void yuv2plane1_ ## bits ## BE_LE ## _vsx(const int16_t *src, \ + uint8_t *dest, int dstW, \ + const uint8_t *dither, int offset) \ +{ \ + yuv2plane1_ ## template_size ## _vsx((const typeX_t *) src, \ + (uint16_t *) dest, dstW, is_be, bits); \ +} + +yuv2NBPS( 9, BE, 1, nbps, int16_t) +yuv2NBPS( 9, LE, 0, nbps, int16_t) +yuv2NBPS(10, BE, 1, nbps, int16_t) +yuv2NBPS(10, LE, 0, nbps, int16_t) +yuv2NBPS(12, BE, 1, nbps, int16_t) +yuv2NBPS(12, LE, 0, nbps, int16_t) +yuv2NBPS(14, BE, 1, nbps, int16_t) +yuv2NBPS(14, LE, 0, nbps, int16_t) + +#endif /* !HAVE_BIGENDIAN */ + #endif /* HAVE_VSX */ av_cold void ff_sws_init_swscale_vsx(SwsContext *c) @@ -158,6 +227,20 @@ av_cold void ff_sws_init_swscale_vsx(SwsContext *c) case 8: c->yuv2plane1 = yuv2plane1_8_vsx; break; +#if !HAVE_BIGENDIAN + case 9: + c->yuv2plane1 = isBE(dstFormat) ? yuv2plane1_9BE_vsx : yuv2plane1_9LE_vsx; + break; + case 10: + c->yuv2plane1 = isBE(dstFormat) ? yuv2plane1_10BE_vsx : yuv2plane1_10LE_vsx; + break; + case 12: + c->yuv2plane1 = isBE(dstFormat) ? yuv2plane1_12BE_vsx : yuv2plane1_12LE_vsx; + break; + case 14: + c->yuv2plane1 = isBE(dstFormat) ? yuv2plane1_14BE_vsx : yuv2plane1_14LE_vsx; + break; +#endif } } #endif /* HAVE_VSX */