@scrypted/libav 1.0.179 → 1.0.181
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/_openh264install/include/wels/codec_api.h +592 -0
- package/_openh264install/include/wels/codec_app_def.h +820 -0
- package/_openh264install/include/wels/codec_def.h +216 -0
- package/_openh264install/include/wels/codec_ver.h +15 -0
- package/_openh264install/lib/libopenh264.a +0 -0
- package/_openh264install/lib/pkgconfig/openh264.pc +10 -0
- package/binding.gyp +4 -0
- package/openh264/.reviewboardrc +3 -0
- package/openh264/.travis.yml +42 -0
- package/openh264/CODING_STYLE +8 -0
- package/openh264/CONTRIBUTORS +55 -0
- package/openh264/LICENSE +23 -0
- package/openh264/Makefile +408 -0
- package/openh264/README.md +197 -0
- package/openh264/RELEASES +467 -0
- package/openh264/SECURITY.md +16 -0
- package/openh264/autotest/codec_comparision/run_MultiCodecComparision.sh +431 -0
- package/openh264/autotest/performanceTest/android/run_AutoTest_android.sh +180 -0
- package/openh264/autotest/performanceTest/ios/fruitstrap +0 -0
- package/openh264/autotest/performanceTest/ios/iFileTransfer +0 -0
- package/openh264/autotest/performanceTest/ios/run_AutoTest_ios.sh +193 -0
- package/openh264/autotest/performanceTest/ios/uiascript.js +19 -0
- package/openh264/autotest/performanceTest/parsePerfData.sh +107 -0
- package/openh264/autotest/performanceTest/run_perfTest.sh +122 -0
- package/openh264/autotest/unitTest/android/run_AutoTest_android.sh +112 -0
- package/openh264/autotest/unitTest/ios/run_AutoTest_ios.sh +167 -0
- package/openh264/autotest/unitTest/run_ParseUTxml.sh +70 -0
- package/openh264/autotest/unitTest/run_unitTest.sh +98 -0
- package/openh264/build.gradle +30 -0
- package/openh264/code-coverage.sh +6 -0
- package/openh264/codec/api/meson.build +1 -0
- package/openh264/codec/api/wels/codec_api.h +592 -0
- package/openh264/codec/api/wels/codec_app_def.h +820 -0
- package/openh264/codec/api/wels/codec_def.h +216 -0
- package/openh264/codec/api/wels/codec_ver.h +15 -0
- package/openh264/codec/api/wels/meson.build +9 -0
- package/openh264/codec/common/arm/arm_arch_common_macro.S +83 -0
- package/openh264/codec/common/arm/copy_mb_neon.S +161 -0
- package/openh264/codec/common/arm/deblocking_neon.S +857 -0
- package/openh264/codec/common/arm/expand_picture_neon.S +154 -0
- package/openh264/codec/common/arm/intra_pred_common_neon.S +82 -0
- package/openh264/codec/common/arm/mc_neon.S +2283 -0
- package/openh264/codec/common/arm64/arm_aarch64_common.h +85 -0
- package/openh264/codec/common/arm64/arm_arch64_common_macro.S +79 -0
- package/openh264/codec/common/arm64/copy_mb_aarch64_neon.S +202 -0
- package/openh264/codec/common/arm64/deblocking_aarch64_neon.S +852 -0
- package/openh264/codec/common/arm64/expand_picture_aarch64_neon.S +150 -0
- package/openh264/codec/common/arm64/intra_pred_common_aarch64_neon.S +58 -0
- package/openh264/codec/common/arm64/mc_aarch64_neon.S +2614 -0
- package/openh264/codec/common/generate_version.sh +28 -0
- package/openh264/codec/common/inc/WelsList.h +287 -0
- package/openh264/codec/common/inc/WelsLock.h +97 -0
- package/openh264/codec/common/inc/WelsTask.h +75 -0
- package/openh264/codec/common/inc/WelsTaskThread.h +83 -0
- package/openh264/codec/common/inc/WelsThread.h +106 -0
- package/openh264/codec/common/inc/WelsThreadLib.h +151 -0
- package/openh264/codec/common/inc/WelsThreadPool.h +124 -0
- package/openh264/codec/common/inc/asmdefs_mmi.h +345 -0
- package/openh264/codec/common/inc/copy_mb.h +103 -0
- package/openh264/codec/common/inc/cpu.h +81 -0
- package/openh264/codec/common/inc/cpu_core.h +103 -0
- package/openh264/codec/common/inc/crt_util_safe_x.h +104 -0
- package/openh264/codec/common/inc/deblocking_common.h +124 -0
- package/openh264/codec/common/inc/expand_pic.h +107 -0
- package/openh264/codec/common/inc/golomb_common.h +166 -0
- package/openh264/codec/common/inc/intra_pred_common.h +86 -0
- package/openh264/codec/common/inc/loongson_intrinsics.h +1919 -0
- package/openh264/codec/common/inc/ls_defines.h +130 -0
- package/openh264/codec/common/inc/macros.h +329 -0
- package/openh264/codec/common/inc/mc.h +407 -0
- package/openh264/codec/common/inc/measure_time.h +88 -0
- package/openh264/codec/common/inc/memory_align.h +116 -0
- package/openh264/codec/common/inc/msa_macros.h +2393 -0
- package/openh264/codec/common/inc/sad_common.h +139 -0
- package/openh264/codec/common/inc/typedefs.h +86 -0
- package/openh264/codec/common/inc/utils.h +95 -0
- package/openh264/codec/common/inc/version.h +10 -0
- package/openh264/codec/common/inc/version_gen.h.template +6 -0
- package/openh264/codec/common/inc/welsCodecTrace.h +64 -0
- package/openh264/codec/common/inc/wels_common_defs.h +373 -0
- package/openh264/codec/common/inc/wels_const_common.h +63 -0
- package/openh264/codec/common/loongarch/copy_mb_lsx.c +194 -0
- package/openh264/codec/common/loongarch/deblock_lsx.c +1489 -0
- package/openh264/codec/common/loongarch/intra_pred_com_lasx.c +112 -0
- package/openh264/codec/common/loongarch/intra_pred_com_lsx.c +91 -0
- package/openh264/codec/common/loongarch/mc_chroma_lsx.c +192 -0
- package/openh264/codec/common/loongarch/mc_horver_lsx.c +654 -0
- package/openh264/codec/common/loongarch/satd_sad_lasx.c +568 -0
- package/openh264/codec/common/meson.build +84 -0
- package/openh264/codec/common/mips/copy_mb_mmi.c +477 -0
- package/openh264/codec/common/mips/copy_mb_msa.c +80 -0
- package/openh264/codec/common/mips/deblock_mmi.c +2826 -0
- package/openh264/codec/common/mips/deblock_msa.c +1024 -0
- package/openh264/codec/common/mips/expand_picture_mmi.c +673 -0
- package/openh264/codec/common/mips/intra_pred_com_mmi.c +548 -0
- package/openh264/codec/common/mips/satd_sad_mmi.c +2154 -0
- package/openh264/codec/common/targets.mk +136 -0
- package/openh264/codec/common/x86/asm_inc.asm +743 -0
- package/openh264/codec/common/x86/cpuid.asm +263 -0
- package/openh264/codec/common/x86/dct.asm +1036 -0
- package/openh264/codec/common/x86/deblock.asm +848 -0
- package/openh264/codec/common/x86/expand_picture.asm +728 -0
- package/openh264/codec/common/x86/intra_pred_com.asm +117 -0
- package/openh264/codec/common/x86/mb_copy.asm +615 -0
- package/openh264/codec/common/x86/mc_chroma.asm +313 -0
- package/openh264/codec/common/x86/mc_luma.asm +4490 -0
- package/openh264/codec/common/x86/satd_sad.asm +2734 -0
- package/openh264/codec/common/x86/vaa.asm +411 -0
- package/openh264/codec/console/common/inc/read_config.h +68 -0
- package/openh264/codec/console/common/meson.build +12 -0
- package/openh264/codec/console/common/targets.mk +20 -0
- package/openh264/codec/console/dec/inc/d3d9_utils.h +148 -0
- package/openh264/codec/console/dec/meson.build +15 -0
- package/openh264/codec/console/dec/targets.mk +20 -0
- package/openh264/codec/console/enc/meson.build +9 -0
- package/openh264/codec/console/enc/targets.mk +19 -0
- package/openh264/codec/console/meson.build +3 -0
- package/openh264/codec/decoder/core/arm/block_add_neon.S +129 -0
- package/openh264/codec/decoder/core/arm/intra_pred_neon.S +635 -0
- package/openh264/codec/decoder/core/arm64/block_add_aarch64_neon.S +131 -0
- package/openh264/codec/decoder/core/arm64/intra_pred_aarch64_neon.S +524 -0
- package/openh264/codec/decoder/core/inc/au_parser.h +199 -0
- package/openh264/codec/decoder/core/inc/bit_stream.h +62 -0
- package/openh264/codec/decoder/core/inc/cabac_decoder.h +111 -0
- package/openh264/codec/decoder/core/inc/deblocking.h +118 -0
- package/openh264/codec/decoder/core/inc/dec_frame.h +156 -0
- package/openh264/codec/decoder/core/inc/dec_golomb.h +344 -0
- package/openh264/codec/decoder/core/inc/decode_mb_aux.h +83 -0
- package/openh264/codec/decoder/core/inc/decode_slice.h +104 -0
- package/openh264/codec/decoder/core/inc/decoder.h +186 -0
- package/openh264/codec/decoder/core/inc/decoder_context.h +590 -0
- package/openh264/codec/decoder/core/inc/decoder_core.h +206 -0
- package/openh264/codec/decoder/core/inc/error_code.h +222 -0
- package/openh264/codec/decoder/core/inc/error_concealment.h +68 -0
- package/openh264/codec/decoder/core/inc/fmo.h +114 -0
- package/openh264/codec/decoder/core/inc/get_intra_predictor.h +191 -0
- package/openh264/codec/decoder/core/inc/manage_dec_ref.h +62 -0
- package/openh264/codec/decoder/core/inc/mb_cache.h +80 -0
- package/openh264/codec/decoder/core/inc/memmgr_nal_unit.h +66 -0
- package/openh264/codec/decoder/core/inc/mv_pred.h +195 -0
- package/openh264/codec/decoder/core/inc/nal_prefix.h +56 -0
- package/openh264/codec/decoder/core/inc/nalu.h +79 -0
- package/openh264/codec/decoder/core/inc/parameter_sets.h +217 -0
- package/openh264/codec/decoder/core/inc/parse_mb_syn_cabac.h +89 -0
- package/openh264/codec/decoder/core/inc/parse_mb_syn_cavlc.h +137 -0
- package/openh264/codec/decoder/core/inc/pic_queue.h +62 -0
- package/openh264/codec/decoder/core/inc/picture.h +112 -0
- package/openh264/codec/decoder/core/inc/rec_mb.h +104 -0
- package/openh264/codec/decoder/core/inc/slice.h +208 -0
- package/openh264/codec/decoder/core/inc/vlc_decoder.h +177 -0
- package/openh264/codec/decoder/core/inc/wels_common_basis.h +317 -0
- package/openh264/codec/decoder/core/inc/wels_const.h +65 -0
- package/openh264/codec/decoder/core/inc/wels_decoder_thread.h +170 -0
- package/openh264/codec/decoder/core/loongarch/mb_aux_lsx.c +296 -0
- package/openh264/codec/decoder/core/mips/dct_mmi.c +786 -0
- package/openh264/codec/decoder/core/x86/dct.asm +72 -0
- package/openh264/codec/decoder/core/x86/intra_pred.asm +1456 -0
- package/openh264/codec/decoder/meson.build +69 -0
- package/openh264/codec/decoder/plus/inc/welsDecoderExt.h +170 -0
- package/openh264/codec/decoder/targets.mk +115 -0
- package/openh264/codec/encoder/core/arm/intra_pred_neon.S +589 -0
- package/openh264/codec/encoder/core/arm/intra_pred_sad_3_opt_neon.S +735 -0
- package/openh264/codec/encoder/core/arm/memory_neon.S +62 -0
- package/openh264/codec/encoder/core/arm/pixel_neon.S +882 -0
- package/openh264/codec/encoder/core/arm/reconstruct_neon.S +903 -0
- package/openh264/codec/encoder/core/arm/svc_motion_estimation.S +366 -0
- package/openh264/codec/encoder/core/arm64/intra_pred_aarch64_neon.S +505 -0
- package/openh264/codec/encoder/core/arm64/intra_pred_sad_3_opt_aarch64_neon.S +611 -0
- package/openh264/codec/encoder/core/arm64/memory_aarch64_neon.S +63 -0
- package/openh264/codec/encoder/core/arm64/pixel_aarch64_neon.S +621 -0
- package/openh264/codec/encoder/core/arm64/reconstruct_aarch64_neon.S +717 -0
- package/openh264/codec/encoder/core/arm64/svc_motion_estimation_aarch64_neon.S +350 -0
- package/openh264/codec/encoder/core/inc/as264_common.h +138 -0
- package/openh264/codec/encoder/core/inc/au_set.h +152 -0
- package/openh264/codec/encoder/core/inc/deblocking.h +92 -0
- package/openh264/codec/encoder/core/inc/decode_mb_aux.h +112 -0
- package/openh264/codec/encoder/core/inc/dq_map.h +58 -0
- package/openh264/codec/encoder/core/inc/encode_mb_aux.h +199 -0
- package/openh264/codec/encoder/core/inc/encoder.h +148 -0
- package/openh264/codec/encoder/core/inc/encoder_context.h +240 -0
- package/openh264/codec/encoder/core/inc/extern.h +124 -0
- package/openh264/codec/encoder/core/inc/get_intra_predictor.h +186 -0
- package/openh264/codec/encoder/core/inc/mb_cache.h +141 -0
- package/openh264/codec/encoder/core/inc/md.h +181 -0
- package/openh264/codec/encoder/core/inc/mt_defs.h +91 -0
- package/openh264/codec/encoder/core/inc/mv_pred.h +191 -0
- package/openh264/codec/encoder/core/inc/nal_encap.h +143 -0
- package/openh264/codec/encoder/core/inc/param_svc.h +576 -0
- package/openh264/codec/encoder/core/inc/parameter_sets.h +181 -0
- package/openh264/codec/encoder/core/inc/paraset_strategy.h +310 -0
- package/openh264/codec/encoder/core/inc/picture.h +137 -0
- package/openh264/codec/encoder/core/inc/picture_handle.h +66 -0
- package/openh264/codec/encoder/core/inc/rc.h +284 -0
- package/openh264/codec/encoder/core/inc/ref_list_mgr_svc.h +148 -0
- package/openh264/codec/encoder/core/inc/sample.h +154 -0
- package/openh264/codec/encoder/core/inc/set_mb_syn_cabac.h +127 -0
- package/openh264/codec/encoder/core/inc/set_mb_syn_cavlc.h +92 -0
- package/openh264/codec/encoder/core/inc/slice.h +212 -0
- package/openh264/codec/encoder/core/inc/slice_multi_threading.h +97 -0
- package/openh264/codec/encoder/core/inc/stat.h +102 -0
- package/openh264/codec/encoder/core/inc/svc_base_layer_md.h +102 -0
- package/openh264/codec/encoder/core/inc/svc_enc_frame.h +136 -0
- package/openh264/codec/encoder/core/inc/svc_enc_golomb.h +123 -0
- package/openh264/codec/encoder/core/inc/svc_enc_macroblock.h +82 -0
- package/openh264/codec/encoder/core/inc/svc_enc_slice_segment.h +212 -0
- package/openh264/codec/encoder/core/inc/svc_encode_mb.h +63 -0
- package/openh264/codec/encoder/core/inc/svc_encode_slice.h +184 -0
- package/openh264/codec/encoder/core/inc/svc_mode_decision.h +94 -0
- package/openh264/codec/encoder/core/inc/svc_motion_estimate.h +369 -0
- package/openh264/codec/encoder/core/inc/svc_set_mb_syn.h +70 -0
- package/openh264/codec/encoder/core/inc/svc_set_mb_syn_cavlc.h +67 -0
- package/openh264/codec/encoder/core/inc/vlc_encoder.h +92 -0
- package/openh264/codec/encoder/core/inc/wels_common_basis.h +160 -0
- package/openh264/codec/encoder/core/inc/wels_const.h +174 -0
- package/openh264/codec/encoder/core/inc/wels_func_ptr_def.h +300 -0
- package/openh264/codec/encoder/core/inc/wels_preprocess.h +251 -0
- package/openh264/codec/encoder/core/inc/wels_task_base.h +79 -0
- package/openh264/codec/encoder/core/inc/wels_task_encoder.h +142 -0
- package/openh264/codec/encoder/core/inc/wels_task_management.h +135 -0
- package/openh264/codec/encoder/core/inc/wels_transpose_matrix.h +58 -0
- package/openh264/codec/encoder/core/loongarch/dct_lasx.c +452 -0
- package/openh264/codec/encoder/core/loongarch/get_intra_predictor_lsx.c +166 -0
- package/openh264/codec/encoder/core/loongarch/quant_lsx.c +153 -0
- package/openh264/codec/encoder/core/loongarch/sample_lasx.c +87 -0
- package/openh264/codec/encoder/core/loongarch/svc_motion_estimate_lsx.c +104 -0
- package/openh264/codec/encoder/core/mips/dct_mmi.c +529 -0
- package/openh264/codec/encoder/core/mips/quant_mmi.c +553 -0
- package/openh264/codec/encoder/core/mips/score_mmi.c +324 -0
- package/openh264/codec/encoder/core/x86/coeff.asm +695 -0
- package/openh264/codec/encoder/core/x86/dct.asm +108 -0
- package/openh264/codec/encoder/core/x86/intra_pred.asm +1129 -0
- package/openh264/codec/encoder/core/x86/matrix_transpose.asm +395 -0
- package/openh264/codec/encoder/core/x86/memzero.asm +132 -0
- package/openh264/codec/encoder/core/x86/quant.asm +507 -0
- package/openh264/codec/encoder/core/x86/sample_sc.asm +1829 -0
- package/openh264/codec/encoder/core/x86/score.asm +363 -0
- package/openh264/codec/encoder/meson.build +98 -0
- package/openh264/codec/encoder/plus/inc/welsEncoderExt.h +129 -0
- package/openh264/codec/encoder/targets.mk +146 -0
- package/openh264/codec/meson.build +9 -0
- package/openh264/codec/processing/interface/IWelsVP.h +310 -0
- package/openh264/codec/processing/meson.build +71 -0
- package/openh264/codec/processing/targets.mk +117 -0
- package/openh264/docs/doxygen/Doxyfile +1865 -0
- package/openh264/docs/doxygen/Home.rest +25 -0
- package/openh264/docs/doxygen/ISVCDecoder.rest +19 -0
- package/openh264/docs/doxygen/ISVCEncoder.rest +16 -0
- package/openh264/docs/doxygen/UsageExampleForDecoder.rest +10 -0
- package/openh264/docs/doxygen/UsageExampleForEncoder.rest +15 -0
- package/openh264/docs/doxygen2rst.py +619 -0
- package/openh264/gmpopenh264.info +4 -0
- package/openh264/gradlew +172 -0
- package/openh264/meson.build +218 -0
- package/openh264/meson_options.txt +1 -0
- package/openh264/module/RefCounted.h +141 -0
- package/openh264/module/gmp-openh264.cpp +1196 -0
- package/openh264/module/targets.mk +14 -0
- package/openh264/module/task_utils.h +38 -0
- package/openh264/module/task_utils.py +207 -0
- package/openh264/module/task_utils_generated.h +2087 -0
- package/openh264/openh264.def +7 -0
- package/openh264/openh264.pc.in +10 -0
- package/openh264/openh264.rc +61 -0
- package/openh264/openh264.rc.template +61 -0
- package/openh264/res/Adobe_PDF_sample_a_1024x768_50Frms.264 +0 -0
- package/openh264/res/BA1_FT_C.264 +0 -0
- package/openh264/res/BA1_Sony_D.jsv +0 -0
- package/openh264/res/BAMQ1_JVC_C.264 +0 -0
- package/openh264/res/BAMQ2_JVC_C.264 +0 -0
- package/openh264/res/BANM_MW_D.264 +0 -0
- package/openh264/res/BASQP1_Sony_C.jsv +0 -0
- package/openh264/res/BA_MW_D.264 +0 -0
- package/openh264/res/BA_MW_D_IDR_LOST.264 +0 -0
- package/openh264/res/BA_MW_D_P_LOST.264 +0 -0
- package/openh264/res/CI1_FT_B.264 +0 -0
- package/openh264/res/CI_MW_D.264 +0 -0
- package/openh264/res/CVFC1_Sony_C.jsv +0 -0
- package/openh264/res/CVPCMNL1_SVA_C.264 +0 -0
- package/openh264/res/CiscoVT2people_160x96_6fps.yuv +1 -0
- package/openh264/res/CiscoVT2people_320x192_12fps.yuv +1 -0
- package/openh264/res/Cisco_Absolute_Power_1280x720_30fps.yuv +3607 -0
- package/openh264/res/Cisco_Adobe_PDF_sample_a_1024x768_CAVLC_Bframe_9.264 +0 -0
- package/openh264/res/Cisco_Men_whisper_640x320_CABAC_Bframe_9.264 +0 -0
- package/openh264/res/Cisco_Men_whisper_640x320_CAVLC_Bframe_9.264 +0 -0
- package/openh264/res/Error_I_P.264 +0 -0
- package/openh264/res/LS_SVA_D.264 +0 -0
- package/openh264/res/MIDR_MW_D.264 +0 -0
- package/openh264/res/MPS_MW_A.264 +0 -0
- package/openh264/res/MR1_BT_A.h264 +0 -0
- package/openh264/res/MR1_MW_A.264 +0 -0
- package/openh264/res/MR2_MW_A.264 +0 -0
- package/openh264/res/MR2_TANDBERG_E.264 +0 -0
- package/openh264/res/NL1_Sony_D.jsv +0 -0
- package/openh264/res/NLMQ1_JVC_C.264 +0 -0
- package/openh264/res/NLMQ2_JVC_C.264 +0 -0
- package/openh264/res/NRF_MW_E.264 +0 -0
- package/openh264/res/QCIF_2P_I_allIPCM.264 +0 -0
- package/openh264/res/SVA_BA1_B.264 +0 -0
- package/openh264/res/SVA_BA2_D.264 +0 -0
- package/openh264/res/SVA_Base_B.264 +0 -0
- package/openh264/res/SVA_CL1_E.264 +0 -0
- package/openh264/res/SVA_FM1_E.264 +0 -0
- package/openh264/res/SVA_NL1_B.264 +0 -0
- package/openh264/res/SVA_NL2_E.264 +0 -0
- package/openh264/res/SarVui.264 +0 -0
- package/openh264/res/Static.264 +0 -0
- package/openh264/res/Static_152_100.yuv +1 -0
- package/openh264/res/VID_1280x544_cabac_temporal_direct.264 +0 -0
- package/openh264/res/VID_1280x544_cavlc_temporal_direct.264 +0 -0
- package/openh264/res/VID_1280x720_cabac_temporal_direct.264 +0 -0
- package/openh264/res/VID_1280x720_cavlc_temporal_direct.264 +0 -0
- package/openh264/res/VID_1920x1080_cabac_temporal_direct.264 +0 -0
- package/openh264/res/VID_1920x1080_cavlc_temporal_direct.264 +0 -0
- package/openh264/res/Zhling_1280x720.264 +0 -0
- package/openh264/res/jm_1080p_allslice.264 +0 -0
- package/openh264/res/sps_subsetsps_bothVUI.264 +0 -0
- package/openh264/res/test_cif_I_CABAC_PCM.264 +0 -0
- package/openh264/res/test_cif_I_CABAC_slice.264 +0 -0
- package/openh264/res/test_cif_P_CABAC_slice.264 +0 -0
- package/openh264/res/test_qcif_cabac.264 +0 -0
- package/openh264/res/test_scalinglist_jm.264 +0 -0
- package/openh264/res/test_vd_1d.264 +0 -0
- package/openh264/res/test_vd_rc.264 +0 -0
- package/openh264/run_Test.sh +91 -0
- package/openh264/settings.gradle +4 -0
- package/openh264/subprojects/gtest.wrap +10 -0
- package/openh264/testbin/AutoBuild_Windows_VS2008.bat +59 -0
- package/openh264/testbin/CmdLineExample.sh +7 -0
- package/openh264/testbin/CmdLineReadMe +22 -0
- package/openh264/testbin/layer2.cfg +37 -0
- package/openh264/testbin/layer2_arbitrary_res.cfg +37 -0
- package/openh264/testbin/layer2_vd.cfg +38 -0
- package/openh264/testbin/layer2_vd_rc.cfg +38 -0
- package/openh264/testbin/welsenc.cfg +71 -0
- package/openh264/testbin/welsenc_arbitrary_res.cfg +70 -0
- package/openh264/testbin/welsenc_ios.cfg +71 -0
- package/openh264/testbin/welsenc_vd_1d.cfg +70 -0
- package/openh264/testbin/welsenc_vd_rc.cfg +69 -0
- package/openh264/ut.def +2 -0
- package/package.json +1 -1
|
@@ -0,0 +1,2826 @@
|
|
|
1
|
+
/*!
|
|
2
|
+
* \copy
|
|
3
|
+
* Copyright (c) 2009-2018, Cisco Systems
|
|
4
|
+
* All rights reserved.
|
|
5
|
+
*
|
|
6
|
+
* Redistribution and use in source and binary forms, with or without
|
|
7
|
+
* modification, are permitted provided that the following conditions
|
|
8
|
+
* are met:
|
|
9
|
+
*
|
|
10
|
+
* * Redistributions of source code must retain the above copyright
|
|
11
|
+
* notice, this list of conditions and the following disclaimer.
|
|
12
|
+
*
|
|
13
|
+
* * Redistributions in binary form must reproduce the above copyright
|
|
14
|
+
* notice, this list of conditions and the following disclaimer in
|
|
15
|
+
* the documentation and/or other materials provided with the
|
|
16
|
+
* distribution.
|
|
17
|
+
*
|
|
18
|
+
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
19
|
+
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
|
20
|
+
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
|
|
21
|
+
* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
|
|
22
|
+
* COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
|
|
23
|
+
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
|
|
24
|
+
* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
|
|
25
|
+
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
|
|
26
|
+
* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
27
|
+
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
|
|
28
|
+
* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
|
29
|
+
* POSSIBILITY OF SUCH DAMAGE.
|
|
30
|
+
*
|
|
31
|
+
*
|
|
32
|
+
* \file deblock_mmi.c
|
|
33
|
+
*
|
|
34
|
+
* \brief Loongson optimize
|
|
35
|
+
*
|
|
36
|
+
* \date 20/07/2018 Created
|
|
37
|
+
*
|
|
38
|
+
*************************************************************************************
|
|
39
|
+
*/
|
|
40
|
+
#include <stdint.h>
|
|
41
|
+
#include "asmdefs_mmi.h"
|
|
42
|
+
|
|
43
|
+
void DeblockLumaLt4V_mmi(uint8_t *pPix, int32_t iStride, int32_t iAlpha,
|
|
44
|
+
int32_t iBeta, int8_t *pTC) {
|
|
45
|
+
unsigned char tmp[512] __attribute__((aligned(32)));
|
|
46
|
+
BACKUP_REG;
|
|
47
|
+
__asm__ volatile (
|
|
48
|
+
".set arch=loongson3a \n\t"
|
|
49
|
+
"dsll $8, %[iStride], 0x1 \n\t"
|
|
50
|
+
"daddu $8, $8, %[iStride] \n\t"
|
|
51
|
+
"dsubu $14, %[pPix], $8 \n\t"
|
|
52
|
+
|
|
53
|
+
"dsll $8, %[iStride], 0x1 \n\t"
|
|
54
|
+
"dsubu $9, %[pPix], $8 \n\t"
|
|
55
|
+
|
|
56
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
57
|
+
"dsubu $13, %[pPix], %[iStride] \n\t"
|
|
58
|
+
"daddu %[iStride], %[iStride], %[pPix] \n\t"
|
|
59
|
+
"daddu $12, $8, %[pPix] \n\t"
|
|
60
|
+
|
|
61
|
+
"punpcklhw $f0, $f0, $f0 \n\t"
|
|
62
|
+
"lb $8, 0x0(%[pTC]) \n\t"
|
|
63
|
+
"punpcklwd $f0, $f0, $f0 \n\t"
|
|
64
|
+
"mov.d $f2, $f0 \n\t"
|
|
65
|
+
"gssqc1 $f2, $f0, 432-112(%[tmp]) \n\t"
|
|
66
|
+
"dmtc1 %[iBeta], $f0 \n\t"
|
|
67
|
+
"lb %[iAlpha], 0x1(%[pTC]) \n\t"
|
|
68
|
+
"dli %[iBeta], 0xFFFF \n\t"
|
|
69
|
+
"punpcklhw $f0, $f0, $f0 \n\t"
|
|
70
|
+
"and $10, %[iAlpha], %[iBeta] \n\t"
|
|
71
|
+
"punpcklwd $f0, $f0, $f0 \n\t"
|
|
72
|
+
"mov.d $f2, $f0 \n\t"
|
|
73
|
+
"and %[iAlpha], %[iAlpha], %[iBeta] \n\t"
|
|
74
|
+
"dmtc1 $10, $f4 \n\t"
|
|
75
|
+
"mov.d $f8, $f4 \n\t"
|
|
76
|
+
"dmtc1 %[iAlpha], $f16 \n\t"
|
|
77
|
+
"and %[iAlpha], $8, %[iBeta] \n\t"
|
|
78
|
+
"dmtc1 %[iAlpha], $f20 \n\t"
|
|
79
|
+
"mov.d $f24, $f20 \n\t"
|
|
80
|
+
"mov.d $f28, $f20 \n\t"
|
|
81
|
+
"gssqc1 $f2, $f0, 432-336(%[tmp]) \n\t"
|
|
82
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
83
|
+
|
|
84
|
+
"lb %[iAlpha], 0x3(%[pTC]) \n\t"
|
|
85
|
+
"lb %[pTC], 0x2(%[pTC]) \n\t"
|
|
86
|
+
"dmtc1 $10, $f12 \n\t"
|
|
87
|
+
"punpcklhw $f0, $f0, $f16 \n\t"
|
|
88
|
+
"and $8, %[iAlpha], %[iBeta] \n\t"
|
|
89
|
+
"punpcklhw $f24, $f24, $f8 \n\t"
|
|
90
|
+
"punpcklhw $f20, $f20, $f4 \n\t"
|
|
91
|
+
"punpcklhw $f0, $f0, $f24 \n\t"
|
|
92
|
+
"punpcklhw $f28, $f28, $f12 \n\t"
|
|
93
|
+
"punpcklhw $f28, $f28, $f20 \n\t"
|
|
94
|
+
"punpckhhw $f2, $f0, $f28 \n\t"
|
|
95
|
+
"punpcklhw $f0, $f0, $f28 \n\t"
|
|
96
|
+
"gssqc1 $f2, $f0, 432-400(%[tmp]) \n\t"
|
|
97
|
+
"dmtc1 $8, $f0 \n\t"
|
|
98
|
+
"and %[iAlpha], %[iAlpha], %[iBeta] \n\t"
|
|
99
|
+
"mov.d $f8, $f0 \n\t"
|
|
100
|
+
"dmtc1 %[iAlpha], $f16 \n\t"
|
|
101
|
+
"and %[iAlpha], %[pTC], %[iBeta] \n\t"
|
|
102
|
+
"dmtc1 $8, $f12 \n\t"
|
|
103
|
+
"dmtc1 %[iAlpha], $f20 \n\t"
|
|
104
|
+
"punpcklhw $f20, $f20, $f0 \n\t"
|
|
105
|
+
|
|
106
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
107
|
+
"dmtc1 %[iAlpha], $f24 \n\t"
|
|
108
|
+
"and %[pTC], %[pTC], %[iBeta] \n\t"
|
|
109
|
+
"punpcklhw $f24, $f24, $f8 \n\t"
|
|
110
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
111
|
+
"dmtc1 %[pTC], $f4 \n\t"
|
|
112
|
+
|
|
113
|
+
"gslqc1 $f10, $f8, 0x0($9) \n\t"
|
|
114
|
+
"punpckhbh $f10, $f8, $f0 \n\t"
|
|
115
|
+
"punpcklbh $f8, $f8, $f0 \n\t"
|
|
116
|
+
|
|
117
|
+
"dli %[iAlpha], 0x4 \n\t"
|
|
118
|
+
"seh %[pTC], %[iAlpha] \n\t"
|
|
119
|
+
"punpcklhw $f28, $f28, $f12 \n\t"
|
|
120
|
+
"punpcklhw $f28, $f28, $f20 \n\t"
|
|
121
|
+
"gslqc1 $f22, $f20, 0x0(%[iStride]) \n\t"
|
|
122
|
+
"gslqc1 $f14, $f12, 0x0($13) \n\t"
|
|
123
|
+
"gsldxc1 $f2, 0x0($12, $0) \n\t"
|
|
124
|
+
"punpckhbh $f22, $f20, $f0 \n\t"
|
|
125
|
+
"punpcklbh $f20, $f20, $f0 \n\t"
|
|
126
|
+
"gssqc1 $f22, $f20, 432-240(%[tmp]) \n\t"
|
|
127
|
+
"punpckhbh $f22, $f2, $f0 \n\t"
|
|
128
|
+
"punpcklbh $f20, $f2, $f0 \n\t"
|
|
129
|
+
"gssqc1 $f22, $f20, 432-352(%[tmp]) \n\t"
|
|
130
|
+
"punpcklhw $f4, $f4, $f16 \n\t"
|
|
131
|
+
"gslqc1 $f18, $f16, 0x0($14) \n\t"
|
|
132
|
+
"punpcklhw $f4, $f4, $f24 \n\t"
|
|
133
|
+
"gslqc1 $f26, $f24, 0x0(%[pPix]) \n\t"
|
|
134
|
+
"punpckhhw $f6, $f4, $f28 \n\t"
|
|
135
|
+
"punpcklhw $f4, $f4, $f28 \n\t"
|
|
136
|
+
"punpckhbh $f26, $f24, $f0 \n\t"
|
|
137
|
+
"punpcklbh $f24, $f24, $f0 \n\t"
|
|
138
|
+
"punpckhbh $f14, $f12, $f0 \n\t"
|
|
139
|
+
"punpcklbh $f12, $f12, $f0 \n\t"
|
|
140
|
+
"punpckhbh $f18, $f16, $f0 \n\t"
|
|
141
|
+
"punpcklbh $f16, $f16, $f0 \n\t"
|
|
142
|
+
"psubh $f28, $f12, $f16 \n\t"
|
|
143
|
+
"psubh $f30, $f14, $f18 \n\t"
|
|
144
|
+
"gssqc1 $f18, $f16, 432-272(%[tmp]) \n\t"
|
|
145
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f16, $f18)
|
|
146
|
+
"gslqc1 $f18, $f16, 432-336(%[tmp]) \n\t"
|
|
147
|
+
"gslqc1 $f2, $f0, 432-352(%[tmp]) \n\t"
|
|
148
|
+
"pcmpgth $f20, $f16, $f28 \n\t"
|
|
149
|
+
"pcmpgth $f22, $f18, $f30 \n\t"
|
|
150
|
+
"gssqc1 $f22, $f20, 432-288(%[tmp]) \n\t"
|
|
151
|
+
"psubh $f28, $f24, $f0 \n\t"
|
|
152
|
+
"psubh $f30, $f26, $f2 \n\t"
|
|
153
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f20, $f22)
|
|
154
|
+
"pcmpgth $f20, $f16, $f28 \n\t"
|
|
155
|
+
"pcmpgth $f22, $f18, $f30 \n\t"
|
|
156
|
+
"gssqc1 $f22, $f20, 432-256(%[tmp]) \n\t"
|
|
157
|
+
"pavgh $f20, $f12, $f24 \n\t"
|
|
158
|
+
"pavgh $f22, $f14, $f26 \n\t"
|
|
159
|
+
"gssqc1 $f22, $f20, 432-304(%[tmp]) \n\t"
|
|
160
|
+
"gslqc1 $f22, $f20, 432-400(%[tmp]) \n\t"
|
|
161
|
+
"gslqc1 $f30, $f28, 432-288(%[tmp]) \n\t"
|
|
162
|
+
"gslqc1 $f2, $f0, 432-256(%[tmp]) \n\t"
|
|
163
|
+
"psubh $f20, $f20, $f28 \n\t"
|
|
164
|
+
"psubh $f22, $f22, $f30 \n\t"
|
|
165
|
+
"psubh $f20, $f20, $f0 \n\t"
|
|
166
|
+
"psubh $f22, $f22, $f2 \n\t"
|
|
167
|
+
"gssqc1 $f22, $f20, 432-224(%[tmp]) \n\t"
|
|
168
|
+
"gslqc1 $f2, $f0, 432-240(%[tmp]) \n\t"
|
|
169
|
+
"psubh $f20, $f24, $f12 \n\t"
|
|
170
|
+
"psubh $f22, $f26, $f14 \n\t"
|
|
171
|
+
"gssqc1 $f26, $f24, 432-32(%[tmp]) \n\t"
|
|
172
|
+
"psubh $f24, $f24, $f0 \n\t"
|
|
173
|
+
"psubh $f26, $f26, $f2 \n\t"
|
|
174
|
+
"gssqc1 $f22, $f20, 432-384(%[tmp]) \n\t"
|
|
175
|
+
WELS_AbsH($f28, $f30, $f20, $f22, $f28, $f30)
|
|
176
|
+
"gslqc1 $f22, $f20, 432-112(%[tmp]) \n\t"
|
|
177
|
+
"pcmpgth $f20, $f20, $f28 \n\t"
|
|
178
|
+
"pcmpgth $f22, $f22, $f30 \n\t"
|
|
179
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
180
|
+
"pcmpgth $f28, $f16, $f24 \n\t"
|
|
181
|
+
"pcmpgth $f30, $f18, $f26 \n\t"
|
|
182
|
+
|
|
183
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
184
|
+
"and $f20, $f20, $f28 \n\t"
|
|
185
|
+
"and $f22, $f22, $f30 \n\t"
|
|
186
|
+
"psubh $f24, $f12, $f8 \n\t"
|
|
187
|
+
"psubh $f26, $f14, $f10 \n\t"
|
|
188
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
189
|
+
"pcmpgth $f28, $f16, $f24 \n\t"
|
|
190
|
+
"pcmpgth $f30, $f18, $f26 \n\t"
|
|
191
|
+
"gslqc1 $f26, $f24, 432-400(%[tmp]) \n\t"
|
|
192
|
+
"and $f20, $f20, $f28 \n\t"
|
|
193
|
+
"and $f22, $f22, $f30 \n\t"
|
|
194
|
+
"pcmpgth $f28, $f24, $f0 \n\t"
|
|
195
|
+
"pcmpgth $f30, $f26, $f0 \n\t"
|
|
196
|
+
"pcmpeqh $f24, $f24, $f0 \n\t"
|
|
197
|
+
"pcmpeqh $f26, $f26, $f0 \n\t"
|
|
198
|
+
"or $f28, $f28, $f24 \n\t"
|
|
199
|
+
"or $f30, $f30, $f26 \n\t"
|
|
200
|
+
"and $f20, $f20, $f28 \n\t"
|
|
201
|
+
"and $f22, $f22, $f30 \n\t"
|
|
202
|
+
"gssqc1 $f22, $f20, 432-320(%[tmp]) \n\t"
|
|
203
|
+
"dmtc1 %[pTC], $f20 \n\t"
|
|
204
|
+
"punpckhhw $f26, $f20, $f20 \n\t"
|
|
205
|
+
"punpcklhw $f24, $f20, $f20 \n\t"
|
|
206
|
+
"punpcklwd $f20, $f24, $f24 \n\t"
|
|
207
|
+
"mov.d $f22, $f20 \n\t"
|
|
208
|
+
"gssqc1 $f22, $f20, 432-336(%[tmp]) \n\t"
|
|
209
|
+
"gslqc1 $f22, $f20, 432-224(%[tmp]) \n\t"
|
|
210
|
+
"psubh $f24, $f0, $f20 \n\t"
|
|
211
|
+
"dli $11, 0x2 \n\t"
|
|
212
|
+
"psubh $f26, $f0, $f22 \n\t"
|
|
213
|
+
"dmtc1 $11, $f28 \n\t"
|
|
214
|
+
"gslqc1 $f22, $f20, 432-384(%[tmp]) \n\t"
|
|
215
|
+
"gslqc1 $f2, $f0, 432-240(%[tmp]) \n\t"
|
|
216
|
+
"psllh $f20, $f20, $f28 \n\t"
|
|
217
|
+
"psllh $f22, $f22, $f28 \n\t"
|
|
218
|
+
"psubh $f28, $f8, $f0 \n\t"
|
|
219
|
+
"psubh $f30, $f10, $f2 \n\t"
|
|
220
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
221
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
222
|
+
"gslqc1 $f22, $f20, 432-336(%[tmp]) \n\t"
|
|
223
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
224
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
225
|
+
"dli $11, 0x3 \n\t"
|
|
226
|
+
"dmtc1 $11, $f20 \n\t"
|
|
227
|
+
"psrah $f28, $f28, $f20 \n\t"
|
|
228
|
+
"psrah $f30, $f30, $f20 \n\t"
|
|
229
|
+
"gslqc1 $f22, $f20, 432-224(%[tmp]) \n\t"
|
|
230
|
+
"pmaxsh $f24, $f24, $f28 \n\t"
|
|
231
|
+
"pmaxsh $f26, $f26, $f30 \n\t"
|
|
232
|
+
"gslqc1 $f2, $f0, 432-320(%[tmp]) \n\t"
|
|
233
|
+
"pminsh $f20, $f20, $f24 \n\t"
|
|
234
|
+
"pminsh $f22, $f22, $f26 \n\t"
|
|
235
|
+
|
|
236
|
+
"and $f20, $f20, $f0 \n\t"
|
|
237
|
+
"and $f22, $f22, $f2 \n\t"
|
|
238
|
+
"gslqc1 $f26, $f24, 432-400(%[tmp]) \n\t"
|
|
239
|
+
"gssqc1 $f22, $f20, 432-64(%[tmp]) \n\t"
|
|
240
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
241
|
+
"gssqc1 $f26, $f24, 432-384(%[tmp]) \n\t"
|
|
242
|
+
"psubh $f20, $f0, $f24 \n\t"
|
|
243
|
+
"psubh $f22, $f0, $f26 \n\t"
|
|
244
|
+
"gssqc1 $f22, $f20, 432-368(%[tmp]) \n\t"
|
|
245
|
+
"mov.d $f24, $f20 \n\t"
|
|
246
|
+
"mov.d $f26, $f22 \n\t"
|
|
247
|
+
"gslqc1 $f22, $f20, 432-272(%[tmp]) \n\t"
|
|
248
|
+
"gslqc1 $f30, $f28, 432-304(%[tmp]) \n\t"
|
|
249
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
250
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
251
|
+
"paddh $f28, $f8, $f8 \n\t"
|
|
252
|
+
"paddh $f30, $f10, $f10 \n\t"
|
|
253
|
+
"psubh $f20, $f20, $f28 \n\t"
|
|
254
|
+
"psubh $f22, $f22, $f30 \n\t"
|
|
255
|
+
"dli $11, 0x1 \n\t"
|
|
256
|
+
"dmtc1 $11, $f28 \n\t"
|
|
257
|
+
"psrah $f20, $f20, $f28 \n\t"
|
|
258
|
+
"psrah $f22, $f22, $f28 \n\t"
|
|
259
|
+
"pmaxsh $f24, $f24, $f20 \n\t"
|
|
260
|
+
"pmaxsh $f26, $f26, $f22 \n\t"
|
|
261
|
+
"gslqc1 $f22, $f20, 432-384(%[tmp]) \n\t"
|
|
262
|
+
"pminsh $f20, $f20, $f24 \n\t"
|
|
263
|
+
"pminsh $f22, $f22, $f26 \n\t"
|
|
264
|
+
|
|
265
|
+
"gslqc1 $f26, $f24, 432-320(%[tmp]) \n\t"
|
|
266
|
+
"gslqc1 $f30, $f28, 432-288(%[tmp]) \n\t"
|
|
267
|
+
"and $f20, $f20, $f24 \n\t"
|
|
268
|
+
"and $f22, $f22, $f26 \n\t"
|
|
269
|
+
"and $f20, $f20, $f28 \n\t"
|
|
270
|
+
"and $f22, $f22, $f30 \n\t"
|
|
271
|
+
"gslqc1 $f26, $f24, 432-240(%[tmp]) \n\t"
|
|
272
|
+
"gssqc1 $f22, $f20, 432-96(%[tmp]) \n\t"
|
|
273
|
+
"gslqc1 $f22, $f20, 432-352(%[tmp]) \n\t"
|
|
274
|
+
"gslqc1 $f30, $f28, 432-304(%[tmp]) \n\t"
|
|
275
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
276
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
277
|
+
"paddh $f28, $f24, $f24 \n\t"
|
|
278
|
+
"paddh $f30, $f26, $f26 \n\t"
|
|
279
|
+
"gslqc1 $f26, $f24, 432-368(%[tmp]) \n\t"
|
|
280
|
+
"dli $11, 0x1 \n\t"
|
|
281
|
+
"psubh $f20, $f20, $f28 \n\t"
|
|
282
|
+
"dmtc1 $11, $f28 \n\t"
|
|
283
|
+
"psubh $f22, $f22, $f30 \n\t"
|
|
284
|
+
|
|
285
|
+
"psrah $f20, $f20, $f28 \n\t"
|
|
286
|
+
"psrah $f22, $f22, $f28 \n\t"
|
|
287
|
+
"gslqc1 $f30, $f28, 0x0(%[iStride]) \n\t"
|
|
288
|
+
"pmaxsh $f24, $f24, $f20 \n\t"
|
|
289
|
+
"pmaxsh $f26, $f26, $f22 \n\t"
|
|
290
|
+
"gslqc1 $f22, $f20, 432-400(%[tmp]) \n\t"
|
|
291
|
+
"pminsh $f20, $f20, $f24 \n\t"
|
|
292
|
+
"pminsh $f22, $f22, $f26 \n\t"
|
|
293
|
+
"gslqc1 $f26, $f24, 432-320(%[tmp]) \n\t"
|
|
294
|
+
"and $f20, $f20, $f24 \n\t"
|
|
295
|
+
"and $f22, $f22, $f26 \n\t"
|
|
296
|
+
"gslqc1 $f26, $f24, 432-256(%[tmp]) \n\t"
|
|
297
|
+
"and $f20, $f20, $f24 \n\t"
|
|
298
|
+
"and $f22, $f22, $f26 \n\t"
|
|
299
|
+
"gslqc1 $f26, $f24, 0x0($9) \n\t"
|
|
300
|
+
"punpcklbh $f28, $f30, $f0 \n\t"
|
|
301
|
+
"punpckhbh $f30, $f30, $f0 \n\t"
|
|
302
|
+
"gssqc1 $f30, $f28, 432-352(%[tmp]) \n\t"
|
|
303
|
+
|
|
304
|
+
"gslqc1 $f30, $f28, 0x0($12) \n\t"
|
|
305
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
306
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
307
|
+
"gssqc1 $f22, $f20, 432-48(%[tmp]) \n\t"
|
|
308
|
+
"gslqc1 $f22, $f20, 0x0($14) \n\t"
|
|
309
|
+
"gssqc1 $f26, $f24, 432-368(%[tmp]) \n\t"
|
|
310
|
+
"gslqc1 $f26, $f24, 0x0($13) \n\t"
|
|
311
|
+
"punpcklbh $f28, $f30, $f0 \n\t"
|
|
312
|
+
"punpckhbh $f30, $f30, $f0 \n\t"
|
|
313
|
+
"punpcklbh $f20, $f22, $f0 \n\t"
|
|
314
|
+
"punpckhbh $f22, $f22, $f0 \n\t"
|
|
315
|
+
"gssqc1 $f30, $f28, 432-384(%[tmp]) \n\t"
|
|
316
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
317
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
318
|
+
"gssqc1 $f26, $f24, 432-400(%[tmp]) \n\t"
|
|
319
|
+
|
|
320
|
+
"gslqc1 $f30, $f28, 432-400(%[tmp]) \n\t"
|
|
321
|
+
"gslqc1 $f26, $f24, 0x0(%[pPix]) \n\t"
|
|
322
|
+
"psubh $f28, $f28, $f20 \n\t"
|
|
323
|
+
"psubh $f30, $f30, $f22 \n\t"
|
|
324
|
+
"gssqc1 $f22, $f20, 432-16(%[tmp]) \n\t"
|
|
325
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f20, $f22)
|
|
326
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
327
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
328
|
+
"pcmpgth $f20, $f16, $f28 \n\t"
|
|
329
|
+
"pcmpgth $f22, $f18, $f30 \n\t"
|
|
330
|
+
"gslqc1 $f30, $f28, 432-384(%[tmp]) \n\t"
|
|
331
|
+
"gssqc1 $f22, $f20, 432-288(%[tmp]) \n\t"
|
|
332
|
+
|
|
333
|
+
"psubh $f28, $f24, $f28 \n\t"
|
|
334
|
+
"psubh $f30, $f26, $f30 \n\t"
|
|
335
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f20, $f22)
|
|
336
|
+
"pcmpgth $f20, $f16, $f28 \n\t"
|
|
337
|
+
"pcmpgth $f22, $f18, $f30 \n\t"
|
|
338
|
+
"gssqc1 $f22, $f20, 432-256(%[tmp]) \n\t"
|
|
339
|
+
|
|
340
|
+
"gslqc1 $f22, $f20, 432-400(%[tmp]) \n\t"
|
|
341
|
+
"gssqc1 $f26, $f24, 432-80(%[tmp]) \n\t"
|
|
342
|
+
"pavgh $f20, $f20, $f24 \n\t"
|
|
343
|
+
"pavgh $f22, $f22, $f26 \n\t"
|
|
344
|
+
"gssqc1 $f22, $f20, 432-304(%[tmp]) \n\t"
|
|
345
|
+
|
|
346
|
+
"gslqc1 $f22, $f20, 432-288(%[tmp]) \n\t"
|
|
347
|
+
"gslqc1 $f30, $f28, 432-256(%[tmp]) \n\t"
|
|
348
|
+
"psubh $f20, $f4, $f20 \n\t"
|
|
349
|
+
"psubh $f22, $f6, $f22 \n\t"
|
|
350
|
+
"psubh $f20, $f20, $f28 \n\t"
|
|
351
|
+
"psubh $f22, $f22, $f30 \n\t"
|
|
352
|
+
"gssqc1 $f22, $f20, 432-224(%[tmp]) \n\t"
|
|
353
|
+
"gslqc1 $f22, $f20, 432-400(%[tmp]) \n\t"
|
|
354
|
+
"gslqc1 $f30, $f28, 432-352(%[tmp]) \n\t"
|
|
355
|
+
"psubh $f20, $f24, $f20 \n\t"
|
|
356
|
+
"psubh $f22, $f26, $f22 \n\t"
|
|
357
|
+
"psubh $f24, $f24, $f28 \n\t"
|
|
358
|
+
"psubh $f26, $f26, $f30 \n\t"
|
|
359
|
+
"gssqc1 $f22, $f20, 432-272(%[tmp]) \n\t"
|
|
360
|
+
"mov.d $f28, $f20 \n\t"
|
|
361
|
+
"mov.d $f30, $f22 \n\t"
|
|
362
|
+
WELS_AbsH($f28, $f30, $f20, $f22, $f0, $f2)
|
|
363
|
+
"gslqc1 $f22, $f20, 432-112(%[tmp]) \n\t"
|
|
364
|
+
"pcmpgth $f20, $f20, $f28 \n\t"
|
|
365
|
+
"pcmpgth $f22, $f22, $f30 \n\t"
|
|
366
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
367
|
+
"pcmpgth $f28, $f16, $f24 \n\t"
|
|
368
|
+
"pcmpgth $f30, $f18, $f26 \n\t"
|
|
369
|
+
"gslqc1 $f26, $f24, 432-368(%[tmp]) \n\t"
|
|
370
|
+
|
|
371
|
+
"and $f20, $f20, $f28 \n\t"
|
|
372
|
+
"and $f22, $f22, $f30 \n\t"
|
|
373
|
+
"gslqc1 $f30, $f28, 432-400(%[tmp]) \n\t"
|
|
374
|
+
"psubh $f28, $f28, $f24 \n\t"
|
|
375
|
+
"psubh $f30, $f30, $f26 \n\t"
|
|
376
|
+
"gslqc1 $f2, $f0, 432-352(%[tmp]) \n\t"
|
|
377
|
+
"psubh $f24, $f24, $f0 \n\t"
|
|
378
|
+
"psubh $f26, $f26, $f2 \n\t"
|
|
379
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f0, $f2)
|
|
380
|
+
"pcmpgth $f16, $f16, $f28 \n\t"
|
|
381
|
+
"pcmpgth $f18, $f18, $f30 \n\t"
|
|
382
|
+
"gslqc1 $f30, $f28, 432-96(%[tmp]) \n\t"
|
|
383
|
+
"and $f20, $f20, $f16 \n\t"
|
|
384
|
+
"and $f22, $f22, $f18 \n\t"
|
|
385
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
386
|
+
|
|
387
|
+
"paddh $f8, $f8, $f28 \n\t"
|
|
388
|
+
"paddh $f10, $f10, $f30 \n\t"
|
|
389
|
+
"pcmpgth $f16, $f4, $f0 \n\t"
|
|
390
|
+
"pcmpgth $f18, $f6, $f0 \n\t"
|
|
391
|
+
"pcmpeqh $f28, $f4, $f0 \n\t"
|
|
392
|
+
"pcmpeqh $f30, $f6, $f0 \n\t"
|
|
393
|
+
"or $f16, $f16, $f28 \n\t"
|
|
394
|
+
"or $f18, $f18, $f30 \n\t"
|
|
395
|
+
"and $f20, $f20, $f16 \n\t"
|
|
396
|
+
"and $f22, $f22, $f18 \n\t"
|
|
397
|
+
"gslqc1 $f18, $f16, 432-224(%[tmp]) \n\t"
|
|
398
|
+
"gssqc1 $f22, $f20, 432-320(%[tmp]) \n\t"
|
|
399
|
+
"gslqc1 $f22, $f20, 432-272(%[tmp]) \n\t"
|
|
400
|
+
"dli $11, 0x2 \n\t"
|
|
401
|
+
"psubh $f28, $f0, $f16 \n\t"
|
|
402
|
+
"psubh $f30, $f0, $f18 \n\t"
|
|
403
|
+
"psubh $f2, $f0, $f6 \n\t"
|
|
404
|
+
"psubh $f0, $f0, $f4 \n\t"
|
|
405
|
+
"dmfc1 %[iAlpha], $f28 \n\t"
|
|
406
|
+
"dmtc1 $11, $f28 \n\t"
|
|
407
|
+
"psllh $f20, $f20, $f28 \n\t"
|
|
408
|
+
"psllh $f22, $f22, $f28 \n\t"
|
|
409
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
410
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
411
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
412
|
+
"gslqc1 $f22, $f20, 432-336(%[tmp]) \n\t"
|
|
413
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
414
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
415
|
+
"gslqc1 $f22, $f20, 432-368(%[tmp]) \n\t"
|
|
416
|
+
"dli $11, 0x3 \n\t"
|
|
417
|
+
"gssqc1 $f2, $f0, 432-336(%[tmp]) \n\t"
|
|
418
|
+
"dmfc1 %[iAlpha], $f0 \n\t"
|
|
419
|
+
"dmtc1 $11, $f0 \n\t"
|
|
420
|
+
"psrah $f24, $f24, $f0 \n\t"
|
|
421
|
+
"psrah $f26, $f26, $f0 \n\t"
|
|
422
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
423
|
+
"pmaxsh $f28, $f28, $f24 \n\t"
|
|
424
|
+
"pmaxsh $f30, $f30, $f26 \n\t"
|
|
425
|
+
"pminsh $f16, $f16, $f28 \n\t"
|
|
426
|
+
"pminsh $f18, $f18, $f30 \n\t"
|
|
427
|
+
"gslqc1 $f30, $f28, 432-320(%[tmp]) \n\t"
|
|
428
|
+
"and $f16, $f16, $f28 \n\t"
|
|
429
|
+
"and $f18, $f18, $f30 \n\t"
|
|
430
|
+
"mov.d $f24, $f0 \n\t"
|
|
431
|
+
"mov.d $f26, $f2 \n\t"
|
|
432
|
+
"gslqc1 $f2, $f0, 432-16(%[tmp]) \n\t"
|
|
433
|
+
"gslqc1 $f30, $f28, 432-304(%[tmp]) \n\t"
|
|
434
|
+
"paddh $f0, $f0, $f28 \n\t"
|
|
435
|
+
"paddh $f2, $f2, $f30 \n\t"
|
|
436
|
+
"gssqc1 $f18, $f16, 432-272(%[tmp]) \n\t"
|
|
437
|
+
"gslqc1 $f18, $f16, 432-368(%[tmp]) \n\t"
|
|
438
|
+
"dli $11, 0x1 \n\t"
|
|
439
|
+
"paddh $f16, $f16, $f16 \n\t"
|
|
440
|
+
"paddh $f18, $f18, $f18 \n\t"
|
|
441
|
+
"psubh $f0, $f0, $f16 \n\t"
|
|
442
|
+
"psubh $f2, $f2, $f18 \n\t"
|
|
443
|
+
|
|
444
|
+
"dmtc1 $11, $f28 \n\t"
|
|
445
|
+
"gslqc1 $f18, $f16, 432-64(%[tmp]) \n\t"
|
|
446
|
+
"psrah $f0, $f0, $f28 \n\t"
|
|
447
|
+
"psrah $f2, $f2, $f28 \n\t"
|
|
448
|
+
"pmaxsh $f24, $f24, $f0 \n\t"
|
|
449
|
+
"pmaxsh $f26, $f26, $f2 \n\t"
|
|
450
|
+
"gslqc1 $f2, $f0, 432-400(%[tmp]) \n\t"
|
|
451
|
+
"pminsh $f28, $f4, $f24 \n\t"
|
|
452
|
+
"pminsh $f30, $f6, $f26 \n\t"
|
|
453
|
+
"gslqc1 $f26, $f24, 432-320(%[tmp]) \n\t"
|
|
454
|
+
"and $f28, $f28, $f24 \n\t"
|
|
455
|
+
"and $f30, $f30, $f26 \n\t"
|
|
456
|
+
"dmfc1 %[iAlpha], $f24 \n\t"
|
|
457
|
+
"dmfc1 %[iBeta], $f26 \n\t"
|
|
458
|
+
"gslqc1 $f26, $f24, 432-288(%[tmp]) \n\t"
|
|
459
|
+
"and $f28, $f28, $f24 \n\t"
|
|
460
|
+
"and $f30, $f30, $f26 \n\t"
|
|
461
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
462
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
463
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
464
|
+
"packushb $f10, $f20, $f22 \n\t"
|
|
465
|
+
"gslqc1 $f22, $f20, 432-272(%[tmp]) \n\t"
|
|
466
|
+
"paddh $f0, $f0, $f20 \n\t"
|
|
467
|
+
"paddh $f2, $f2, $f22 \n\t"
|
|
468
|
+
"paddh $f12, $f12, $f16 \n\t"
|
|
469
|
+
"paddh $f14, $f14, $f18 \n\t"
|
|
470
|
+
"packushb $f12, $f12, $f14 \n\t"
|
|
471
|
+
"packushb $f14, $f0, $f2 \n\t"
|
|
472
|
+
|
|
473
|
+
"gslqc1 $f2, $f0, 432-32(%[tmp]) \n\t"
|
|
474
|
+
"psubh $f0, $f0, $f16 \n\t"
|
|
475
|
+
"psubh $f2, $f2, $f18 \n\t"
|
|
476
|
+
"gslqc1 $f18, $f16, 432-80(%[tmp]) \n\t"
|
|
477
|
+
"psubh $f16, $f16, $f20 \n\t"
|
|
478
|
+
"gslqc1 $f26, $f24, 432-48(%[tmp]) \n\t"
|
|
479
|
+
"psubh $f18, $f18, $f22 \n\t"
|
|
480
|
+
|
|
481
|
+
"gslqc1 $f22, $f20, 432-240(%[tmp]) \n\t"
|
|
482
|
+
"paddh $f20, $f20, $f24 \n\t"
|
|
483
|
+
"paddh $f22, $f22, $f26 \n\t"
|
|
484
|
+
"gslqc1 $f26, $f24, 432-304(%[tmp]) \n\t"
|
|
485
|
+
"packushb $f0, $f0, $f2 \n\t"
|
|
486
|
+
"packushb $f2, $f16, $f18 \n\t"
|
|
487
|
+
"gslqc1 $f18, $f16, 432-384(%[tmp]) \n\t"
|
|
488
|
+
"paddh $f16, $f16, $f24 \n\t"
|
|
489
|
+
"paddh $f18, $f18, $f26 \n\t"
|
|
490
|
+
"gssqc1 $f2, $f0, 480-208(%[tmp]) \n\t"
|
|
491
|
+
"gslqc1 $f2, $f0, 432-352(%[tmp]) \n\t"
|
|
492
|
+
"mov.d $f28, $f0 \n\t"
|
|
493
|
+
"mov.d $f30, $f2 \n\t"
|
|
494
|
+
"paddh $f0, $f0, $f0 \n\t"
|
|
495
|
+
"paddh $f2, $f2, $f2 \n\t"
|
|
496
|
+
|
|
497
|
+
"dmtc1 %[iAlpha], $f24 \n\t"
|
|
498
|
+
"dmtc1 %[iBeta], $f26 \n\t"
|
|
499
|
+
|
|
500
|
+
"psubh $f16, $f16, $f0 \n\t"
|
|
501
|
+
"psubh $f18, $f18, $f2 \n\t"
|
|
502
|
+
"dli $11, 0x1 \n\t"
|
|
503
|
+
"gslqc1 $f2, $f0, 432-336(%[tmp]) \n\t"
|
|
504
|
+
"gssqc1 $f10, $f8, 0x0($9) \n\t"
|
|
505
|
+
"dmtc1 $11, $f8 \n\t"
|
|
506
|
+
"psrah $f16, $f16, $f8 \n\t"
|
|
507
|
+
"psrah $f18, $f18, $f8 \n\t"
|
|
508
|
+
"pmaxsh $f0, $f0, $f16 \n\t"
|
|
509
|
+
"pmaxsh $f2, $f2, $f18 \n\t"
|
|
510
|
+
"pminsh $f4, $f4, $f0 \n\t"
|
|
511
|
+
"pminsh $f6, $f6, $f2 \n\t"
|
|
512
|
+
"gslqc1 $f2, $f0, 480-208(%[tmp]) \n\t"
|
|
513
|
+
|
|
514
|
+
"gslqc1 $f10, $f8, 428-256+4(%[tmp]) \n\t"
|
|
515
|
+
"and $f4, $f4, $f24 \n\t"
|
|
516
|
+
"and $f6, $f6, $f26 \n\t"
|
|
517
|
+
"and $f4, $f4, $f8 \n\t"
|
|
518
|
+
"and $f6, $f6, $f10 \n\t"
|
|
519
|
+
"gssqc1 $f14, $f12, 0x0($13) \n\t"
|
|
520
|
+
"paddh $f28, $f28, $f4 \n\t"
|
|
521
|
+
"paddh $f30, $f30, $f6 \n\t"
|
|
522
|
+
"packushb $f20, $f20, $f22 \n\t"
|
|
523
|
+
"packushb $f22, $f28, $f30 \n\t"
|
|
524
|
+
"gssqc1 $f2, $f0, 0x0(%[pPix]) \n\t"
|
|
525
|
+
"gssqc1 $f22, $f20, 0x0(%[iStride]) \n\t"
|
|
526
|
+
: [pPix]"+&r"((unsigned char *)pPix)
|
|
527
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha), [iBeta]"r"(iBeta),
|
|
528
|
+
[pTC]"r"((unsigned char *)pTC), [tmp]"r"((unsigned char *)tmp)
|
|
529
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$13", "$14", "$f0", "$f2",
|
|
530
|
+
"$f4", "$f6", "$f8", "$f10", "$f12", "$f14", "$f16", "$f18", "$f20",
|
|
531
|
+
"$f22", "$f24", "$f26", "$f28", "$f30"
|
|
532
|
+
);
|
|
533
|
+
RECOVER_REG;
|
|
534
|
+
}
|
|
535
|
+
|
|
536
|
+
void DeblockLumaTransposeH2V_mmi(uint8_t *pPixY, int32_t iStride,
|
|
537
|
+
uint8_t *pDst) {
|
|
538
|
+
BACKUP_REG;
|
|
539
|
+
__asm__ volatile(
|
|
540
|
+
".set arch=loongson3a \n\t"
|
|
541
|
+
"dsll $8, %[iStride], 0x3 \n\t"
|
|
542
|
+
"daddu $8, $8, %[pPixY] \n\t"
|
|
543
|
+
|
|
544
|
+
"daddu $9, %[pPixY], %[iStride] \n\t"
|
|
545
|
+
"daddu $10, $8, %[iStride] \n\t"
|
|
546
|
+
"gsldlc1 $f0, 0x7(%[pPixY]) \n\t"
|
|
547
|
+
"gsldlc1 $f2, 0x7($8) \n\t"
|
|
548
|
+
"gsldlc1 $f4, 0x7($9) \n\t"
|
|
549
|
+
"gsldlc1 $f6, 0x7($10) \n\t"
|
|
550
|
+
"gsldrc1 $f0, 0x0(%[pPixY]) \n\t"
|
|
551
|
+
"gsldrc1 $f2, 0x0($8) \n\t"
|
|
552
|
+
"gsldrc1 $f4, 0x0($9) \n\t"
|
|
553
|
+
"gsldrc1 $f6, 0x0($10) \n\t"
|
|
554
|
+
"daddu %[pPixY], $9, %[iStride] \n\t"
|
|
555
|
+
"daddu $8, $10, %[iStride] \n\t"
|
|
556
|
+
"daddu $9, %[pPixY], %[iStride] \n\t"
|
|
557
|
+
"daddu $10, $8, %[iStride] \n\t"
|
|
558
|
+
"gsldlc1 $f8, 0x7(%[pPixY]) \n\t"
|
|
559
|
+
"gsldlc1 $f10, 0x7($8) \n\t"
|
|
560
|
+
"gsldlc1 $f12, 0x7($9) \n\t"
|
|
561
|
+
"gsldlc1 $f14, 0x7($10) \n\t"
|
|
562
|
+
"gsldrc1 $f8, 0x0(%[pPixY]) \n\t"
|
|
563
|
+
"gsldrc1 $f10, 0x0($8) \n\t"
|
|
564
|
+
"gsldrc1 $f12, 0x0($9) \n\t"
|
|
565
|
+
"gsldrc1 $f14, 0x0($10) \n\t"
|
|
566
|
+
|
|
567
|
+
"daddu %[pPixY], $9, %[iStride] \n\t"
|
|
568
|
+
"daddu $8, $10, %[iStride] \n\t"
|
|
569
|
+
"daddu $9, %[pPixY], %[iStride] \n\t"
|
|
570
|
+
"daddu $10, $8, %[iStride] \n\t"
|
|
571
|
+
"gsldlc1 $f16, 0x7(%[pPixY]) \n\t"
|
|
572
|
+
"gsldlc1 $f18, 0x7($8) \n\t"
|
|
573
|
+
"gsldlc1 $f20, 0x7($9) \n\t"
|
|
574
|
+
"gsldlc1 $f22, 0x7($10) \n\t"
|
|
575
|
+
"gsldrc1 $f16, 0x0(%[pPixY]) \n\t"
|
|
576
|
+
"gsldrc1 $f18, 0x0($8) \n\t"
|
|
577
|
+
"gsldrc1 $f20, 0x0($9) \n\t"
|
|
578
|
+
"gsldrc1 $f22, 0x0($10) \n\t"
|
|
579
|
+
"daddu %[pPixY], $9, %[iStride] \n\t"
|
|
580
|
+
"daddu $8, $10, %[iStride] \n\t"
|
|
581
|
+
"daddu $9, %[pPixY], %[iStride] \n\t"
|
|
582
|
+
"daddu $10, $8, %[iStride] \n\t"
|
|
583
|
+
"gsldlc1 $f24, 0x7(%[pPixY]) \n\t"
|
|
584
|
+
"gsldlc1 $f26, 0x7($8) \n\t"
|
|
585
|
+
|
|
586
|
+
"gsldlc1 $f28, 0x7($9) \n\t"
|
|
587
|
+
"gsldlc1 $f30, 0x7($10) \n\t"
|
|
588
|
+
"gsldrc1 $f24, 0x0(%[pPixY]) \n\t"
|
|
589
|
+
"gsldrc1 $f26, 0x0($8) \n\t"
|
|
590
|
+
"gsldrc1 $f28, 0x0($9) \n\t"
|
|
591
|
+
"gsldrc1 $f30, 0x0($10) \n\t"
|
|
592
|
+
|
|
593
|
+
MMI_TransTwo8x8B($f0, $f2, $f4, $f6, $f8, $f10, $f12,
|
|
594
|
+
$f14, $f16, $f18, $f20, $f22, $f24,
|
|
595
|
+
$f26, $f28, $f30, $9, $10)
|
|
596
|
+
|
|
597
|
+
"gssqc1 $f18, $f16, 0x0(%[pDst]) \n\t"
|
|
598
|
+
"gssqc1 $f10, $f8, 0x10(%[pDst]) \n\t"
|
|
599
|
+
"gssqc1 $f14, $f12, 0x20(%[pDst]) \n\t"
|
|
600
|
+
"gssqc1 $f30, $f28, 0x30(%[pDst]) \n\t"
|
|
601
|
+
"gssqc1 $f22, $f20, 0x40(%[pDst]) \n\t"
|
|
602
|
+
"gssqc1 $f6, $f4, 0x50(%[pDst]) \n\t"
|
|
603
|
+
"gssqc1 $f26, $f24, 0x60(%[pDst]) \n\t"
|
|
604
|
+
"gssqc1 $f2, $f0, 0x70(%[pDst]) \n\t"
|
|
605
|
+
: [pPixY] "+&r"((unsigned char *)pPixY)
|
|
606
|
+
: [iStride] "r"((int)iStride), [pDst] "r"((unsigned char *)pDst)
|
|
607
|
+
: "memory", "$8", "$9", "$10", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10",
|
|
608
|
+
"$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28",
|
|
609
|
+
"$f30"
|
|
610
|
+
);
|
|
611
|
+
RECOVER_REG;
|
|
612
|
+
}
|
|
613
|
+
|
|
614
|
+
void DeblockLumaTransposeV2H_mmi(uint8_t *pPixY, int32_t iStride,
|
|
615
|
+
uint8_t *pSrc) {
|
|
616
|
+
BACKUP_REG;
|
|
617
|
+
__asm__ volatile(
|
|
618
|
+
".set arch=loongson3a \n\t"
|
|
619
|
+
"gslqc1 $f2, $f0, 0x0(%[pSrc]) \n\t"
|
|
620
|
+
"gslqc1 $f6, $f4, 0x10(%[pSrc]) \n\t"
|
|
621
|
+
"gslqc1 $f10, $f8, 0x20(%[pSrc]) \n\t"
|
|
622
|
+
"gslqc1 $f14, $f12, 0x30(%[pSrc]) \n\t"
|
|
623
|
+
"gslqc1 $f18, $f16, 0x40(%[pSrc]) \n\t"
|
|
624
|
+
"gslqc1 $f22, $f20, 0x50(%[pSrc]) \n\t"
|
|
625
|
+
"gslqc1 $f26, $f24, 0x60(%[pSrc]) \n\t"
|
|
626
|
+
"gslqc1 $f30, $f28, 0x70(%[pSrc]) \n\t"
|
|
627
|
+
|
|
628
|
+
MMI_TransTwo8x8B($f0, $f2, $f4, $f6, $f8, $f10, $f12,
|
|
629
|
+
$f14, $f16, $f18, $f20, $f22, $f24,
|
|
630
|
+
$f26, $f28, $f30, $9, $10)
|
|
631
|
+
|
|
632
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
633
|
+
"gssdlc1 $f16, 0x7(%[pPixY]) \n\t"
|
|
634
|
+
"gssdlc1 $f8, 0x7($8) \n\t"
|
|
635
|
+
"gssdrc1 $f16, 0x0(%[pPixY]) \n\t"
|
|
636
|
+
"gssdrc1 $f8, 0x0($8) \n\t"
|
|
637
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
638
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
639
|
+
"gssdlc1 $f12, 0x7(%[pPixY]) \n\t"
|
|
640
|
+
"gssdlc1 $f28, 0x7($8) \n\t"
|
|
641
|
+
"gssdrc1 $f12, 0x0(%[pPixY]) \n\t"
|
|
642
|
+
"gssdrc1 $f28, 0x0($8) \n\t"
|
|
643
|
+
|
|
644
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
645
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
646
|
+
"gssdlc1 $f20, 0x7(%[pPixY]) \n\t"
|
|
647
|
+
"gssdlc1 $f4, 0x7($8) \n\t"
|
|
648
|
+
"gssdrc1 $f20, 0x0(%[pPixY]) \n\t"
|
|
649
|
+
"gssdrc1 $f4, 0x0($8) \n\t"
|
|
650
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
651
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
652
|
+
"gssdlc1 $f24, 0x7(%[pPixY]) \n\t"
|
|
653
|
+
"gssdlc1 $f0, 0x7($8) \n\t"
|
|
654
|
+
"gssdrc1 $f24, 0x0(%[pPixY]) \n\t"
|
|
655
|
+
"gssdrc1 $f0, 0x0($8) \n\t"
|
|
656
|
+
|
|
657
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
658
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
659
|
+
"gssdlc1 $f18, 0x7(%[pPixY]) \n\t"
|
|
660
|
+
"gssdlc1 $f10, 0x7($8) \n\t"
|
|
661
|
+
"gssdrc1 $f18, 0x0(%[pPixY]) \n\t"
|
|
662
|
+
"gssdrc1 $f10, 0x0($8) \n\t"
|
|
663
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
664
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
665
|
+
"gssdlc1 $f14, 0x7(%[pPixY]) \n\t"
|
|
666
|
+
"gssdlc1 $f30, 0x7($8) \n\t"
|
|
667
|
+
"gssdrc1 $f14, 0x0(%[pPixY]) \n\t"
|
|
668
|
+
"gssdrc1 $f30, 0x0($8) \n\t"
|
|
669
|
+
|
|
670
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
671
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
672
|
+
"gssdlc1 $f22, 0x7(%[pPixY]) \n\t"
|
|
673
|
+
"gssdlc1 $f6, 0x7($8) \n\t"
|
|
674
|
+
"gssdrc1 $f22, 0x0(%[pPixY]) \n\t"
|
|
675
|
+
"gssdrc1 $f6, 0x0($8) \n\t"
|
|
676
|
+
"daddu %[pPixY], $8, %[iStride] \n\t"
|
|
677
|
+
"daddu $8, %[pPixY], %[iStride] \n\t"
|
|
678
|
+
"gssdlc1 $f26, 0x7(%[pPixY]) \n\t"
|
|
679
|
+
"gssdlc1 $f2, 0x7($8) \n\t"
|
|
680
|
+
"gssdrc1 $f26, 0x0(%[pPixY]) \n\t"
|
|
681
|
+
"gssdrc1 $f2, 0x0($8) \n\t"
|
|
682
|
+
: [pPixY] "+&r"((unsigned char *)pPixY)
|
|
683
|
+
: [iStride] "r"((int)iStride), [pSrc] "r"((unsigned char *)pSrc)
|
|
684
|
+
: "memory", "$8", "$9", "$10", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10",
|
|
685
|
+
"$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28",
|
|
686
|
+
"$f30"
|
|
687
|
+
);
|
|
688
|
+
RECOVER_REG;
|
|
689
|
+
}
|
|
690
|
+
|
|
691
|
+
void DeblockLumaEq4V_mmi(uint8_t *pPix, int32_t iStride, int32_t iAlpha,
|
|
692
|
+
int32_t iBeta) {
|
|
693
|
+
unsigned char tmp[720] __attribute__((aligned(32)));
|
|
694
|
+
BACKUP_REG;
|
|
695
|
+
__asm__ volatile (
|
|
696
|
+
".set arch=loongson3a \n\t"
|
|
697
|
+
"dsll $11, %[iStride], 0x2 \n\t"
|
|
698
|
+
"xor $f8, $f8, $f8 \n\t"
|
|
699
|
+
"daddu $14, %[iStride], %[pPix] \n\t"
|
|
700
|
+
"dsubu $8, %[pPix], $11 \n\t"
|
|
701
|
+
"gslqc1 $f14, $f12, 0x0($8) \n\t"
|
|
702
|
+
"gslqc1 $f22, $f20, 0x0(%[pPix]) \n\t"
|
|
703
|
+
"daddu $9, %[iStride], %[iStride] \n\t"
|
|
704
|
+
"daddu $10, $9, %[iStride] \n\t"
|
|
705
|
+
"move $12, $9 \n\t"
|
|
706
|
+
"dsubu $8, %[pPix], $9 \n\t"
|
|
707
|
+
"gslqc1 $f6, $f4, 0x0($8) \n\t"
|
|
708
|
+
"dsubu $9, %[pPix], %[iStride] \n\t"
|
|
709
|
+
"gslqc1 $f18, $f16, 0x0($9) \n\t"
|
|
710
|
+
"daddu $13, %[iStride], %[pPix] \n\t"
|
|
711
|
+
|
|
712
|
+
"move %[iStride], $12 \n\t"
|
|
713
|
+
"daddu $15, $12, %[pPix] \n\t"
|
|
714
|
+
|
|
715
|
+
"daddu $12, %[pPix], $10 \n\t"
|
|
716
|
+
"dsubu $11, %[pPix], $10 \n\t"
|
|
717
|
+
|
|
718
|
+
"gslqc1 $f26, $f24, 0x0($11) \n\t"
|
|
719
|
+
"daddu %[iStride], %[iStride], %[pPix] \n\t"
|
|
720
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
721
|
+
|
|
722
|
+
"punpcklhw $f28, $f0, $f0 \n\t"
|
|
723
|
+
"punpcklwd $f0, $f28, $f28 \n\t"
|
|
724
|
+
"mov.d $f2, $f0 \n\t"
|
|
725
|
+
"gssqc1 $f2, $f0, 640-320(%[tmp]) \n\t"
|
|
726
|
+
"dmtc1 %[iBeta], $f0 \n\t"
|
|
727
|
+
"gsldxc1 $f10, 0x0($15, $0) \n\t"
|
|
728
|
+
"punpcklhw $f28, $f0, $f0 \n\t"
|
|
729
|
+
"punpcklwd $f0, $f28, $f28 \n\t"
|
|
730
|
+
"punpckhbh $f30, $f10, $f8 \n\t"
|
|
731
|
+
"mov.d $f2, $f0 \n\t"
|
|
732
|
+
|
|
733
|
+
"punpcklbh $f28, $f10, $f8 \n\t"
|
|
734
|
+
"gssqc1 $f2, $f0, 640-512(%[tmp]) \n\t"
|
|
735
|
+
"gssqc1 $f30, $f28, 640-416(%[tmp]) \n\t"
|
|
736
|
+
"mov.d $f0, $f4 \n\t"
|
|
737
|
+
"gssqc1 $f22, $f20, 704-272(%[tmp]) \n\t"
|
|
738
|
+
"gssqc1 $f6, $f4, 672-272(%[tmp]) \n\t"
|
|
739
|
+
"mov.d $f4, $f16 \n\t"
|
|
740
|
+
"punpckhbh $f22, $f20, $f8 \n\t"
|
|
741
|
+
"punpcklbh $f20, $f20, $f8 \n\t"
|
|
742
|
+
"punpckhbh $f6, $f4, $f8 \n\t"
|
|
743
|
+
"punpcklbh $f4, $f4, $f8 \n\t"
|
|
744
|
+
|
|
745
|
+
"psubh $f28, $f20, $f4 \n\t"
|
|
746
|
+
"psubh $f30, $f22, $f6 \n\t"
|
|
747
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f2, $f10)
|
|
748
|
+
"gssqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
749
|
+
"punpckhbh $f2, $f0, $f8 \n\t"
|
|
750
|
+
"punpcklbh $f0, $f0, $f8 \n\t"
|
|
751
|
+
"gssqc1 $f18, $f16, 688-272(%[tmp]) \n\t"
|
|
752
|
+
"gslqc1 $f18, $f16, 0x0($14) \n\t"
|
|
753
|
+
"gssqc1 $f2, $f0, 640-480(%[tmp]) \n\t"
|
|
754
|
+
|
|
755
|
+
"psubh $f28, $f4, $f0 \n\t"
|
|
756
|
+
"psubh $f30, $f6, $f2 \n\t"
|
|
757
|
+
|
|
758
|
+
"gslqc1 $f2, $f0, 640-512(%[tmp]) \n\t"
|
|
759
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f18, $f10)
|
|
760
|
+
"punpckhbh $f18, $f16, $f8 \n\t"
|
|
761
|
+
"punpcklbh $f16, $f16, $f8 \n\t"
|
|
762
|
+
"pcmpgth $f0, $f0, $f28 \n\t"
|
|
763
|
+
"pcmpgth $f2, $f2, $f30 \n\t"
|
|
764
|
+
"gssqc1 $f18, $f16, 640-384(%[tmp]) \n\t"
|
|
765
|
+
"psubh $f28, $f20, $f16 \n\t"
|
|
766
|
+
"psubh $f30, $f22, $f18 \n\t"
|
|
767
|
+
"gslqc1 $f18, $f16, 640-512(%[tmp]) \n\t"
|
|
768
|
+
"gssqc1 $f26, $f24, 656-272(%[tmp]) \n\t"
|
|
769
|
+
"punpckhbh $f26, $f24, $f8 \n\t"
|
|
770
|
+
"punpcklbh $f24, $f24, $f8 \n\t"
|
|
771
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
772
|
+
"gssqc1 $f26, $f24, 640-368(%[tmp]) \n\t"
|
|
773
|
+
"gssqc1 $f6, $f4, 640-144(%[tmp]) \n\t"
|
|
774
|
+
"gssqc1 $f22, $f20, 640-400(%[tmp]) \n\t"
|
|
775
|
+
"pcmpgth $f16, $f16, $f28 \n\t"
|
|
776
|
+
"pcmpgth $f18, $f18, $f30 \n\t"
|
|
777
|
+
"and $f0, $f0, $f16 \n\t"
|
|
778
|
+
"and $f2, $f2, $f18 \n\t"
|
|
779
|
+
"gslqc1 $f18, $f16, 640-320(%[tmp]) \n\t"
|
|
780
|
+
"gslqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
781
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
782
|
+
"dli %[iBeta], 0x2 \n\t"
|
|
783
|
+
"pcmpgth $f16, $f16, $f28 \n\t"
|
|
784
|
+
"pcmpgth $f18, $f18, $f30 \n\t"
|
|
785
|
+
"and $f0, $f0, $f16 \n\t"
|
|
786
|
+
"and $f2, $f2, $f18 \n\t"
|
|
787
|
+
"dmtc1 %[iAlpha], $f16 \n\t"
|
|
788
|
+
"dmtc1 %[iBeta], $f10 \n\t"
|
|
789
|
+
"gssqc1 $f2, $f0, 736-272(%[tmp]) \n\t"
|
|
790
|
+
"gslqc1 $f2, $f0, 640-320(%[tmp]) \n\t"
|
|
791
|
+
|
|
792
|
+
"punpcklhw $f28, $f16, $f16 \n\t"
|
|
793
|
+
"psrah $f16, $f0, $f10 \n\t"
|
|
794
|
+
"psrah $f18, $f2, $f10 \n\t"
|
|
795
|
+
"punpcklwd $f28, $f28, $f28 \n\t"
|
|
796
|
+
"mov.d $f30, $f28 \n\t"
|
|
797
|
+
"gslqc1 $f10, $f8, 640-560(%[tmp]) \n\t"
|
|
798
|
+
"paddh $f16, $f16, $f28 \n\t"
|
|
799
|
+
"paddh $f18, $f18, $f30 \n\t"
|
|
800
|
+
"gssqc1 $f18, $f16, 640-576(%[tmp]) \n\t"
|
|
801
|
+
"pcmpgth $f16, $f16, $f8 \n\t"
|
|
802
|
+
"pcmpgth $f18, $f18, $f10 \n\t"
|
|
803
|
+
"gssqc1 $f18, $f16, 640-560(%[tmp]) \n\t"
|
|
804
|
+
|
|
805
|
+
"gssqc1 $f30, $f28, 640-624(%[tmp]) \n\t"
|
|
806
|
+
"gslqc1 $f18, $f16, 640-512(%[tmp]) \n\t"
|
|
807
|
+
"psubh $f28, $f4, $f24 \n\t"
|
|
808
|
+
"psubh $f30, $f6, $f26 \n\t"
|
|
809
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
810
|
+
"gslqc1 $f10, $f8, 640-560(%[tmp]) \n\t"
|
|
811
|
+
"pcmpgth $f16, $f16, $f28 \n\t"
|
|
812
|
+
"pcmpgth $f18, $f18, $f30 \n\t"
|
|
813
|
+
|
|
814
|
+
"gslqc1 $f2, $f0, 640-416(%[tmp]) \n\t"
|
|
815
|
+
"and $f16, $f16, $f8 \n\t"
|
|
816
|
+
"and $f18, $f18, $f10 \n\t"
|
|
817
|
+
"gssqc1 $f18, $f16, 640-544(%[tmp]) \n\t"
|
|
818
|
+
"gslqc1 $f18, $f16, 640-512(%[tmp]) \n\t"
|
|
819
|
+
"psubh $f28, $f20, $f0 \n\t"
|
|
820
|
+
"psubh $f30, $f22, $f2 \n\t"
|
|
821
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
822
|
+
"gslqc1 $f10, $f8, 640-560(%[tmp]) \n\t"
|
|
823
|
+
"pcmpgth $f16, $f16, $f28 \n\t"
|
|
824
|
+
"pcmpgth $f18, $f18, $f30 \n\t"
|
|
825
|
+
|
|
826
|
+
"and $f16, $f16, $f8 \n\t"
|
|
827
|
+
"and $f18, $f18, $f10 \n\t"
|
|
828
|
+
"gssqc1 $f18, $f16, 640-560(%[tmp]) \n\t"
|
|
829
|
+
|
|
830
|
+
"gslqc1 $f18, $f16, 640-544(%[tmp]) \n\t"
|
|
831
|
+
"xor $f8, $f8, $f8 \n\t"
|
|
832
|
+
"pandn $f16, $f16, $f24 \n\t"
|
|
833
|
+
"dli %[iAlpha], 0x4 \n\t"
|
|
834
|
+
"pandn $f18, $f18, $f26 \n\t"
|
|
835
|
+
"gssqc1 $f18, $f16, 640-16(%[tmp]) \n\t"
|
|
836
|
+
"dmtc1 %[iAlpha], $f16 \n\t"
|
|
837
|
+
"punpcklhw $f28, $f16, $f16 \n\t"
|
|
838
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
839
|
+
"punpckhbh $f18, $f12, $f8 \n\t"
|
|
840
|
+
"dmtc1 %[iAlpha], $f30 \n\t"
|
|
841
|
+
"punpcklbh $f16, $f12, $f8 \n\t"
|
|
842
|
+
"psllh $f16, $f16, $f30 \n\t"
|
|
843
|
+
"psllh $f18, $f18, $f30 \n\t"
|
|
844
|
+
"paddh $f16, $f16, $f24 \n\t"
|
|
845
|
+
"paddh $f18, $f18, $f26 \n\t"
|
|
846
|
+
"gslqc1 $f2, $f0, 640-480(%[tmp]) \n\t"
|
|
847
|
+
"paddh $f16, $f16, $f24 \n\t"
|
|
848
|
+
"paddh $f18, $f18, $f26 \n\t"
|
|
849
|
+
"paddh $f16, $f16, $f24 \n\t"
|
|
850
|
+
"paddh $f18, $f18, $f26 \n\t"
|
|
851
|
+
"paddh $f16, $f16, $f0 \n\t"
|
|
852
|
+
"paddh $f18, $f18, $f2 \n\t"
|
|
853
|
+
|
|
854
|
+
"gslqc1 $f26, $f24, 640-560(%[tmp]) \n\t"
|
|
855
|
+
"punpcklwd $f28, $f28, $f28 \n\t"
|
|
856
|
+
"mov.d $f30, $f28 \n\t"
|
|
857
|
+
"paddh $f16, $f16, $f4 \n\t"
|
|
858
|
+
"paddh $f18, $f18, $f6 \n\t"
|
|
859
|
+
"gssqc1 $f30, $f28, 640-592(%[tmp]) \n\t"
|
|
860
|
+
"paddh $f16, $f16, $f20 \n\t"
|
|
861
|
+
"paddh $f18, $f18, $f22 \n\t"
|
|
862
|
+
"paddh $f16, $f16, $f28 \n\t"
|
|
863
|
+
"paddh $f18, $f18, $f30 \n\t"
|
|
864
|
+
"gslqc1 $f30, $f28, 640-416(%[tmp]) \n\t"
|
|
865
|
+
"gslqc1 $f2, $f0, 640-384(%[tmp]) \n\t"
|
|
866
|
+
"pandn $f24, $f24, $f28 \n\t"
|
|
867
|
+
"pandn $f26, $f26, $f30 \n\t"
|
|
868
|
+
"gssqc1 $f26, $f24, 640-80(%[tmp]) \n\t"
|
|
869
|
+
"gslqc1 $f26, $f24, 0x0($12) \n\t"
|
|
870
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
871
|
+
"punpckhbh $f26, $f24, $f8 \n\t"
|
|
872
|
+
"punpcklbh $f24, $f24, $f8 \n\t"
|
|
873
|
+
"psllh $f24, $f24, $f10 \n\t"
|
|
874
|
+
"psllh $f26, $f26, $f10 \n\t"
|
|
875
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
876
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
877
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
878
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
879
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
880
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
881
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
882
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
883
|
+
|
|
884
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
885
|
+
"gslqc1 $f30, $f28, 640-480(%[tmp]) \n\t"
|
|
886
|
+
"gslqc1 $f2, $f0, 640-592(%[tmp]) \n\t"
|
|
887
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
888
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
889
|
+
"paddh $f24, $f24, $f4 \n\t"
|
|
890
|
+
"paddh $f26, $f26, $f6 \n\t"
|
|
891
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
892
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
893
|
+
"gslqc1 $f2, $f0, 640-560(%[tmp]) \n\t"
|
|
894
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
895
|
+
"psrah $f24, $f24, $f10 \n\t"
|
|
896
|
+
"psrah $f26, $f26, $f10 \n\t"
|
|
897
|
+
"and $f24, $f24, $f0 \n\t"
|
|
898
|
+
"and $f26, $f26, $f2 \n\t"
|
|
899
|
+
"gssqc1 $f26, $f24, 640-112(%[tmp]) \n\t"
|
|
900
|
+
"gslqc1 $f26, $f24, 640-544(%[tmp]) \n\t"
|
|
901
|
+
"pandn $f24, $f24, $f28 \n\t"
|
|
902
|
+
"pandn $f26, $f26, $f30 \n\t"
|
|
903
|
+
"gssqc1 $f26, $f24, 640-336(%[tmp]) \n\t"
|
|
904
|
+
"gslqc1 $f26, $f24, 640-544(%[tmp]) \n\t"
|
|
905
|
+
"gssqc1 $f26, $f24, 640-528(%[tmp]) \n\t"
|
|
906
|
+
"gslqc1 $f26, $f24, 640-368(%[tmp]) \n\t"
|
|
907
|
+
"gslqc1 $f2, $f0, 640-544(%[tmp]) \n\t"
|
|
908
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
909
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
910
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
911
|
+
"psrah $f16, $f16, $f10 \n\t"
|
|
912
|
+
"psrah $f18, $f18, $f10 \n\t"
|
|
913
|
+
"and $f16, $f16, $f0 \n\t"
|
|
914
|
+
"and $f18, $f18, $f2 \n\t"
|
|
915
|
+
"gslqc1 $f2, $f0, 640-624(%[tmp]) \n\t"
|
|
916
|
+
"paddh $f28, $f4, $f20 \n\t"
|
|
917
|
+
"paddh $f30, $f6, $f22 \n\t"
|
|
918
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
919
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
920
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
921
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
922
|
+
"gslqc1 $f30, $f28, 640-528(%[tmp]) \n\t"
|
|
923
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
924
|
+
|
|
925
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
926
|
+
"paddh $f20, $f20, $f4 \n\t"
|
|
927
|
+
"paddh $f22, $f22, $f6 \n\t"
|
|
928
|
+
"psrah $f24, $f24, $f10 \n\t"
|
|
929
|
+
"psrah $f26, $f26, $f10 \n\t"
|
|
930
|
+
"and $f28, $f28, $f24 \n\t"
|
|
931
|
+
"and $f30, $f30, $f26 \n\t"
|
|
932
|
+
|
|
933
|
+
"gslqc1 $f26, $f24, 640-384(%[tmp]) \n\t"
|
|
934
|
+
"gssqc1 $f30, $f28, 640-64(%[tmp]) \n\t"
|
|
935
|
+
"gslqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
936
|
+
"pandn $f28, $f28, $f24 \n\t"
|
|
937
|
+
"pandn $f30, $f30, $f26 \n\t"
|
|
938
|
+
"gssqc1 $f30, $f28, 640-304(%[tmp]) \n\t"
|
|
939
|
+
"gslqc1 $f30, $f28, 640-416(%[tmp]) \n\t"
|
|
940
|
+
"gslqc1 $f10, $f8, 640-624(%[tmp]) \n\t"
|
|
941
|
+
"paddh $f28, $f28, $f24 \n\t"
|
|
942
|
+
"paddh $f30, $f30, $f26 \n\t"
|
|
943
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
944
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
945
|
+
"paddh $f28, $f28, $f8 \n\t"
|
|
946
|
+
"paddh $f30, $f30, $f10 \n\t"
|
|
947
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
948
|
+
"gslqc1 $f22, $f20, 640-560(%[tmp]) \n\t"
|
|
949
|
+
"psrah $f28, $f28, $f10 \n\t"
|
|
950
|
+
"psrah $f30, $f30, $f10 \n\t"
|
|
951
|
+
"and $f20, $f20, $f28 \n\t"
|
|
952
|
+
"and $f22, $f22, $f30 \n\t"
|
|
953
|
+
"gssqc1 $f22, $f20, 640-32(%[tmp]) \n\t"
|
|
954
|
+
|
|
955
|
+
"gslqc1 $f22, $f20, 640-480(%[tmp]) \n\t"
|
|
956
|
+
"gslqc1 $f2, $f0, 640-592(%[tmp]) \n\t"
|
|
957
|
+
"gslqc1 $f10, $f8, 640-624(%[tmp]) \n\t"
|
|
958
|
+
"paddh $f28, $f20, $f20 \n\t"
|
|
959
|
+
"paddh $f30, $f22, $f22 \n\t"
|
|
960
|
+
"paddh $f20, $f4, $f24 \n\t"
|
|
961
|
+
"paddh $f22, $f6, $f26 \n\t"
|
|
962
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
963
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
964
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
965
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
966
|
+
"paddh $f28, $f28, $f8 \n\t"
|
|
967
|
+
"paddh $f30, $f30, $f10 \n\t"
|
|
968
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
969
|
+
"gslqc1 $f22, $f20, 640-544(%[tmp]) \n\t"
|
|
970
|
+
"psrah $f28, $f28, $f10 \n\t"
|
|
971
|
+
"psrah $f30, $f30, $f10 \n\t"
|
|
972
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
973
|
+
"pandn $f20, $f20, $f28 \n\t"
|
|
974
|
+
"pandn $f22, $f22, $f30 \n\t"
|
|
975
|
+
"gslqc1 $f30, $f28, 640-480(%[tmp]) \n\t"
|
|
976
|
+
"paddh $f28, $f28, $f4 \n\t"
|
|
977
|
+
"paddh $f30, $f30, $f6 \n\t"
|
|
978
|
+
"gslqc1 $f6, $f4, 640-400(%[tmp]) \n\t"
|
|
979
|
+
"paddh $f28, $f28, $f4 \n\t"
|
|
980
|
+
"paddh $f30, $f30, $f6 \n\t"
|
|
981
|
+
"gslqc1 $f6, $f4, 640-544(%[tmp]) \n\t"
|
|
982
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
983
|
+
"gssqc1 $f22, $f20, 640-352(%[tmp]) \n\t"
|
|
984
|
+
"gslqc1 $f22, $f20, 640-368(%[tmp]) \n\t"
|
|
985
|
+
"psllh $f28, $f28, $f10 \n\t"
|
|
986
|
+
"psllh $f30, $f30, $f10 \n\t"
|
|
987
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
988
|
+
"paddh $f28, $f28, $f24 \n\t"
|
|
989
|
+
"paddh $f30, $f30, $f26 \n\t"
|
|
990
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
991
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
992
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
993
|
+
|
|
994
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
995
|
+
"gslqc1 $f30, $f28, 640-400(%[tmp]) \n\t"
|
|
996
|
+
"psrah $f20, $f20, $f10 \n\t"
|
|
997
|
+
"psrah $f22, $f22, $f10 \n\t"
|
|
998
|
+
"and $f4, $f4, $f20 \n\t"
|
|
999
|
+
"and $f6, $f6, $f22 \n\t"
|
|
1000
|
+
"gslqc1 $f22, $f20, 640-480(%[tmp]) \n\t"
|
|
1001
|
+
"gssqc1 $f6, $f4, 640-96(%[tmp]) \n\t"
|
|
1002
|
+
"gslqc1 $f6, $f4, 640-384(%[tmp]) \n\t"
|
|
1003
|
+
"gslqc1 $f10, $f8, 640-400(%[tmp]) \n\t"
|
|
1004
|
+
"paddh $f24, $f4, $f4 \n\t"
|
|
1005
|
+
"paddh $f26, $f6, $f6 \n\t"
|
|
1006
|
+
"paddh $f4, $f4, $f8 \n\t"
|
|
1007
|
+
"paddh $f6, $f6, $f10 \n\t"
|
|
1008
|
+
"gslqc1 $f10, $f8, 640-144(%[tmp]) \n\t"
|
|
1009
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
1010
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
1011
|
+
"paddh $f4, $f4, $f8 \n\t"
|
|
1012
|
+
"paddh $f6, $f6, $f10 \n\t"
|
|
1013
|
+
"gslqc1 $f10, $f8, 640-592(%[tmp]) \n\t"
|
|
1014
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
1015
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
1016
|
+
"paddh $f20, $f20, $f8 \n\t"
|
|
1017
|
+
"paddh $f22, $f22, $f10 \n\t"
|
|
1018
|
+
"gslqc1 $f10, $f8, 640-624(%[tmp]) \n\t"
|
|
1019
|
+
"paddh $f24, $f24, $f8 \n\t"
|
|
1020
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
1021
|
+
"paddh $f26, $f26, $f10 \n\t"
|
|
1022
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
1023
|
+
"gslqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
1024
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
1025
|
+
"psrah $f24, $f24, $f8 \n\t"
|
|
1026
|
+
"psrah $f26, $f26, $f8 \n\t"
|
|
1027
|
+
"psllh $f4, $f4, $f10 \n\t"
|
|
1028
|
+
"psllh $f6, $f6, $f10 \n\t"
|
|
1029
|
+
"paddh $f4, $f4, $f20 \n\t"
|
|
1030
|
+
"paddh $f6, $f6, $f22 \n\t"
|
|
1031
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
1032
|
+
|
|
1033
|
+
"gslqc1 $f22, $f20, 656-272(%[tmp]) \n\t"
|
|
1034
|
+
"pandn $f28, $f28, $f24 \n\t"
|
|
1035
|
+
"pandn $f30, $f30, $f26 \n\t"
|
|
1036
|
+
"gslqc1 $f26, $f24, 640-416(%[tmp]) \n\t"
|
|
1037
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
1038
|
+
"paddh $f24, $f24, $f4 \n\t"
|
|
1039
|
+
"paddh $f26, $f26, $f6 \n\t"
|
|
1040
|
+
"gslqc1 $f6, $f4, 640-560(%[tmp]) \n\t"
|
|
1041
|
+
"psrah $f24, $f24, $f10 \n\t"
|
|
1042
|
+
"psrah $f26, $f26, $f10 \n\t"
|
|
1043
|
+
"and $f4, $f4, $f24 \n\t"
|
|
1044
|
+
"and $f6, $f6, $f26 \n\t"
|
|
1045
|
+
|
|
1046
|
+
"xor $f8, $f8, $f8 \n\t"
|
|
1047
|
+
"gslqc1 $f26, $f24, 704-272(%[tmp]) \n\t"
|
|
1048
|
+
"gssqc1 $f6, $f4, 640-128(%[tmp]) \n\t"
|
|
1049
|
+
"gslqc1 $f6, $f4, 672-272(%[tmp]) \n\t"
|
|
1050
|
+
"punpcklbh $f4, $f6, $f8 \n\t"
|
|
1051
|
+
"punpckhbh $f6, $f6, $f8 \n\t"
|
|
1052
|
+
"gssqc1 $f6, $f4, 640-448(%[tmp]) \n\t"
|
|
1053
|
+
"gslqc1 $f6, $f4, 688-272(%[tmp]) \n\t"
|
|
1054
|
+
"punpcklbh $f4, $f6, $f8 \n\t"
|
|
1055
|
+
"punpckhbh $f6, $f6, $f8 \n\t"
|
|
1056
|
+
"punpcklbh $f24, $f26, $f8 \n\t"
|
|
1057
|
+
"punpckhbh $f26, $f26, $f8 \n\t"
|
|
1058
|
+
"gssqc1 $f30, $f28, 640-288(%[tmp]) \n\t"
|
|
1059
|
+
"punpcklbh $f20, $f22, $f8 \n\t"
|
|
1060
|
+
"punpckhbh $f22, $f22, $f8 \n\t"
|
|
1061
|
+
"gslqc1 $f30, $f28, 0x0($14) \n\t"
|
|
1062
|
+
"gssqc1 $f6, $f4, 640-496(%[tmp]) \n\t"
|
|
1063
|
+
"gssqc1 $f26, $f24, 640-432(%[tmp]) \n\t"
|
|
1064
|
+
|
|
1065
|
+
"gsldxc1 $f0, 0x8($15, $0) \n\t"
|
|
1066
|
+
"punpcklbh $f28, $f30, $f8 \n\t"
|
|
1067
|
+
"punpckhbh $f30, $f30, $f8 \n\t"
|
|
1068
|
+
"gssqc1 $f30, $f28, 640-464(%[tmp]) \n\t"
|
|
1069
|
+
|
|
1070
|
+
"punpcklbh $f28, $f0, $f8 \n\t"
|
|
1071
|
+
"punpckhbh $f30, $f0, $f8 \n\t"
|
|
1072
|
+
"gslqc1 $f10, $f8, 640-464(%[tmp]) \n\t"
|
|
1073
|
+
"gssqc1 $f30, $f28, 640-528(%[tmp]) \n\t"
|
|
1074
|
+
|
|
1075
|
+
"psubh $f28, $f24, $f4 \n\t"
|
|
1076
|
+
"psubh $f30, $f26, $f6 \n\t"
|
|
1077
|
+
"psubh $f24, $f24, $f8 \n\t"
|
|
1078
|
+
"psubh $f26, $f26, $f10 \n\t"
|
|
1079
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
1080
|
+
"gslqc1 $f10, $f8, 640-16(%[tmp]) \n\t"
|
|
1081
|
+
"gssqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
1082
|
+
"or $f16, $f16, $f8 \n\t"
|
|
1083
|
+
"or $f18, $f18, $f10 \n\t"
|
|
1084
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
1085
|
+
"gslqc1 $f30, $f28, 640-448(%[tmp]) \n\t"
|
|
1086
|
+
"psubh $f28, $f4, $f28 \n\t"
|
|
1087
|
+
"psubh $f30, $f6, $f30 \n\t"
|
|
1088
|
+
|
|
1089
|
+
"gslqc1 $f2, $f0, 640-512(%[tmp]) \n\t"
|
|
1090
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
1091
|
+
"pcmpgth $f4, $f0, $f28 \n\t"
|
|
1092
|
+
"pcmpgth $f6, $f2, $f30 \n\t"
|
|
1093
|
+
"pcmpgth $f28, $f0, $f24 \n\t"
|
|
1094
|
+
"pcmpgth $f30, $f2, $f26 \n\t"
|
|
1095
|
+
"gslqc1 $f26, $f24, 640-320(%[tmp]) \n\t"
|
|
1096
|
+
"and $f4, $f4, $f28 \n\t"
|
|
1097
|
+
"and $f6, $f6, $f30 \n\t"
|
|
1098
|
+
"gslqc1 $f30, $f28, 640-560(%[tmp]) \n\t"
|
|
1099
|
+
"pcmpgth $f24, $f24, $f28 \n\t"
|
|
1100
|
+
"pcmpgth $f26, $f26, $f30 \n\t"
|
|
1101
|
+
"and $f4, $f4, $f24 \n\t"
|
|
1102
|
+
"and $f6, $f6, $f26 \n\t"
|
|
1103
|
+
|
|
1104
|
+
"gslqc1 $f26, $f24, 640-576(%[tmp]) \n\t"
|
|
1105
|
+
"pcmpgth $f24, $f24, $f28 \n\t"
|
|
1106
|
+
"pcmpgth $f26, $f26, $f30 \n\t"
|
|
1107
|
+
"xor $f8, $f8, $f8 \n\t"
|
|
1108
|
+
"gslqc1 $f30, $f28, 640-496(%[tmp]) \n\t"
|
|
1109
|
+
"punpcklbh $f12, $f14, $f8 \n\t"
|
|
1110
|
+
"punpckhbh $f14, $f14, $f8 \n\t"
|
|
1111
|
+
"gssqc1 $f26, $f24, 640-560(%[tmp]) \n\t"
|
|
1112
|
+
"gslqc1 $f26, $f24, 640-512(%[tmp]) \n\t"
|
|
1113
|
+
"psubh $f28, $f28, $f20 \n\t"
|
|
1114
|
+
"psubh $f30, $f30, $f22 \n\t"
|
|
1115
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
1116
|
+
"pcmpgth $f24, $f24, $f28 \n\t"
|
|
1117
|
+
"pcmpgth $f26, $f26, $f30 \n\t"
|
|
1118
|
+
|
|
1119
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
1120
|
+
"gslqc1 $f10, $f8, 640-560(%[tmp]) \n\t"
|
|
1121
|
+
"and $f24, $f24, $f8 \n\t"
|
|
1122
|
+
"and $f26, $f26, $f10 \n\t"
|
|
1123
|
+
"gslqc1 $f30, $f28, 640-432(%[tmp]) \n\t"
|
|
1124
|
+
"gslqc1 $f10, $f8, 640-528(%[tmp]) \n\t"
|
|
1125
|
+
"psubh $f28, $f28, $f8 \n\t"
|
|
1126
|
+
"psubh $f30, $f30, $f10 \n\t"
|
|
1127
|
+
"dmtc1 %[iAlpha], $f10 \n\t"
|
|
1128
|
+
|
|
1129
|
+
"psllh $f12, $f12, $f10 \n\t"
|
|
1130
|
+
"psllh $f14, $f14, $f10 \n\t"
|
|
1131
|
+
"gssqc1 $f26, $f24, 640-544(%[tmp]) \n\t"
|
|
1132
|
+
"gslqc1 $f26, $f24, 640-512(%[tmp]) \n\t"
|
|
1133
|
+
|
|
1134
|
+
"gslqc1 $f10, $f8, 640-448(%[tmp]) \n\t"
|
|
1135
|
+
"paddh $f12, $f12, $f20 \n\t"
|
|
1136
|
+
"paddh $f14, $f14, $f22 \n\t"
|
|
1137
|
+
"paddh $f12, $f12, $f20 \n\t"
|
|
1138
|
+
"paddh $f14, $f14, $f22 \n\t"
|
|
1139
|
+
"paddh $f12, $f12, $f20 \n\t"
|
|
1140
|
+
"paddh $f14, $f14, $f22 \n\t"
|
|
1141
|
+
"paddh $f12, $f12, $f8 \n\t"
|
|
1142
|
+
"paddh $f14, $f14, $f10 \n\t"
|
|
1143
|
+
"gslqc1 $f10, $f8, 640-496(%[tmp]) \n\t"
|
|
1144
|
+
"gslqc1 $f2, $f0, 640-560(%[tmp]) \n\t"
|
|
1145
|
+
"paddh $f12, $f12, $f8 \n\t"
|
|
1146
|
+
"paddh $f14, $f14, $f10 \n\t"
|
|
1147
|
+
WELS_AbsH($f28, $f30, $f28, $f30, $f8, $f10)
|
|
1148
|
+
"pcmpgth $f24, $f24, $f28 \n\t"
|
|
1149
|
+
"pcmpgth $f26, $f26, $f30 \n\t"
|
|
1150
|
+
"and $f24, $f24, $f0 \n\t"
|
|
1151
|
+
"and $f26, $f26, $f2 \n\t"
|
|
1152
|
+
"gssqc1 $f26, $f24, 640-560(%[tmp]) \n\t"
|
|
1153
|
+
"gslqc1 $f10, $f8, 640-544(%[tmp]) \n\t"
|
|
1154
|
+
|
|
1155
|
+
"gslqc1 $f2, $f0, 736-272(%[tmp]) \n\t"
|
|
1156
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
1157
|
+
"gslqc1 $f30, $f28, 640-368(%[tmp]) \n\t"
|
|
1158
|
+
"and $f24, $f0, $f16 \n\t"
|
|
1159
|
+
"and $f26, $f2, $f18 \n\t"
|
|
1160
|
+
"pandn $f16, $f0, $f28 \n\t"
|
|
1161
|
+
"pandn $f18, $f2, $f30 \n\t"
|
|
1162
|
+
"or $f24, $f24, $f16 \n\t"
|
|
1163
|
+
"or $f26, $f26, $f18 \n\t"
|
|
1164
|
+
"gslqc1 $f18, $f16, 640-432(%[tmp]) \n\t"
|
|
1165
|
+
"paddh $f12, $f12, $f16 \n\t"
|
|
1166
|
+
"paddh $f14, $f14, $f18 \n\t"
|
|
1167
|
+
"gslqc1 $f30, $f28, 640-592(%[tmp]) \n\t"
|
|
1168
|
+
"paddh $f12, $f12, $f28 \n\t"
|
|
1169
|
+
"paddh $f14, $f14, $f30 \n\t"
|
|
1170
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1171
|
+
"psrah $f12, $f12, $f28 \n\t"
|
|
1172
|
+
"psrah $f14, $f14, $f28 \n\t"
|
|
1173
|
+
"and $f12, $f12, $f8 \n\t"
|
|
1174
|
+
"and $f14, $f14, $f10 \n\t"
|
|
1175
|
+
"pandn $f8, $f8, $f20 \n\t"
|
|
1176
|
+
"pandn $f10, $f10, $f22 \n\t"
|
|
1177
|
+
"or $f12, $f12, $f8 \n\t"
|
|
1178
|
+
"or $f14, $f14, $f10 \n\t"
|
|
1179
|
+
"and $f28, $f4, $f12 \n\t"
|
|
1180
|
+
"and $f30, $f6, $f14 \n\t"
|
|
1181
|
+
"gslqc1 $f14, $f12, 640-64(%[tmp]) \n\t"
|
|
1182
|
+
"gslqc1 $f10, $f8, 640-336(%[tmp]) \n\t"
|
|
1183
|
+
"or $f12, $f12, $f8 \n\t"
|
|
1184
|
+
"or $f14, $f14, $f10 \n\t"
|
|
1185
|
+
"pandn $f8, $f4, $f20 \n\t"
|
|
1186
|
+
"pandn $f10, $f6, $f22 \n\t"
|
|
1187
|
+
"or $f28, $f28, $f8 \n\t"
|
|
1188
|
+
"or $f30, $f30, $f10 \n\t"
|
|
1189
|
+
|
|
1190
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
1191
|
+
"and $f8, $f0, $f12 \n\t"
|
|
1192
|
+
"and $f10, $f2, $f14 \n\t"
|
|
1193
|
+
"gslqc1 $f14, $f12, 640-480(%[tmp]) \n\t"
|
|
1194
|
+
"pandn $f12, $f0, $f12 \n\t"
|
|
1195
|
+
"pandn $f14, $f2, $f14 \n\t"
|
|
1196
|
+
"or $f8, $f8, $f12 \n\t"
|
|
1197
|
+
"or $f10, $f10, $f14 \n\t"
|
|
1198
|
+
"packushb $f24, $f24, $f26 \n\t"
|
|
1199
|
+
"packushb $f26, $f28, $f30 \n\t"
|
|
1200
|
+
"gssqc1 $f10, $f8, 640-336(%[tmp]) \n\t"
|
|
1201
|
+
"gssqc1 $f26, $f24, 656-272(%[tmp]) \n\t"
|
|
1202
|
+
"gslqc1 $f26, $f24, 640-544(%[tmp]) \n\t"
|
|
1203
|
+
"gslqc1 $f10, $f8, 640-448(%[tmp]) \n\t"
|
|
1204
|
+
"paddh $f8, $f20, $f8 \n\t"
|
|
1205
|
+
"paddh $f10, $f22, $f10 \n\t"
|
|
1206
|
+
"gslqc1 $f30, $f28, 640-496(%[tmp]) \n\t"
|
|
1207
|
+
"paddh $f28, $f28, $f16 \n\t"
|
|
1208
|
+
"paddh $f30, $f30, $f18 \n\t"
|
|
1209
|
+
"paddh $f8, $f8, $f28 \n\t"
|
|
1210
|
+
"paddh $f10, $f10, $f30 \n\t"
|
|
1211
|
+
"gslqc1 $f30, $f28, 640-624(%[tmp]) \n\t"
|
|
1212
|
+
"paddh $f8, $f8, $f28 \n\t"
|
|
1213
|
+
"paddh $f10, $f10, $f30 \n\t"
|
|
1214
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1215
|
+
"psrah $f8, $f8, $f28 \n\t"
|
|
1216
|
+
"psrah $f10, $f10, $f28 \n\t"
|
|
1217
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
1218
|
+
"gslqc1 $f30, $f28, 640-544(%[tmp]) \n\t"
|
|
1219
|
+
"and $f24, $f24, $f8 \n\t"
|
|
1220
|
+
"and $f26, $f26, $f10 \n\t"
|
|
1221
|
+
"gslqc1 $f10, $f8, 640-448(%[tmp]) \n\t"
|
|
1222
|
+
"pandn $f28, $f28, $f8 \n\t"
|
|
1223
|
+
"pandn $f30, $f30, $f10 \n\t"
|
|
1224
|
+
"or $f24, $f24, $f28 \n\t"
|
|
1225
|
+
"or $f26, $f26, $f30 \n\t"
|
|
1226
|
+
"and $f12, $f4, $f24 \n\t"
|
|
1227
|
+
"and $f14, $f6, $f26 \n\t"
|
|
1228
|
+
"pandn $f24, $f4, $f8 \n\t"
|
|
1229
|
+
"pandn $f26, $f6, $f10 \n\t"
|
|
1230
|
+
"gslqc1 $f30, $f28, 640-496(%[tmp]) \n\t"
|
|
1231
|
+
"paddh $f8, $f8, $f28 \n\t"
|
|
1232
|
+
"paddh $f10, $f10, $f30 \n\t"
|
|
1233
|
+
"paddh $f8, $f8, $f16 \n\t"
|
|
1234
|
+
"paddh $f10, $f10, $f18 \n\t"
|
|
1235
|
+
"or $f12, $f12, $f24 \n\t"
|
|
1236
|
+
"or $f14, $f14, $f26 \n\t"
|
|
1237
|
+
"gslqc1 $f26, $f24, 640-336(%[tmp]) \n\t"
|
|
1238
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1239
|
+
"packushb $f24, $f24, $f26 \n\t"
|
|
1240
|
+
"packushb $f26, $f12, $f14 \n\t"
|
|
1241
|
+
"psllh $f8, $f8, $f28 \n\t"
|
|
1242
|
+
"psllh $f10, $f10, $f28 \n\t"
|
|
1243
|
+
"gssqc1 $f26, $f24, 672-272(%[tmp]) \n\t"
|
|
1244
|
+
"gslqc1 $f26, $f24, 640-96(%[tmp]) \n\t"
|
|
1245
|
+
"gslqc1 $f30, $f28, 640-352(%[tmp]) \n\t"
|
|
1246
|
+
"or $f24, $f24, $f28 \n\t"
|
|
1247
|
+
"or $f26, $f26, $f30 \n\t"
|
|
1248
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
1249
|
+
|
|
1250
|
+
"and $f12, $f0, $f24 \n\t"
|
|
1251
|
+
"and $f14, $f2, $f26 \n\t"
|
|
1252
|
+
"gslqc1 $f26, $f24, 640-144(%[tmp]) \n\t"
|
|
1253
|
+
"pandn $f24, $f0, $f24 \n\t"
|
|
1254
|
+
"pandn $f26, $f2, $f26 \n\t"
|
|
1255
|
+
"or $f12, $f12, $f24 \n\t"
|
|
1256
|
+
"or $f14, $f14, $f26 \n\t"
|
|
1257
|
+
"gslqc1 $f26, $f24, 640-544(%[tmp]) \n\t"
|
|
1258
|
+
"gssqc1 $f14, $f12, 640-352(%[tmp]) \n\t"
|
|
1259
|
+
"gslqc1 $f14, $f12, 640-464(%[tmp]) \n\t"
|
|
1260
|
+
"gslqc1 $f30, $f28, 640-592(%[tmp]) \n\t"
|
|
1261
|
+
"paddh $f12, $f12, $f28 \n\t"
|
|
1262
|
+
"paddh $f14, $f14, $f30 \n\t"
|
|
1263
|
+
"paddh $f8, $f8, $f12 \n\t"
|
|
1264
|
+
"paddh $f10, $f10, $f14 \n\t"
|
|
1265
|
+
"gslqc1 $f14, $f12, 640-448(%[tmp]) \n\t"
|
|
1266
|
+
"paddh $f20, $f20, $f8 \n\t"
|
|
1267
|
+
"paddh $f22, $f22, $f10 \n\t"
|
|
1268
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1269
|
+
"gslqc1 $f10, $f8, 640-496(%[tmp]) \n\t"
|
|
1270
|
+
"psrah $f20, $f20, $f28 \n\t"
|
|
1271
|
+
"psrah $f22, $f22, $f28 \n\t"
|
|
1272
|
+
"and $f24, $f24, $f20 \n\t"
|
|
1273
|
+
"and $f26, $f26, $f22 \n\t"
|
|
1274
|
+
"gslqc1 $f22, $f20, 640-464(%[tmp]) \n\t"
|
|
1275
|
+
"paddh $f8, $f8, $f20 \n\t"
|
|
1276
|
+
"paddh $f10, $f10, $f22 \n\t"
|
|
1277
|
+
"gslqc1 $f30, $f28, 640-432(%[tmp]) \n\t"
|
|
1278
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
1279
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
1280
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
1281
|
+
"paddh $f16, $f12, $f12 \n\t"
|
|
1282
|
+
"paddh $f18, $f14, $f14 \n\t"
|
|
1283
|
+
"paddh $f16, $f16, $f8 \n\t"
|
|
1284
|
+
"paddh $f18, $f18, $f10 \n\t"
|
|
1285
|
+
"gslqc1 $f30, $f28, 640-624(%[tmp]) \n\t"
|
|
1286
|
+
"paddh $f16, $f16, $f28 \n\t"
|
|
1287
|
+
"paddh $f18, $f18, $f30 \n\t"
|
|
1288
|
+
"gslqc1 $f10, $f8, 640-544(%[tmp]) \n\t"
|
|
1289
|
+
"gslqc1 $f30, $f28, 640-592(%[tmp]) \n\t"
|
|
1290
|
+
"paddh $f12, $f12, $f28 \n\t"
|
|
1291
|
+
"paddh $f14, $f14, $f30 \n\t"
|
|
1292
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1293
|
+
"psrah $f16, $f16, $f28 \n\t"
|
|
1294
|
+
"psrah $f18, $f18, $f28 \n\t"
|
|
1295
|
+
"pandn $f8, $f8, $f16 \n\t"
|
|
1296
|
+
"pandn $f10, $f10, $f18 \n\t"
|
|
1297
|
+
"or $f24, $f24, $f8 \n\t"
|
|
1298
|
+
"or $f26, $f26, $f10 \n\t"
|
|
1299
|
+
"and $f28, $f4, $f24 \n\t"
|
|
1300
|
+
"and $f30, $f6, $f26 \n\t"
|
|
1301
|
+
"gslqc1 $f26, $f24, 640-496(%[tmp]) \n\t"
|
|
1302
|
+
"pandn $f8, $f4, $f24 \n\t"
|
|
1303
|
+
"pandn $f10, $f6, $f26 \n\t"
|
|
1304
|
+
"or $f28, $f28, $f8 \n\t"
|
|
1305
|
+
"or $f30, $f30, $f10 \n\t"
|
|
1306
|
+
"gslqc1 $f10, $f8, 640-352(%[tmp]) \n\t"
|
|
1307
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
1308
|
+
"packushb $f10, $f28, $f30 \n\t"
|
|
1309
|
+
"gssqc1 $f10, $f8, 688-272(%[tmp]) \n\t"
|
|
1310
|
+
"gslqc1 $f10, $f8, 640-128(%[tmp]) \n\t"
|
|
1311
|
+
"gslqc1 $f30, $f28, 640-288(%[tmp]) \n\t"
|
|
1312
|
+
"or $f8, $f8, $f28 \n\t"
|
|
1313
|
+
"or $f10, $f10, $f30 \n\t"
|
|
1314
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
1315
|
+
|
|
1316
|
+
"and $f16, $f0, $f8 \n\t"
|
|
1317
|
+
"and $f18, $f2, $f10 \n\t"
|
|
1318
|
+
"paddh $f20, $f20, $f24 \n\t"
|
|
1319
|
+
"paddh $f22, $f22, $f26 \n\t"
|
|
1320
|
+
"gslqc1 $f30, $f28, 640-400(%[tmp]) \n\t"
|
|
1321
|
+
"pandn $f8, $f0, $f28 \n\t"
|
|
1322
|
+
"pandn $f10, $f2, $f30 \n\t"
|
|
1323
|
+
"or $f16, $f16, $f8 \n\t"
|
|
1324
|
+
"or $f18, $f18, $f10 \n\t"
|
|
1325
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1326
|
+
"gslqc1 $f10, $f8, 640-528(%[tmp]) \n\t"
|
|
1327
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
1328
|
+
"psllh $f20, $f20, $f28 \n\t"
|
|
1329
|
+
"psllh $f22, $f22, $f28 \n\t"
|
|
1330
|
+
"paddh $f20, $f20, $f12 \n\t"
|
|
1331
|
+
"paddh $f22, $f22, $f14 \n\t"
|
|
1332
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1333
|
+
"gslqc1 $f14, $f12, 640-560(%[tmp]) \n\t"
|
|
1334
|
+
"paddh $f8, $f8, $f20 \n\t"
|
|
1335
|
+
"paddh $f10, $f10, $f22 \n\t"
|
|
1336
|
+
"psrah $f8, $f8, $f28 \n\t"
|
|
1337
|
+
"psrah $f10, $f10, $f28 \n\t"
|
|
1338
|
+
"gssqc1 $f18, $f16, 640-288(%[tmp]) \n\t"
|
|
1339
|
+
"gslqc1 $f18, $f16, 640-560(%[tmp]) \n\t"
|
|
1340
|
+
"and $f16, $f16, $f8 \n\t"
|
|
1341
|
+
"and $f18, $f18, $f10 \n\t"
|
|
1342
|
+
"gslqc1 $f10, $f8, 640-464(%[tmp]) \n\t"
|
|
1343
|
+
"paddh $f20, $f8, $f8 \n\t"
|
|
1344
|
+
"paddh $f22, $f10, $f10 \n\t"
|
|
1345
|
+
"gslqc1 $f10, $f8, 640-432(%[tmp]) \n\t"
|
|
1346
|
+
"gslqc1 $f30, $f28, 640-448(%[tmp]) \n\t"
|
|
1347
|
+
"paddh $f8, $f8, $f28 \n\t"
|
|
1348
|
+
"paddh $f10, $f10, $f30 \n\t"
|
|
1349
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
1350
|
+
"paddh $f20, $f20, $f8 \n\t"
|
|
1351
|
+
"paddh $f22, $f22, $f10 \n\t"
|
|
1352
|
+
"gslqc1 $f30, $f28, 640-624(%[tmp]) \n\t"
|
|
1353
|
+
"paddh $f20, $f20, $f28 \n\t"
|
|
1354
|
+
"paddh $f22, $f22, $f30 \n\t"
|
|
1355
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1356
|
+
"gslqc1 $f26, $f24, 640-560(%[tmp]) \n\t"
|
|
1357
|
+
"psrah $f20, $f20, $f28 \n\t"
|
|
1358
|
+
"psrah $f22, $f22, $f28 \n\t"
|
|
1359
|
+
"pandn $f12, $f12, $f20 \n\t"
|
|
1360
|
+
"pandn $f14, $f14, $f22 \n\t"
|
|
1361
|
+
"or $f16, $f16, $f12 \n\t"
|
|
1362
|
+
"or $f18, $f18, $f14 \n\t"
|
|
1363
|
+
"gslqc1 $f14, $f12, 640-32(%[tmp]) \n\t"
|
|
1364
|
+
"gslqc1 $f30, $f28, 640-304(%[tmp]) \n\t"
|
|
1365
|
+
"or $f12, $f12, $f28 \n\t"
|
|
1366
|
+
"or $f14, $f14, $f30 \n\t"
|
|
1367
|
+
"and $f28, $f4, $f16 \n\t"
|
|
1368
|
+
"and $f30, $f6, $f18 \n\t"
|
|
1369
|
+
"gslqc1 $f18, $f16, 640-432(%[tmp]) \n\t"
|
|
1370
|
+
"gslqc1 $f22, $f20, 640-464(%[tmp]) \n\t"
|
|
1371
|
+
"pandn $f8, $f4, $f16 \n\t"
|
|
1372
|
+
"pandn $f10, $f6, $f18 \n\t"
|
|
1373
|
+
"or $f28, $f28, $f8 \n\t"
|
|
1374
|
+
"or $f30, $f30, $f10 \n\t"
|
|
1375
|
+
"gslqc1 $f10, $f8, 640-496(%[tmp]) \n\t"
|
|
1376
|
+
"paddh $f16, $f16, $f8 \n\t"
|
|
1377
|
+
"paddh $f18, $f18, $f10 \n\t"
|
|
1378
|
+
"gslqc1 $f10, $f8, 640-288(%[tmp]) \n\t"
|
|
1379
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
1380
|
+
"packushb $f10, $f28, $f30 \n\t"
|
|
1381
|
+
"dli %[iAlpha], 0x2 \n\t"
|
|
1382
|
+
"gssqc1 $f10, $f8, 704-272(%[tmp]) \n\t"
|
|
1383
|
+
|
|
1384
|
+
"and $f8, $f0, $f12 \n\t"
|
|
1385
|
+
"and $f10, $f2, $f14 \n\t"
|
|
1386
|
+
"gslqc1 $f30, $f28, 640-384(%[tmp]) \n\t"
|
|
1387
|
+
"pandn $f12, $f0, $f28 \n\t"
|
|
1388
|
+
"pandn $f14, $f2, $f30 \n\t"
|
|
1389
|
+
"or $f8, $f8, $f12 \n\t"
|
|
1390
|
+
"or $f10, $f10, $f14 \n\t"
|
|
1391
|
+
"gssqc1 $f10, $f8, 640-304(%[tmp]) \n\t"
|
|
1392
|
+
"gslqc1 $f10, $f8, 640-528(%[tmp]) \n\t"
|
|
1393
|
+
"gslqc1 $f30, $f28, 640-464(%[tmp]) \n\t"
|
|
1394
|
+
"paddh $f12, $f8, $f28 \n\t"
|
|
1395
|
+
"paddh $f14, $f10, $f30 \n\t"
|
|
1396
|
+
"paddh $f12, $f12, $f16 \n\t"
|
|
1397
|
+
"paddh $f14, $f14, $f18 \n\t"
|
|
1398
|
+
"gslqc1 $f30, $f28, 640-624(%[tmp]) \n\t"
|
|
1399
|
+
"paddh $f12, $f12, $f28 \n\t"
|
|
1400
|
+
"paddh $f14, $f14, $f30 \n\t"
|
|
1401
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1402
|
+
"psrah $f12, $f12, $f28 \n\t"
|
|
1403
|
+
"psrah $f14, $f14, $f28 \n\t"
|
|
1404
|
+
"and $f24, $f24, $f12 \n\t"
|
|
1405
|
+
"and $f26, $f26, $f14 \n\t"
|
|
1406
|
+
"gslqc1 $f14, $f12, 640-560(%[tmp]) \n\t"
|
|
1407
|
+
"pandn $f16, $f12, $f20 \n\t"
|
|
1408
|
+
"pandn $f18, $f14, $f22 \n\t"
|
|
1409
|
+
"or $f24, $f24, $f16 \n\t"
|
|
1410
|
+
"or $f26, $f26, $f18 \n\t"
|
|
1411
|
+
"and $f28, $f4, $f24 \n\t"
|
|
1412
|
+
"and $f30, $f6, $f26 \n\t"
|
|
1413
|
+
"gslqc1 $f26, $f24, 640-304(%[tmp]) \n\t"
|
|
1414
|
+
"pandn $f16, $f4, $f20 \n\t"
|
|
1415
|
+
"pandn $f18, $f6, $f22 \n\t"
|
|
1416
|
+
"or $f28, $f28, $f16 \n\t"
|
|
1417
|
+
"or $f30, $f30, $f18 \n\t"
|
|
1418
|
+
"dli %[iAlpha], 0x1 \n\t"
|
|
1419
|
+
|
|
1420
|
+
"packushb $f24, $f24, $f26 \n\t"
|
|
1421
|
+
"packushb $f26, $f28, $f30 \n\t"
|
|
1422
|
+
"gslqc1 $f30, $f28, 640-112(%[tmp]) \n\t"
|
|
1423
|
+
"gslqc1 $f18, $f16, 640-80(%[tmp]) \n\t"
|
|
1424
|
+
"or $f28, $f28, $f16 \n\t"
|
|
1425
|
+
"or $f30, $f30, $f18 \n\t"
|
|
1426
|
+
"and $f16, $f0, $f28 \n\t"
|
|
1427
|
+
"and $f18, $f2, $f30 \n\t"
|
|
1428
|
+
"gslqc1 $f30, $f28, 640-416(%[tmp]) \n\t"
|
|
1429
|
+
"pandn $f0, $f0, $f28 \n\t"
|
|
1430
|
+
"pandn $f2, $f2, $f30 \n\t"
|
|
1431
|
+
"or $f16, $f16, $f0 \n\t"
|
|
1432
|
+
"or $f18, $f18, $f2 \n\t"
|
|
1433
|
+
"xor $f28, $f28, $f28 \n\t"
|
|
1434
|
+
"xor $f30, $f30, $f30 \n\t"
|
|
1435
|
+
"gslqc1 $f2, $f0, 0x0($12) \n\t"
|
|
1436
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1437
|
+
"punpcklbh $f0, $f2, $f30 \n\t"
|
|
1438
|
+
"punpckhbh $f2, $f2, $f30 \n\t"
|
|
1439
|
+
"psllh $f0, $f0, $f28 \n\t"
|
|
1440
|
+
"psllh $f2, $f2, $f28 \n\t"
|
|
1441
|
+
"paddh $f0, $f0, $f8 \n\t"
|
|
1442
|
+
"paddh $f2, $f2, $f10 \n\t"
|
|
1443
|
+
"paddh $f0, $f0, $f8 \n\t"
|
|
1444
|
+
"paddh $f2, $f2, $f10 \n\t"
|
|
1445
|
+
"paddh $f0, $f0, $f8 \n\t"
|
|
1446
|
+
"paddh $f2, $f2, $f10 \n\t"
|
|
1447
|
+
"paddh $f0, $f0, $f20 \n\t"
|
|
1448
|
+
"paddh $f2, $f2, $f22 \n\t"
|
|
1449
|
+
"dli %[iAlpha], 0x3 \n\t"
|
|
1450
|
+
"gslqc1 $f30, $f28, 640-432(%[tmp]) \n\t"
|
|
1451
|
+
"paddh $f0, $f0, $f28 \n\t"
|
|
1452
|
+
"paddh $f2, $f2, $f30 \n\t"
|
|
1453
|
+
"gslqc1 $f30, $f28, 640-496(%[tmp]) \n\t"
|
|
1454
|
+
"paddh $f0, $f0, $f28 \n\t"
|
|
1455
|
+
"paddh $f2, $f2, $f30 \n\t"
|
|
1456
|
+
"gslqc1 $f30, $f28, 640-592(%[tmp]) \n\t"
|
|
1457
|
+
"paddh $f0, $f0, $f28 \n\t"
|
|
1458
|
+
"paddh $f2, $f2, $f30 \n\t"
|
|
1459
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1460
|
+
"psrah $f0, $f0, $f28 \n\t"
|
|
1461
|
+
"psrah $f2, $f2, $f28 \n\t"
|
|
1462
|
+
"and $f0, $f0, $f12 \n\t"
|
|
1463
|
+
"and $f2, $f2, $f14 \n\t"
|
|
1464
|
+
"pandn $f12, $f12, $f8 \n\t"
|
|
1465
|
+
"pandn $f14, $f14, $f10 \n\t"
|
|
1466
|
+
"or $f0, $f0, $f12 \n\t"
|
|
1467
|
+
"or $f2, $f2, $f14 \n\t"
|
|
1468
|
+
"and $f28, $f4, $f0 \n\t"
|
|
1469
|
+
"and $f30, $f6, $f2 \n\t"
|
|
1470
|
+
|
|
1471
|
+
"gslqc1 $f2, $f0, 656-272(%[tmp]) \n\t"
|
|
1472
|
+
"gssqc1 $f2, $f0, 0x0($11) \n\t"
|
|
1473
|
+
|
|
1474
|
+
"gslqc1 $f2, $f0, 672-272(%[tmp]) \n\t"
|
|
1475
|
+
|
|
1476
|
+
"gssqc1 $f2, $f0, 0x0($8) \n\t"
|
|
1477
|
+
"gslqc1 $f2, $f0, 688-272(%[tmp]) \n\t"
|
|
1478
|
+
"gssqc1 $f2, $f0, 0x0($9) \n\t"
|
|
1479
|
+
"gslqc1 $f2, $f0, 704-272(%[tmp]) \n\t"
|
|
1480
|
+
|
|
1481
|
+
"pandn $f4, $f4, $f8 \n\t"
|
|
1482
|
+
"pandn $f6, $f6, $f10 \n\t"
|
|
1483
|
+
"gssqc1 $f2, $f0, 0x0(%[pPix]) \n\t"
|
|
1484
|
+
"or $f28, $f28, $f4 \n\t"
|
|
1485
|
+
"or $f30, $f30, $f6 \n\t"
|
|
1486
|
+
"packushb $f16, $f16, $f18 \n\t"
|
|
1487
|
+
"packushb $f18, $f28, $f30 \n\t"
|
|
1488
|
+
"gssqc1 $f26, $f24, 0x0($13) \n\t"
|
|
1489
|
+
"gssqc1 $f18, $f16, 0x0(%[iStride]) \n\t"
|
|
1490
|
+
: [pPix]"+&r"((unsigned char *)pPix)
|
|
1491
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha),
|
|
1492
|
+
[iBeta]"r"(iBeta), [tmp]"r"((unsigned char *)tmp)
|
|
1493
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$13", "$14", "$15", "$f0",
|
|
1494
|
+
"$f2", "$f4", "$f6", "$f8", "$f10", "$f12", "$f14", "$f16", "$f18", "$f20",
|
|
1495
|
+
"$f22", "$f24", "$f26", "$f28", "$f30"
|
|
1496
|
+
);
|
|
1497
|
+
RECOVER_REG;
|
|
1498
|
+
}
|
|
1499
|
+
|
|
1500
|
+
void DeblockChromaLt4V_mmi(uint8_t *pPixCb, uint8_t *pPixCr, int32_t iStride,
|
|
1501
|
+
int32_t iAlpha, int32_t iBeta, int8_t *pTC) {
|
|
1502
|
+
unsigned char tmp[256] __attribute__((aligned(32)));
|
|
1503
|
+
BACKUP_REG;
|
|
1504
|
+
__asm__ volatile (
|
|
1505
|
+
".set arch=loongson3a \n\t"
|
|
1506
|
+
"lb $8, 0x2(%[pTC]) \n\t"
|
|
1507
|
+
"lb $9, 0x3(%[pTC]) \n\t"
|
|
1508
|
+
"move $11, $8 \n\t"
|
|
1509
|
+
"lb $8, 0x1(%[pTC]) \n\t"
|
|
1510
|
+
"lb %[pTC], 0x0(%[pTC]) \n\t"
|
|
1511
|
+
"move $12, %[pTC] \n\t"
|
|
1512
|
+
"and %[pTC], $9, 0xFFFF \n\t"
|
|
1513
|
+
"dmtc1 %[pTC], $f4 \n\t"
|
|
1514
|
+
"and %[pTC], $9, 0xFFFF \n\t"
|
|
1515
|
+
"dmtc1 %[pTC], $f8 \n\t"
|
|
1516
|
+
"move %[pTC], $11 \n\t"
|
|
1517
|
+
"and $9, %[pTC], 0xFFFF \n\t"
|
|
1518
|
+
"and %[pTC], %[pTC], 0xFFFF \n\t"
|
|
1519
|
+
"dmtc1 %[pTC], $f16 \n\t"
|
|
1520
|
+
"and %[pTC], $8, 0xFFFF \n\t"
|
|
1521
|
+
"dmtc1 %[pTC], $f20 \n\t"
|
|
1522
|
+
"dmtc1 $9, $f12 \n\t"
|
|
1523
|
+
"and %[pTC], $8, 0xFFFF \n\t"
|
|
1524
|
+
"dmtc1 %[pTC], $f24 \n\t"
|
|
1525
|
+
"move %[pTC], $12 \n\t"
|
|
1526
|
+
"and $9, %[pTC], 0xFFFF \n\t"
|
|
1527
|
+
"and %[pTC], %[pTC], 0xFFFF \n\t"
|
|
1528
|
+
"punpcklhw $f24, $f24, $f8 \n\t"
|
|
1529
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
1530
|
+
"xor $f2, $f2, $f2 \n\t"
|
|
1531
|
+
"gssqc1 $f2, $f0, 0x40(%[tmp]) \n\t"
|
|
1532
|
+
"dmtc1 $9, $f28 \n\t"
|
|
1533
|
+
"dmtc1 %[pTC], $f0 \n\t"
|
|
1534
|
+
"daddu %[pTC], %[iStride], %[iStride] \n\t"
|
|
1535
|
+
"dsubu $9, %[pPixCb], %[pTC] \n\t"
|
|
1536
|
+
"punpcklhw $f20, $f20, $f4 \n\t"
|
|
1537
|
+
"gslqc1 $f6, $f4, 0x40(%[tmp]) \n\t"
|
|
1538
|
+
"punpcklhw $f0, $f0, $f16 \n\t"
|
|
1539
|
+
"gsldxc1 $f16, 0x0(%[iStride], %[pPixCr]) \n\t"
|
|
1540
|
+
"punpcklhw $f28, $f28, $f12 \n\t"
|
|
1541
|
+
"gsldxc1 $f12, 0x0(%[pPixCb], $0) \n\t"
|
|
1542
|
+
"punpcklhw $f0, $f0, $f24 \n\t"
|
|
1543
|
+
"gsldxc1 $f24, 0x0($9, $0) \n\t"
|
|
1544
|
+
"punpcklhw $f28, $f28, $f20 \n\t"
|
|
1545
|
+
"punpckhhw $f2, $f0, $f28 \n\t"
|
|
1546
|
+
"punpcklhw $f0, $f0, $f28 \n\t"
|
|
1547
|
+
"dsubu $9, %[pPixCr], %[pTC] \n\t"
|
|
1548
|
+
"psubh $f8, $f4, $f0 \n\t"
|
|
1549
|
+
"psubh $f10, $f6, $f2 \n\t"
|
|
1550
|
+
"gssqc1 $f10, $f8, 0x60(%[tmp]) \n\t"
|
|
1551
|
+
"gsldxc1 $f8, 0x0($9, $0) \n\t"
|
|
1552
|
+
"mov.d $f26, $f8 \n\t"
|
|
1553
|
+
"dsubu %[pTC], %[pPixCb], %[iStride] \n\t"
|
|
1554
|
+
"gsldxc1 $f28, 0x0(%[pTC], $0) \n\t"
|
|
1555
|
+
"dsubu $9, %[pPixCr], %[iStride] \n\t"
|
|
1556
|
+
"gsldxc1 $f8, 0x0($9, $0) \n\t"
|
|
1557
|
+
"mov.d $f30, $f8 \n\t"
|
|
1558
|
+
"gsldxc1 $f8, 0x0(%[pPixCr], $0) \n\t"
|
|
1559
|
+
"mov.d $f14, $f8 \n\t"
|
|
1560
|
+
"gsldxc1 $f8, 0x0(%[iStride], %[pPixCb]) \n\t"
|
|
1561
|
+
"mov.d $f10, $f16 \n\t"
|
|
1562
|
+
"gssqc1 $f10, $f8, 0xE0(%[tmp]) \n\t"
|
|
1563
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
1564
|
+
"punpcklhw $f16, $f8, $f8 \n\t"
|
|
1565
|
+
"dmtc1 %[iBeta], $f8 \n\t"
|
|
1566
|
+
"punpcklhw $f20, $f8, $f8 \n\t"
|
|
1567
|
+
"punpcklwd $f8, $f20, $f20 \n\t"
|
|
1568
|
+
"mov.d $f10, $f8 \n\t"
|
|
1569
|
+
"gssqc1 $f10, $f8, 0x50(%[tmp]) \n\t"
|
|
1570
|
+
"punpckhbh $f10, $f24, $f4 \n\t"
|
|
1571
|
+
"punpcklbh $f8, $f24, $f4 \n\t"
|
|
1572
|
+
"gssqc1 $f14, $f12, 0xd0(%[tmp]) \n\t"
|
|
1573
|
+
"punpcklwd $f16, $f16, $f16 \n\t"
|
|
1574
|
+
"mov.d $f18, $f16 \n\t"
|
|
1575
|
+
"gssqc1 $f10, $f8, 0x30(%[tmp]) \n\t"
|
|
1576
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
1577
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
1578
|
+
"gssqc1 $f26, $f24, 0x80(%[tmp]) \n\t"
|
|
1579
|
+
"gslqc1 $f26, $f24, 0xd0(%[tmp]) \n\t"
|
|
1580
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
1581
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
1582
|
+
"gssqc1 $f26, $f24, 0x70(%[tmp]) \n\t"
|
|
1583
|
+
"gslqc1 $f26, $f24, 0xe0(%[tmp]) \n\t"
|
|
1584
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
1585
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
1586
|
+
"gssqc1 $f26, $f24, 0x90(%[tmp]) \n\t"
|
|
1587
|
+
"gslqc1 $f22, $f20, 0xe0(%[tmp]) \n\t"
|
|
1588
|
+
"mov.d $f8, $f28 \n\t"
|
|
1589
|
+
"mov.d $f10, $f30 \n\t"
|
|
1590
|
+
"punpcklbh $f28, $f30, $f6 \n\t"
|
|
1591
|
+
"punpckhbh $f30, $f30, $f6 \n\t"
|
|
1592
|
+
"punpckhbh $f22, $f20, $f4 \n\t"
|
|
1593
|
+
"punpcklbh $f20, $f20, $f4 \n\t"
|
|
1594
|
+
"gssqc1 $f30, $f28, 0xa0(%[tmp]) \n\t"
|
|
1595
|
+
"punpckhbh $f14, $f12, $f4 \n\t"
|
|
1596
|
+
"punpcklbh $f12, $f12, $f4 \n\t"
|
|
1597
|
+
"dli %[iBeta], 0x4 \n\t"
|
|
1598
|
+
"punpckhbh $f10, $f8, $f4 \n\t"
|
|
1599
|
+
"punpcklbh $f8, $f8, $f4 \n\t"
|
|
1600
|
+
"dmtc1 %[iBeta], $f24 \n\t"
|
|
1601
|
+
"punpcklhw $f28, $f24, $f24 \n\t"
|
|
1602
|
+
"punpcklwd $f24, $f28, $f28 \n\t"
|
|
1603
|
+
"mov.d $f26, $f24 \n\t"
|
|
1604
|
+
"gslqc1 $f30, $f28, 0x30(%[tmp]) \n\t"
|
|
1605
|
+
"gssqc1 $f26, $f24, 0x20(%[tmp]) \n\t"
|
|
1606
|
+
"psubh $f28, $f28, $f20 \n\t"
|
|
1607
|
+
"psubh $f30, $f30, $f22 \n\t"
|
|
1608
|
+
"pcmpgth $f24, $f0, $f4 \n\t"
|
|
1609
|
+
"pcmpgth $f26, $f2, $f6 \n\t"
|
|
1610
|
+
"gslqc1 $f6, $f4, 0x60(%[tmp]) \n\t"
|
|
1611
|
+
"gssqc1 $f26, $f24, 0x40(%[tmp]) \n\t"
|
|
1612
|
+
"psubh $f24, $f12, $f8 \n\t"
|
|
1613
|
+
"psubh $f26, $f14, $f10 \n\t"
|
|
1614
|
+
"dmfc1 %[iAlpha], $f12 \n\t"
|
|
1615
|
+
"dmfc1 %[iBeta], $f14 \n\t"
|
|
1616
|
+
"dli $10, 0x2 \n\t"
|
|
1617
|
+
"dmtc1 $10, $f12 \n\t"
|
|
1618
|
+
"dli $10, 0x3 \n\t"
|
|
1619
|
+
"dmtc1 $10, $f14 \n\t"
|
|
1620
|
+
"psllh $f24, $f24, $f12 \n\t"
|
|
1621
|
+
"psllh $f26, $f26, $f12 \n\t"
|
|
1622
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
1623
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
1624
|
+
"gslqc1 $f30, $f28, 0x20(%[tmp]) \n\t"
|
|
1625
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
1626
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
1627
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
1628
|
+
"psrah $f24, $f24, $f14 \n\t"
|
|
1629
|
+
"psrah $f26, $f26, $f14 \n\t"
|
|
1630
|
+
"dmtc1 %[iAlpha], $f12 \n\t"
|
|
1631
|
+
"dmtc1 %[iBeta], $f14 \n\t"
|
|
1632
|
+
"pmaxsh $f4, $f4, $f24 \n\t"
|
|
1633
|
+
"pmaxsh $f6, $f6, $f26 \n\t"
|
|
1634
|
+
"gssqc1 $f2, $f0, 0x10(%[tmp]) \n\t"
|
|
1635
|
+
"gslqc1 $f26, $f24, 0x10(%[tmp]) \n\t"
|
|
1636
|
+
"pminsh $f24, $f24, $f4 \n\t"
|
|
1637
|
+
"pminsh $f26, $f26, $f6 \n\t"
|
|
1638
|
+
"gssqc1 $f26, $f24, 0x10(%[tmp]) \n\t"
|
|
1639
|
+
"psubh $f4, $f8, $f12 \n\t"
|
|
1640
|
+
"psubh $f6, $f10, $f14 \n\t"
|
|
1641
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f24, $f26)
|
|
1642
|
+
"pcmpgth $f24, $f16, $f4 \n\t"
|
|
1643
|
+
"pcmpgth $f26, $f18, $f6 \n\t"
|
|
1644
|
+
"gslqc1 $f6, $f4, 0x30(%[tmp]) \n\t"
|
|
1645
|
+
"psubh $f4, $f4, $f8 \n\t"
|
|
1646
|
+
"psubh $f6, $f6, $f10 \n\t"
|
|
1647
|
+
"dmfc1 %[iAlpha], $f8 \n\t"
|
|
1648
|
+
"dmfc1 %[iBeta], $f10 \n\t"
|
|
1649
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f8, $f10)
|
|
1650
|
+
"pcmpgth $f28, $f28, $f4 \n\t"
|
|
1651
|
+
"pcmpgth $f30, $f30, $f6 \n\t"
|
|
1652
|
+
"gslqc1 $f6, $f4, 0x50(%[tmp]) \n\t"
|
|
1653
|
+
"and $f24, $f24, $f28 \n\t"
|
|
1654
|
+
"and $f26, $f26, $f30 \n\t"
|
|
1655
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
1656
|
+
"psubh $f20, $f20, $f12 \n\t"
|
|
1657
|
+
"psubh $f22, $f22, $f14 \n\t"
|
|
1658
|
+
WELS_AbsH($f20, $f22, $f20, $f22, $f8, $f10)
|
|
1659
|
+
"pcmpgth $f4, $f4, $f20 \n\t"
|
|
1660
|
+
"pcmpgth $f6, $f6, $f22 \n\t"
|
|
1661
|
+
"gslqc1 $f22, $f20, 0x80(%[tmp]) \n\t"
|
|
1662
|
+
"gslqc1 $f10, $f8, 0x90(%[tmp]) \n\t"
|
|
1663
|
+
"psubh $f20, $f20, $f8 \n\t"
|
|
1664
|
+
"psubh $f22, $f22, $f10 \n\t"
|
|
1665
|
+
"and $f24, $f24, $f4 \n\t"
|
|
1666
|
+
"and $f26, $f26, $f6 \n\t"
|
|
1667
|
+
"gslqc1 $f10, $f8, 0x40(%[tmp]) \n\t"
|
|
1668
|
+
"and $f24, $f24, $f8 \n\t"
|
|
1669
|
+
"and $f26, $f26, $f10 \n\t"
|
|
1670
|
+
"gslqc1 $f6, $f4, 0x10(%[tmp]) \n\t"
|
|
1671
|
+
"and $f4, $f4, $f24 \n\t"
|
|
1672
|
+
"and $f6, $f6, $f26 \n\t"
|
|
1673
|
+
"gslqc1 $f26, $f24, 0x70(%[tmp]) \n\t"
|
|
1674
|
+
"gssqc1 $f6, $f4, 0x30(%[tmp]) \n\t"
|
|
1675
|
+
"gslqc1 $f6, $f4, 0xa0(%[tmp]) \n\t"
|
|
1676
|
+
"psubh $f24, $f24, $f4 \n\t"
|
|
1677
|
+
"psubh $f26, $f26, $f6 \n\t"
|
|
1678
|
+
"dli $10, 0x2 \n\t"
|
|
1679
|
+
"dmtc1 $10, $f8 \n\t"
|
|
1680
|
+
"psllh $f24, $f24, $f8 \n\t"
|
|
1681
|
+
"psllh $f26, $f26, $f8 \n\t"
|
|
1682
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
1683
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
1684
|
+
"dli $10, 0x3 \n\t"
|
|
1685
|
+
"gslqc1 $f10, $f8, 0x20(%[tmp]) \n\t"
|
|
1686
|
+
"paddh $f24, $f24, $f8 \n\t"
|
|
1687
|
+
"paddh $f26, $f26, $f10 \n\t"
|
|
1688
|
+
"dmtc1 $10, $f8 \n\t"
|
|
1689
|
+
"gslqc1 $f22, $f20, 0x60(%[tmp]) \n\t"
|
|
1690
|
+
"psrah $f24, $f24, $f8 \n\t"
|
|
1691
|
+
"psrah $f26, $f26, $f8 \n\t"
|
|
1692
|
+
"pmaxsh $f20, $f20, $f24 \n\t"
|
|
1693
|
+
"pmaxsh $f22, $f22, $f26 \n\t"
|
|
1694
|
+
"pminsh $f0, $f0, $f20 \n\t"
|
|
1695
|
+
"pminsh $f2, $f2, $f22 \n\t"
|
|
1696
|
+
"gslqc1 $f22, $f20, 0x70(%[tmp]) \n\t"
|
|
1697
|
+
"psubh $f24, $f4, $f20 \n\t"
|
|
1698
|
+
"psubh $f26, $f6, $f22 \n\t"
|
|
1699
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f8, $f10)
|
|
1700
|
+
"pcmpgth $f16, $f16, $f24 \n\t"
|
|
1701
|
+
"pcmpgth $f18, $f18, $f26 \n\t"
|
|
1702
|
+
"gslqc1 $f26, $f24, 0x80(%[tmp]) \n\t"
|
|
1703
|
+
"psubh $f24, $f24, $f4 \n\t"
|
|
1704
|
+
"psubh $f26, $f26, $f6 \n\t"
|
|
1705
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f8, $f10)
|
|
1706
|
+
"pcmpgth $f28, $f28, $f24 \n\t"
|
|
1707
|
+
"pcmpgth $f30, $f30, $f26 \n\t"
|
|
1708
|
+
"gslqc1 $f26, $f24, 0x90(%[tmp]) \n\t"
|
|
1709
|
+
"and $f16, $f16, $f28 \n\t"
|
|
1710
|
+
"and $f18, $f18, $f30 \n\t"
|
|
1711
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
1712
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
1713
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
1714
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f8, $f10)
|
|
1715
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
1716
|
+
"dmtc1 %[iBeta], $f10 \n\t"
|
|
1717
|
+
"pcmpgth $f28, $f28, $f24 \n\t"
|
|
1718
|
+
"pcmpgth $f30, $f30, $f26 \n\t"
|
|
1719
|
+
"and $f16, $f16, $f28 \n\t"
|
|
1720
|
+
"and $f18, $f18, $f30 \n\t"
|
|
1721
|
+
"gslqc1 $f26, $f24, 0x40(%[tmp]) \n\t"
|
|
1722
|
+
"and $f16, $f16, $f24 \n\t"
|
|
1723
|
+
"and $f18, $f18, $f26 \n\t"
|
|
1724
|
+
"and $f0, $f0, $f16 \n\t"
|
|
1725
|
+
"and $f2, $f2, $f18 \n\t"
|
|
1726
|
+
"gslqc1 $f18, $f16, 0x30(%[tmp]) \n\t"
|
|
1727
|
+
"paddh $f8, $f8, $f16 \n\t"
|
|
1728
|
+
"paddh $f10, $f10, $f18 \n\t"
|
|
1729
|
+
"paddh $f4, $f4, $f0 \n\t"
|
|
1730
|
+
"paddh $f6, $f6, $f2 \n\t"
|
|
1731
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
1732
|
+
"packushb $f10, $f4, $f6 \n\t"
|
|
1733
|
+
"gssdxc1 $f8, 0x0(%[pTC], $0) \n\t"
|
|
1734
|
+
"psubh $f12, $f12, $f16 \n\t"
|
|
1735
|
+
"psubh $f14, $f14, $f18 \n\t"
|
|
1736
|
+
"psubh $f20, $f20, $f0 \n\t"
|
|
1737
|
+
"psubh $f22, $f22, $f2 \n\t"
|
|
1738
|
+
"packushb $f12, $f12, $f14 \n\t"
|
|
1739
|
+
"packushb $f14, $f20, $f22 \n\t"
|
|
1740
|
+
"gssdxc1 $f12, 0x0(%[pPixCb], $0) \n\t"
|
|
1741
|
+
"gssdxc1 $f10, 0x0($9, $0) \n\t"
|
|
1742
|
+
"gssdxc1 $f14, 0x0(%[pPixCr], $0) \n\t"
|
|
1743
|
+
: [pPixCb]"+&r"((unsigned char *)pPixCb), [pPixCr]"+&r"((unsigned char *)pPixCr)
|
|
1744
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha), [iBeta]"r"(iBeta),
|
|
1745
|
+
[pTC]"r"((unsigned char *)pTC), [tmp]"r"((unsigned char *)tmp)
|
|
1746
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$f0", "$f2", "$f4", "$f6", "$f8",
|
|
1747
|
+
"$f10", "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26",
|
|
1748
|
+
"$f28", "$f30"
|
|
1749
|
+
);
|
|
1750
|
+
RECOVER_REG;
|
|
1751
|
+
}
|
|
1752
|
+
|
|
1753
|
+
void DeblockChromaEq4V_mmi(uint8_t *pPixCb, uint8_t *pPixCr, int32_t iStride,
|
|
1754
|
+
int32_t iAlpha, int32_t iBeta) {
|
|
1755
|
+
unsigned char tmp[128] __attribute__((aligned(32)));
|
|
1756
|
+
BACKUP_REG;
|
|
1757
|
+
__asm__ volatile (
|
|
1758
|
+
".set arch=loongson3a \n\t"
|
|
1759
|
+
"daddu $8, %[iStride], %[iStride] \n\t"
|
|
1760
|
+
"dsubu $9, %[pPixCb], $8 \n\t"
|
|
1761
|
+
"gsldxc1 $f16, 0x0(%[pPixCr], $0) \n\t"
|
|
1762
|
+
"gsldxc1 $f20, 0x0(%[iStride], %[pPixCr]) \n\t"
|
|
1763
|
+
"gsldxc1 $f4, 0x0($9, $0) \n\t"
|
|
1764
|
+
"dsubu $9, %[pPixCr], $8 \n\t"
|
|
1765
|
+
"gsldxc1 $f8, 0x0($9, $0) \n\t"
|
|
1766
|
+
"mov.d $f6, $f8 \n\t"
|
|
1767
|
+
"dsubu $8, %[pPixCb], %[iStride] \n\t"
|
|
1768
|
+
"gsldxc1 $f8, 0x0($8, $0) \n\t"
|
|
1769
|
+
"dsubu $9, %[pPixCr], %[iStride] \n\t"
|
|
1770
|
+
"gsldxc1 $f12, 0x0($9, $0) \n\t"
|
|
1771
|
+
"mov.d $f10, $f12 \n\t"
|
|
1772
|
+
"gsldxc1 $f12, 0x0(%[pPixCb], $0) \n\t"
|
|
1773
|
+
"mov.d $f14, $f16 \n\t"
|
|
1774
|
+
"gsldxc1 $f16, 0x0(%[iStride], %[pPixCb]) \n\t"
|
|
1775
|
+
"mov.d $f18, $f20 \n\t"
|
|
1776
|
+
"dmtc1 %[iAlpha], $f20 \n\t"
|
|
1777
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
1778
|
+
"xor $f2, $f2, $f2 \n\t"
|
|
1779
|
+
"punpcklhw $f24, $f20, $f20 \n\t"
|
|
1780
|
+
"punpcklwd $f20, $f24, $f24 \n\t"
|
|
1781
|
+
"mov.d $f22, $f20 \n\t"
|
|
1782
|
+
"dmtc1 %[iBeta], $f24 \n\t"
|
|
1783
|
+
"punpcklhw $f28, $f24, $f24 \n\t"
|
|
1784
|
+
"punpcklwd $f24, $f28, $f28 \n\t"
|
|
1785
|
+
"mov.d $f26, $f24 \n\t"
|
|
1786
|
+
"mov.d $f28, $f4 \n\t"
|
|
1787
|
+
"punpcklbh $f4, $f6, $f2 \n\t"
|
|
1788
|
+
"punpckhbh $f6, $f6, $f2 \n\t"
|
|
1789
|
+
"punpckhbh $f30, $f28, $f0 \n\t"
|
|
1790
|
+
"punpcklbh $f28, $f28, $f0 \n\t"
|
|
1791
|
+
"gssqc1 $f6, $f4, 0x40(%[tmp]) \n\t"
|
|
1792
|
+
"gssqc1 $f30, $f28, 0x60(%[tmp]) \n\t"
|
|
1793
|
+
"punpckhbh $f30, $f8, $f0 \n\t"
|
|
1794
|
+
"punpcklbh $f28, $f8, $f0 \n\t"
|
|
1795
|
+
"gssqc1 $f30, $f28, 0x10(%[tmp]) \n\t"
|
|
1796
|
+
"punpckhbh $f30, $f12, $f0 \n\t"
|
|
1797
|
+
"punpcklbh $f28, $f12, $f0 \n\t"
|
|
1798
|
+
"punpcklbh $f12, $f14, $f2 \n\t"
|
|
1799
|
+
"punpckhbh $f14, $f14, $f2 \n\t"
|
|
1800
|
+
"gssqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
1801
|
+
"mov.d $f28, $f16 \n\t"
|
|
1802
|
+
"punpcklbh $f16, $f18, $f2 \n\t"
|
|
1803
|
+
"punpckhbh $f18, $f18, $f2 \n\t"
|
|
1804
|
+
"punpcklbh $f8, $f10, $f2 \n\t"
|
|
1805
|
+
"punpckhbh $f10, $f10, $f2 \n\t"
|
|
1806
|
+
"punpckhbh $f30, $f28, $f0 \n\t"
|
|
1807
|
+
"punpcklbh $f28, $f28, $f0 \n\t"
|
|
1808
|
+
"gssqc1 $f14, $f12, 0x30(%[tmp]) \n\t"
|
|
1809
|
+
"gslqc1 $f14, $f12, 0x10(%[tmp]) \n\t"
|
|
1810
|
+
"gslqc1 $f2, $f0, 0x50(%[tmp]) \n\t"
|
|
1811
|
+
"psubh $f4, $f12, $f0 \n\t"
|
|
1812
|
+
"psubh $f6, $f14, $f2 \n\t"
|
|
1813
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f0, $f2)
|
|
1814
|
+
"gssqc1 $f18, $f16, 0x20(%[tmp]) \n\t"
|
|
1815
|
+
"pcmpgth $f0, $f20, $f4 \n\t"
|
|
1816
|
+
"pcmpgth $f2, $f22, $f6 \n\t"
|
|
1817
|
+
"gslqc1 $f6, $f4, 0x60(%[tmp]) \n\t"
|
|
1818
|
+
"psubh $f4, $f4, $f12 \n\t"
|
|
1819
|
+
"psubh $f6, $f6, $f14 \n\t"
|
|
1820
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f16, $f18)
|
|
1821
|
+
"pcmpgth $f16, $f24, $f4 \n\t"
|
|
1822
|
+
"pcmpgth $f18, $f26, $f6 \n\t"
|
|
1823
|
+
"and $f0, $f0, $f16 \n\t"
|
|
1824
|
+
"and $f2, $f2, $f18 \n\t"
|
|
1825
|
+
"gslqc1 $f18, $f16, 0x50(%[tmp]) \n\t"
|
|
1826
|
+
"psubh $f4, $f28, $f16 \n\t"
|
|
1827
|
+
"psubh $f6, $f30, $f18 \n\t"
|
|
1828
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f16, $f18)
|
|
1829
|
+
"pcmpgth $f16, $f24, $f4 \n\t"
|
|
1830
|
+
"pcmpgth $f18, $f26, $f6 \n\t"
|
|
1831
|
+
"gslqc1 $f6, $f4, 0x30(%[tmp]) \n\t"
|
|
1832
|
+
"psubh $f4, $f8, $f4 \n\t"
|
|
1833
|
+
"psubh $f6, $f10, $f6 \n\t"
|
|
1834
|
+
"dmfc1 %[iAlpha], $f28 \n\t"
|
|
1835
|
+
"dmfc1 %[iBeta], $f30 \n\t"
|
|
1836
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f28, $f30)
|
|
1837
|
+
"pcmpgth $f20, $f20, $f4 \n\t"
|
|
1838
|
+
"pcmpgth $f22, $f22, $f6 \n\t"
|
|
1839
|
+
"gslqc1 $f6, $f4, 0x40(%[tmp]) \n\t"
|
|
1840
|
+
"and $f0, $f0, $f16 \n\t"
|
|
1841
|
+
"and $f2, $f2, $f18 \n\t"
|
|
1842
|
+
"psubh $f4, $f4, $f8 \n\t"
|
|
1843
|
+
"psubh $f6, $f6, $f10 \n\t"
|
|
1844
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f16, $f18)
|
|
1845
|
+
"pcmpgth $f16, $f24, $f4 \n\t"
|
|
1846
|
+
"pcmpgth $f18, $f26, $f6 \n\t"
|
|
1847
|
+
"gslqc1 $f6, $f4, 0x20(%[tmp]) \n\t"
|
|
1848
|
+
"gslqc1 $f30, $f28, 0x30(%[tmp]) \n\t"
|
|
1849
|
+
"psubh $f4, $f4, $f28 \n\t"
|
|
1850
|
+
"psubh $f6, $f6, $f30 \n\t"
|
|
1851
|
+
"and $f20, $f20, $f16 \n\t"
|
|
1852
|
+
"and $f22, $f22, $f18 \n\t"
|
|
1853
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f28, $f30)
|
|
1854
|
+
"dmtc1 %[iAlpha], $f28 \n\t"
|
|
1855
|
+
"dmtc1 %[iBeta], $f30 \n\t"
|
|
1856
|
+
"pcmpgth $f24, $f24, $f4 \n\t"
|
|
1857
|
+
"pcmpgth $f26, $f26, $f6 \n\t"
|
|
1858
|
+
"and $f20, $f20, $f24 \n\t"
|
|
1859
|
+
"and $f22, $f22, $f26 \n\t"
|
|
1860
|
+
"dli %[iBeta], 0x2 \n\t"
|
|
1861
|
+
"dmtc1 %[iBeta], $f4 \n\t"
|
|
1862
|
+
"punpcklhw $f16, $f4, $f4 \n\t"
|
|
1863
|
+
"punpcklwd $f4, $f16, $f16 \n\t"
|
|
1864
|
+
"mov.d $f6, $f4 \n\t"
|
|
1865
|
+
"gslqc1 $f18, $f16, 0x60(%[tmp]) \n\t"
|
|
1866
|
+
"paddh $f24, $f16, $f16 \n\t"
|
|
1867
|
+
"paddh $f26, $f18, $f18 \n\t"
|
|
1868
|
+
"paddh $f24, $f24, $f12 \n\t"
|
|
1869
|
+
"paddh $f26, $f26, $f14 \n\t"
|
|
1870
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
1871
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
1872
|
+
"gssqc1 $f6, $f4, 0x10(%[tmp]) \n\t"
|
|
1873
|
+
"gslqc1 $f18, $f16, 0x10(%[tmp]) \n\t"
|
|
1874
|
+
"paddh $f24, $f24, $f16 \n\t"
|
|
1875
|
+
"paddh $f26, $f26, $f18 \n\t"
|
|
1876
|
+
"dmtc1 %[iBeta], $f16 \n\t"
|
|
1877
|
+
"psrah $f24, $f24, $f16 \n\t"
|
|
1878
|
+
"psrah $f26, $f26, $f16 \n\t"
|
|
1879
|
+
"pandn $f16, $f0, $f12 \n\t"
|
|
1880
|
+
"pandn $f18, $f2, $f14 \n\t"
|
|
1881
|
+
"gslqc1 $f14, $f12, 0x40(%[tmp]) \n\t"
|
|
1882
|
+
"and $f4, $f0, $f24 \n\t"
|
|
1883
|
+
"and $f6, $f2, $f26 \n\t"
|
|
1884
|
+
"or $f4, $f4, $f16 \n\t"
|
|
1885
|
+
"or $f6, $f6, $f18 \n\t"
|
|
1886
|
+
"paddh $f24, $f12, $f12 \n\t"
|
|
1887
|
+
"paddh $f26, $f14, $f14 \n\t"
|
|
1888
|
+
"gslqc1 $f14, $f12, 0x10(%[tmp]) \n\t"
|
|
1889
|
+
"paddh $f24, $f24, $f8 \n\t"
|
|
1890
|
+
"paddh $f26, $f26, $f10 \n\t"
|
|
1891
|
+
"gslqc1 $f18, $f16, 0x20(%[tmp]) \n\t"
|
|
1892
|
+
"paddh $f24, $f24, $f16 \n\t"
|
|
1893
|
+
"paddh $f26, $f26, $f18 \n\t"
|
|
1894
|
+
"dmtc1 %[iBeta], $f16 \n\t"
|
|
1895
|
+
"paddh $f24, $f24, $f12 \n\t"
|
|
1896
|
+
"paddh $f26, $f26, $f14 \n\t"
|
|
1897
|
+
"psrah $f24, $f24, $f16 \n\t"
|
|
1898
|
+
"psrah $f26, $f26, $f16 \n\t"
|
|
1899
|
+
"and $f16, $f20, $f24 \n\t"
|
|
1900
|
+
"and $f18, $f22, $f26 \n\t"
|
|
1901
|
+
"pandn $f24, $f20, $f8 \n\t"
|
|
1902
|
+
"pandn $f26, $f22, $f10 \n\t"
|
|
1903
|
+
"or $f16, $f16, $f24 \n\t"
|
|
1904
|
+
"or $f18, $f18, $f26 \n\t"
|
|
1905
|
+
"packushb $f4, $f4, $f6 \n\t"
|
|
1906
|
+
"packushb $f6, $f16, $f18 \n\t"
|
|
1907
|
+
"gslqc1 $f18, $f16, 0x50(%[tmp]) \n\t"
|
|
1908
|
+
"paddh $f24, $f28, $f28 \n\t"
|
|
1909
|
+
"paddh $f26, $f30, $f30 \n\t"
|
|
1910
|
+
"paddh $f24, $f24, $f16 \n\t"
|
|
1911
|
+
"paddh $f26, $f26, $f18 \n\t"
|
|
1912
|
+
"gslqc1 $f10, $f8, 0x60(%[tmp]) \n\t"
|
|
1913
|
+
"paddh $f24, $f24, $f8 \n\t"
|
|
1914
|
+
"paddh $f26, $f26, $f10 \n\t"
|
|
1915
|
+
"dmtc1 %[iBeta], $f28 \n\t"
|
|
1916
|
+
"paddh $f24, $f24, $f12 \n\t"
|
|
1917
|
+
"paddh $f26, $f26, $f14 \n\t"
|
|
1918
|
+
"psrah $f24, $f24, $f28 \n\t"
|
|
1919
|
+
"psrah $f26, $f26, $f28 \n\t"
|
|
1920
|
+
"and $f8, $f0, $f24 \n\t"
|
|
1921
|
+
"and $f10, $f2, $f26 \n\t"
|
|
1922
|
+
"pandn $f0, $f0, $f16 \n\t"
|
|
1923
|
+
"pandn $f2, $f2, $f18 \n\t"
|
|
1924
|
+
"or $f8, $f8, $f0 \n\t"
|
|
1925
|
+
"or $f10, $f10, $f2 \n\t"
|
|
1926
|
+
"gslqc1 $f2, $f0, 0x20(%[tmp]) \n\t"
|
|
1927
|
+
"paddh $f24, $f0, $f0 \n\t"
|
|
1928
|
+
"paddh $f26, $f2, $f2 \n\t"
|
|
1929
|
+
"gslqc1 $f2, $f0, 0x30(%[tmp]) \n\t"
|
|
1930
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
1931
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
1932
|
+
"gslqc1 $f18, $f16, 0x40(%[tmp]) \n\t"
|
|
1933
|
+
"paddh $f24, $f24, $f16 \n\t"
|
|
1934
|
+
"paddh $f26, $f26, $f18 \n\t"
|
|
1935
|
+
"paddh $f24, $f24, $f12 \n\t"
|
|
1936
|
+
"paddh $f26, $f26, $f14 \n\t"
|
|
1937
|
+
"gssdxc1 $f4, 0x0($8, $0) \n\t"
|
|
1938
|
+
"psrah $f24, $f24, $f28 \n\t"
|
|
1939
|
+
"psrah $f26, $f26, $f28 \n\t"
|
|
1940
|
+
"and $f16, $f20, $f24 \n\t"
|
|
1941
|
+
"and $f18, $f22, $f26 \n\t"
|
|
1942
|
+
"pandn $f20, $f20, $f0 \n\t"
|
|
1943
|
+
"pandn $f22, $f22, $f2 \n\t"
|
|
1944
|
+
"or $f16, $f16, $f20 \n\t"
|
|
1945
|
+
"or $f18, $f18, $f22 \n\t"
|
|
1946
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
1947
|
+
"packushb $f10, $f16, $f18 \n\t"
|
|
1948
|
+
"gssdxc1 $f8, 0x0(%[pPixCb], $0) \n\t"
|
|
1949
|
+
"gssdxc1 $f6, 0x0($9, $0) \n\t"
|
|
1950
|
+
"gssdxc1 $f10, 0x0(%[pPixCr], $0) \n\t"
|
|
1951
|
+
: [pPixCb]"+&r"((unsigned char *)pPixCb), [pPixCr]"+&r"((unsigned char *)pPixCr)
|
|
1952
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha),
|
|
1953
|
+
[iBeta]"r"(iBeta), [tmp]"r"((unsigned char *)tmp)
|
|
1954
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$f0", "$f2", "$f4", "$f6", "$f8",
|
|
1955
|
+
"$f10", "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26",
|
|
1956
|
+
"$f28", "$f30"
|
|
1957
|
+
);
|
|
1958
|
+
RECOVER_REG;
|
|
1959
|
+
}
|
|
1960
|
+
|
|
1961
|
+
void DeblockChromaEq4H_mmi(uint8_t *pPixCb, uint8_t *pPixCr, int32_t iStride,
|
|
1962
|
+
int32_t iAlpha, int32_t iBeta) {
|
|
1963
|
+
unsigned char tmp[256] __attribute__((aligned(32)));
|
|
1964
|
+
BACKUP_REG;
|
|
1965
|
+
__asm__ volatile (
|
|
1966
|
+
".set arch=loongson3a \n\t"
|
|
1967
|
+
"daddiu %[pPixCb], %[pPixCb], -0x2 \n\t"
|
|
1968
|
+
"daddiu %[pPixCr], %[pPixCr], -0x2 \n\t"
|
|
1969
|
+
"move $9, %[pPixCb] \n\t"
|
|
1970
|
+
"move $10, %[pPixCr] \n\t"
|
|
1971
|
+
"dsll $11, %[iStride], 0x2 \n\t"
|
|
1972
|
+
"daddu %[pPixCb], %[pPixCb], $11 \n\t"
|
|
1973
|
+
"daddu %[pPixCr], %[pPixCr], $11 \n\t"
|
|
1974
|
+
"daddiu $11, %[tmp], 0x80 \n\t"
|
|
1975
|
+
"gsldlc1 $f0, 0x7($9) \n\t"
|
|
1976
|
+
"gsldrc1 $f0, 0x0($9) \n\t"
|
|
1977
|
+
"daddu $12, $9, %[iStride] \n\t"
|
|
1978
|
+
"gsldlc1 $f4, 0x7($12) \n\t"
|
|
1979
|
+
"gsldrc1 $f4, 0x0($12) \n\t"
|
|
1980
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
1981
|
+
"gsldlc1 $f8, 0x7($12) \n\t"
|
|
1982
|
+
"gsldrc1 $f8, 0x0($12) \n\t"
|
|
1983
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
1984
|
+
"gsldlc1 $f12, 0x7($12) \n\t"
|
|
1985
|
+
"gsldlc1 $f16, 0x7($10) \n\t"
|
|
1986
|
+
"gsldrc1 $f12, 0x0($12) \n\t"
|
|
1987
|
+
"gsldrc1 $f16, 0x0($10) \n\t"
|
|
1988
|
+
"daddu $12, $10, %[iStride] \n\t"
|
|
1989
|
+
"gsldlc1 $f20, 0x7($12) \n\t"
|
|
1990
|
+
"gsldrc1 $f20, 0x0($12) \n\t"
|
|
1991
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
1992
|
+
"gsldlc1 $f24, 0x7($12) \n\t"
|
|
1993
|
+
"gsldrc1 $f24, 0x0($12) \n\t"
|
|
1994
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
1995
|
+
"gsldlc1 $f28, 0x7($12) \n\t"
|
|
1996
|
+
"gsldrc1 $f28, 0x0($12) \n\t"
|
|
1997
|
+
"punpcklwd $f0, $f0, $f16 \n\t"
|
|
1998
|
+
"punpcklwd $f4, $f4, $f20 \n\t"
|
|
1999
|
+
"punpcklwd $f8, $f8, $f24 \n\t"
|
|
2000
|
+
"punpcklwd $f12, $f12, $f28 \n\t"
|
|
2001
|
+
"gsldlc1 $f16, 0x7(%[pPixCb]) \n\t"
|
|
2002
|
+
"gsldlc1 $f20, 0x7(%[pPixCr]) \n\t"
|
|
2003
|
+
"gsldrc1 $f16, 0x0(%[pPixCb]) \n\t"
|
|
2004
|
+
"gsldrc1 $f20, 0x0(%[pPixCr]) \n\t"
|
|
2005
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2006
|
+
"mov.d $f2, $f16 \n\t"
|
|
2007
|
+
"daddu $12, %[pPixCb], %[iStride] \n\t"
|
|
2008
|
+
"daddu $13, %[pPixCr], %[iStride] \n\t"
|
|
2009
|
+
"gsldlc1 $f16, 0x7($12) \n\t"
|
|
2010
|
+
"gsldlc1 $f20, 0x7($13) \n\t"
|
|
2011
|
+
"gsldrc1 $f16, 0x0($12) \n\t"
|
|
2012
|
+
"gsldrc1 $f20, 0x0($13) \n\t"
|
|
2013
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2014
|
+
"mov.d $f6, $f16 \n\t"
|
|
2015
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2016
|
+
"daddu $13, $13, %[iStride] \n\t"
|
|
2017
|
+
"gsldlc1 $f16, 0x7($12) \n\t"
|
|
2018
|
+
"gsldlc1 $f20, 0x7($13) \n\t"
|
|
2019
|
+
"gsldrc1 $f16, 0x0($12) \n\t"
|
|
2020
|
+
"gsldrc1 $f20, 0x0($13) \n\t"
|
|
2021
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2022
|
+
"mov.d $f10, $f16 \n\t"
|
|
2023
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2024
|
+
"daddu $13, $13, %[iStride] \n\t"
|
|
2025
|
+
"gsldlc1 $f16, 0x7($12) \n\t"
|
|
2026
|
+
"gsldlc1 $f20, 0x7($13) \n\t"
|
|
2027
|
+
"gsldrc1 $f16, 0x0($12) \n\t"
|
|
2028
|
+
"gsldrc1 $f20, 0x0($13) \n\t"
|
|
2029
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2030
|
+
"mov.d $f14, $f16 \n\t"
|
|
2031
|
+
"punpcklbh $f24, $f2, $f6 \n\t"
|
|
2032
|
+
"punpckhbh $f26, $f2, $f6 \n\t"
|
|
2033
|
+
"punpckhbh $f2, $f0, $f4 \n\t"
|
|
2034
|
+
"punpcklbh $f0, $f0, $f4 \n\t"
|
|
2035
|
+
"punpcklbh $f28, $f10, $f14 \n\t"
|
|
2036
|
+
"punpckhbh $f30, $f10, $f14 \n\t"
|
|
2037
|
+
"punpckhbh $f10, $f8, $f12 \n\t"
|
|
2038
|
+
"punpcklbh $f8, $f8, $f12 \n\t"
|
|
2039
|
+
"punpcklhw $f16, $f2, $f10 \n\t"
|
|
2040
|
+
"punpckhhw $f18, $f2, $f10 \n\t"
|
|
2041
|
+
"punpckhhw $f2, $f0, $f8 \n\t"
|
|
2042
|
+
"punpcklhw $f0, $f0, $f8 \n\t"
|
|
2043
|
+
"punpcklhw $f20, $f26, $f30 \n\t"
|
|
2044
|
+
"punpckhhw $f22, $f26, $f30 \n\t"
|
|
2045
|
+
"punpckhhw $f26, $f24, $f28 \n\t"
|
|
2046
|
+
"punpcklhw $f24, $f24, $f28 \n\t"
|
|
2047
|
+
"punpcklwd $f4, $f2, $f26 \n\t"
|
|
2048
|
+
"punpckhwd $f6, $f2, $f26 \n\t"
|
|
2049
|
+
"punpckhwd $f2, $f0, $f24 \n\t"
|
|
2050
|
+
"punpcklwd $f0, $f0, $f24 \n\t"
|
|
2051
|
+
"punpcklwd $f8, $f18, $f22 \n\t"
|
|
2052
|
+
"punpckhwd $f10, $f18, $f22 \n\t"
|
|
2053
|
+
"punpckhwd $f18, $f16, $f20 \n\t"
|
|
2054
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2055
|
+
"mov.d $f20, $f2 \n\t"
|
|
2056
|
+
"mov.d $f22, $f18 \n\t"
|
|
2057
|
+
"mov.d $f2, $f16 \n\t"
|
|
2058
|
+
"mov.d $f24, $f6 \n\t"
|
|
2059
|
+
"mov.d $f26, $f10 \n\t"
|
|
2060
|
+
"mov.d $f6, $f8 \n\t"
|
|
2061
|
+
"gssqc1 $f2, $f0, 0x0($11) \n\t"
|
|
2062
|
+
"gssqc1 $f22, $f20, 0x10($11) \n\t"
|
|
2063
|
+
"gssqc1 $f6, $f4, 0x20($11) \n\t"
|
|
2064
|
+
"gssqc1 $f26, $f24, 0x30($11) \n\t"
|
|
2065
|
+
"gslqc1 $f26, $f24, 0x80(%[tmp]) \n\t"
|
|
2066
|
+
"gslqc1 $f18, $f16, 0x90(%[tmp]) \n\t"
|
|
2067
|
+
"gslqc1 $f22, $f20, 0xa0(%[tmp]) \n\t"
|
|
2068
|
+
"gslqc1 $f30, $f28, 0xb0(%[tmp]) \n\t"
|
|
2069
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
2070
|
+
"dmtc1 %[iAlpha], $f4 \n\t"
|
|
2071
|
+
"punpcklhw $f8, $f4, $f4 \n\t"
|
|
2072
|
+
"punpcklwd $f4, $f8, $f8 \n\t"
|
|
2073
|
+
"mov.d $f6, $f4 \n\t"
|
|
2074
|
+
"dmtc1 %[iBeta], $f8 \n\t"
|
|
2075
|
+
"punpcklhw $f12, $f8, $f8 \n\t"
|
|
2076
|
+
"punpcklwd $f8, $f12, $f12 \n\t"
|
|
2077
|
+
"mov.d $f10, $f8 \n\t"
|
|
2078
|
+
"mov.d $f12, $f24 \n\t"
|
|
2079
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
2080
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
2081
|
+
"gssqc1 $f26, $f24, 0x60(%[tmp]) \n\t"
|
|
2082
|
+
"gslqc1 $f26, $f24, 0x90(%[tmp]) \n\t"
|
|
2083
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
2084
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
2085
|
+
"gssqc1 $f26, $f24, 0x30(%[tmp]) \n\t"
|
|
2086
|
+
"gslqc1 $f26, $f24, 0xa0(%[tmp]) \n\t"
|
|
2087
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
2088
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
2089
|
+
"gssqc1 $f26, $f24, 0x40(%[tmp]) \n\t"
|
|
2090
|
+
"gslqc1 $f26, $f24, 0xb0(%[tmp]) \n\t"
|
|
2091
|
+
"punpcklbh $f24, $f26, $f0 \n\t"
|
|
2092
|
+
"punpckhbh $f26, $f26, $f0 \n\t"
|
|
2093
|
+
"gssqc1 $f26, $f24, 0x70(%[tmp]) \n\t"
|
|
2094
|
+
"punpckhbh $f30, $f28, $f0 \n\t"
|
|
2095
|
+
"punpcklbh $f28, $f28, $f0 \n\t"
|
|
2096
|
+
"punpckhbh $f18, $f16, $f0 \n\t"
|
|
2097
|
+
"punpcklbh $f16, $f16, $f0 \n\t"
|
|
2098
|
+
"punpckhbh $f22, $f20, $f0 \n\t"
|
|
2099
|
+
"punpcklbh $f20, $f20, $f0 \n\t"
|
|
2100
|
+
"punpckhbh $f14, $f12, $f0 \n\t"
|
|
2101
|
+
"punpcklbh $f12, $f12, $f0 \n\t"
|
|
2102
|
+
"gssqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
2103
|
+
"psubh $f24, $f16, $f20 \n\t"
|
|
2104
|
+
"psubh $f26, $f18, $f22 \n\t"
|
|
2105
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f0, $f2)
|
|
2106
|
+
"pcmpgth $f0, $f4, $f24 \n\t"
|
|
2107
|
+
"pcmpgth $f2, $f6, $f26 \n\t"
|
|
2108
|
+
"psubh $f24, $f12, $f16 \n\t"
|
|
2109
|
+
"psubh $f26, $f14, $f18 \n\t"
|
|
2110
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
2111
|
+
"pcmpgth $f28, $f8, $f24 \n\t"
|
|
2112
|
+
"pcmpgth $f30, $f10, $f26 \n\t"
|
|
2113
|
+
"gslqc1 $f26, $f24, 0x50(%[tmp]) \n\t"
|
|
2114
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
2115
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
2116
|
+
"and $f0, $f0, $f28 \n\t"
|
|
2117
|
+
"and $f2, $f2, $f30 \n\t"
|
|
2118
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f28, $f30)
|
|
2119
|
+
"dmfc1 %[iAlpha], $f20 \n\t"
|
|
2120
|
+
"dmfc1 %[iBeta], $f22 \n\t"
|
|
2121
|
+
"pcmpgth $f28, $f8, $f24 \n\t"
|
|
2122
|
+
"pcmpgth $f30, $f10, $f26 \n\t"
|
|
2123
|
+
"gslqc1 $f26, $f24, 0x30(%[tmp]) \n\t"
|
|
2124
|
+
"gslqc1 $f22, $f20, 0x40(%[tmp]) \n\t"
|
|
2125
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
2126
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
2127
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f20, $f22)
|
|
2128
|
+
"pcmpgth $f4, $f4, $f24 \n\t"
|
|
2129
|
+
"pcmpgth $f6, $f6, $f26 \n\t"
|
|
2130
|
+
"gslqc1 $f26, $f24, 0x60(%[tmp]) \n\t"
|
|
2131
|
+
"gslqc1 $f22, $f20, 0x30(%[tmp]) \n\t"
|
|
2132
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
2133
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
2134
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f20, $f22)
|
|
2135
|
+
"and $f0, $f0, $f28 \n\t"
|
|
2136
|
+
"and $f2, $f2, $f30 \n\t"
|
|
2137
|
+
"pcmpgth $f28, $f8, $f24 \n\t"
|
|
2138
|
+
"pcmpgth $f30, $f10, $f26 \n\t"
|
|
2139
|
+
"gslqc1 $f26, $f24, 0x70(%[tmp]) \n\t"
|
|
2140
|
+
"gslqc1 $f22, $f20, 0x40(%[tmp]) \n\t"
|
|
2141
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
2142
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
2143
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f20, $f22)
|
|
2144
|
+
"dli $8, 0x2 \n\t"
|
|
2145
|
+
"and $f4, $f4, $f28 \n\t"
|
|
2146
|
+
"and $f6, $f6, $f30 \n\t"
|
|
2147
|
+
"pcmpgth $f8, $f8, $f24 \n\t"
|
|
2148
|
+
"pcmpgth $f10, $f10, $f26 \n\t"
|
|
2149
|
+
"and $f4, $f4, $f8 \n\t"
|
|
2150
|
+
"and $f6, $f6, $f10 \n\t"
|
|
2151
|
+
"dmtc1 $8, $f8 \n\t"
|
|
2152
|
+
"punpcklhw $f24, $f8, $f8 \n\t"
|
|
2153
|
+
"punpcklwd $f8, $f24, $f24 \n\t"
|
|
2154
|
+
"mov.d $f10, $f8 \n\t"
|
|
2155
|
+
"gssqc1 $f10, $f8, 0x20(%[tmp]) \n\t"
|
|
2156
|
+
"paddh $f8, $f12, $f12 \n\t"
|
|
2157
|
+
"paddh $f10, $f14, $f14 \n\t"
|
|
2158
|
+
"paddh $f8, $f8, $f16 \n\t"
|
|
2159
|
+
"paddh $f10, $f10, $f18 \n\t"
|
|
2160
|
+
"gslqc1 $f22, $f20, 0x50(%[tmp]) \n\t"
|
|
2161
|
+
"paddh $f8, $f8, $f20 \n\t"
|
|
2162
|
+
"paddh $f10, $f10, $f22 \n\t"
|
|
2163
|
+
"gslqc1 $f26, $f24, 0x20(%[tmp]) \n\t"
|
|
2164
|
+
"paddh $f8, $f8, $f24 \n\t"
|
|
2165
|
+
"paddh $f10, $f10, $f26 \n\t"
|
|
2166
|
+
"dmtc1 $8, $f20 \n\t"
|
|
2167
|
+
"psrah $f8, $f8, $f20 \n\t"
|
|
2168
|
+
"psrah $f10, $f10, $f20 \n\t"
|
|
2169
|
+
"and $f24, $f0, $f8 \n\t"
|
|
2170
|
+
"and $f26, $f2, $f10 \n\t"
|
|
2171
|
+
"pandn $f8, $f0, $f16 \n\t"
|
|
2172
|
+
"pandn $f10, $f2, $f18 \n\t"
|
|
2173
|
+
"or $f24, $f24, $f8 \n\t"
|
|
2174
|
+
"or $f26, $f26, $f10 \n\t"
|
|
2175
|
+
"gslqc1 $f10, $f8, 0x60(%[tmp]) \n\t"
|
|
2176
|
+
"paddh $f28, $f8, $f8 \n\t"
|
|
2177
|
+
"paddh $f30, $f10, $f10 \n\t"
|
|
2178
|
+
"gslqc1 $f22, $f20, 0x30(%[tmp]) \n\t"
|
|
2179
|
+
"paddh $f28, $f28, $f20 \n\t"
|
|
2180
|
+
"paddh $f30, $f30, $f22 \n\t"
|
|
2181
|
+
"gslqc1 $f18, $f16, 0x70(%[tmp]) \n\t"
|
|
2182
|
+
"paddh $f28, $f28, $f16 \n\t"
|
|
2183
|
+
"paddh $f30, $f30, $f18 \n\t"
|
|
2184
|
+
"gslqc1 $f10, $f8, 0x20(%[tmp]) \n\t"
|
|
2185
|
+
"paddh $f28, $f28, $f8 \n\t"
|
|
2186
|
+
"paddh $f30, $f30, $f10 \n\t"
|
|
2187
|
+
"pandn $f8, $f4, $f20 \n\t"
|
|
2188
|
+
"pandn $f10, $f6, $f22 \n\t"
|
|
2189
|
+
"dmtc1 $8, $f20 \n\t"
|
|
2190
|
+
"psrah $f28, $f28, $f20 \n\t"
|
|
2191
|
+
"psrah $f30, $f30, $f20 \n\t"
|
|
2192
|
+
"and $f16, $f4, $f28 \n\t"
|
|
2193
|
+
"and $f18, $f6, $f30 \n\t"
|
|
2194
|
+
"or $f16, $f16, $f8 \n\t"
|
|
2195
|
+
"or $f18, $f18, $f10 \n\t"
|
|
2196
|
+
"gslqc1 $f10, $f8, 0x50(%[tmp]) \n\t"
|
|
2197
|
+
"packushb $f24, $f24, $f26 \n\t"
|
|
2198
|
+
"packushb $f26, $f16, $f18 \n\t"
|
|
2199
|
+
"gssqc1 $f26, $f24, 0x90(%[tmp]) \n\t"
|
|
2200
|
+
"paddh $f24, $f8, $f8 \n\t"
|
|
2201
|
+
"paddh $f26, $f10, $f10 \n\t"
|
|
2202
|
+
"dmtc1 %[iAlpha], $f20 \n\t"
|
|
2203
|
+
"dmtc1 %[iBeta], $f22 \n\t"
|
|
2204
|
+
"gslqc1 $f10, $f8, 0x20(%[tmp]) \n\t"
|
|
2205
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
2206
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
2207
|
+
"paddh $f24, $f24, $f12 \n\t"
|
|
2208
|
+
"paddh $f26, $f26, $f14 \n\t"
|
|
2209
|
+
"mov.d $f16, $f0 \n\t"
|
|
2210
|
+
"mov.d $f18, $f2 \n\t"
|
|
2211
|
+
"pandn $f0, $f0, $f20 \n\t"
|
|
2212
|
+
"pandn $f2, $f2, $f22 \n\t"
|
|
2213
|
+
"dmtc1 $8, $f20 \n\t"
|
|
2214
|
+
"paddh $f24, $f24, $f8 \n\t"
|
|
2215
|
+
"paddh $f26, $f26, $f10 \n\t"
|
|
2216
|
+
"psrah $f24, $f24, $f20 \n\t"
|
|
2217
|
+
"psrah $f26, $f26, $f20 \n\t"
|
|
2218
|
+
"and $f16, $f16, $f24 \n\t"
|
|
2219
|
+
"and $f18, $f18, $f26 \n\t"
|
|
2220
|
+
"or $f16, $f16, $f0 \n\t"
|
|
2221
|
+
"or $f18, $f18, $f2 \n\t"
|
|
2222
|
+
"gslqc1 $f2, $f0, 0x70(%[tmp]) \n\t"
|
|
2223
|
+
"paddh $f20, $f0, $f0 \n\t"
|
|
2224
|
+
"paddh $f22, $f2, $f2 \n\t"
|
|
2225
|
+
"gslqc1 $f2, $f0, 0x40(%[tmp]) \n\t"
|
|
2226
|
+
"paddh $f20, $f20, $f0 \n\t"
|
|
2227
|
+
"paddh $f22, $f22, $f2 \n\t"
|
|
2228
|
+
"gslqc1 $f14, $f12, 0x60(%[tmp]) \n\t"
|
|
2229
|
+
"paddh $f20, $f20, $f12 \n\t"
|
|
2230
|
+
"paddh $f22, $f22, $f14 \n\t"
|
|
2231
|
+
"paddh $f20, $f20, $f8 \n\t"
|
|
2232
|
+
"paddh $f22, $f22, $f10 \n\t"
|
|
2233
|
+
"dmtc1 $8, $f8 \n\t"
|
|
2234
|
+
"psrah $f20, $f20, $f8 \n\t"
|
|
2235
|
+
"psrah $f22, $f22, $f8 \n\t"
|
|
2236
|
+
"and $f12, $f4, $f20 \n\t"
|
|
2237
|
+
"and $f14, $f6, $f22 \n\t"
|
|
2238
|
+
"pandn $f4, $f4, $f0 \n\t"
|
|
2239
|
+
"pandn $f6, $f6, $f2 \n\t"
|
|
2240
|
+
"or $f12, $f12, $f4 \n\t"
|
|
2241
|
+
"or $f14, $f14, $f6 \n\t"
|
|
2242
|
+
"packushb $f16, $f16, $f18 \n\t"
|
|
2243
|
+
"packushb $f18, $f12, $f14 \n\t"
|
|
2244
|
+
"gssqc1 $f18, $f16, 0xa0(%[tmp]) \n\t"
|
|
2245
|
+
"gslqc1 $f2, $f0, 0x0($11) \n\t"
|
|
2246
|
+
"gslqc1 $f6, $f4, 0x10($11) \n\t"
|
|
2247
|
+
"gslqc1 $f10, $f8, 0x20($11) \n\t"
|
|
2248
|
+
"gslqc1 $f14, $f12, 0x30($11) \n\t"
|
|
2249
|
+
"mov.d $f26, $f2 \n\t"
|
|
2250
|
+
"punpckhbh $f2, $f0, $f4 \n\t"
|
|
2251
|
+
"punpcklbh $f0, $f0, $f4 \n\t"
|
|
2252
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
2253
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
2254
|
+
"mov.d $f30, $f10 \n\t"
|
|
2255
|
+
"punpckhbh $f10, $f8, $f12 \n\t"
|
|
2256
|
+
"punpcklbh $f8, $f8, $f12 \n\t"
|
|
2257
|
+
"punpcklbh $f28, $f30, $f14 \n\t"
|
|
2258
|
+
"punpckhbh $f30, $f30, $f14 \n\t"
|
|
2259
|
+
"punpcklhw $f16, $f2, $f10 \n\t"
|
|
2260
|
+
"punpckhhw $f18, $f2, $f10 \n\t"
|
|
2261
|
+
"punpcklhw $f20, $f26, $f30 \n\t"
|
|
2262
|
+
"punpckhhw $f22, $f26, $f30 \n\t"
|
|
2263
|
+
"punpckhhw $f2, $f0, $f8 \n\t"
|
|
2264
|
+
"punpcklhw $f0, $f0, $f8 \n\t"
|
|
2265
|
+
"punpckhhw $f26, $f24, $f28 \n\t"
|
|
2266
|
+
"punpcklhw $f24, $f24, $f28 \n\t"
|
|
2267
|
+
"punpcklwd $f4, $f2, $f26 \n\t"
|
|
2268
|
+
"punpckhwd $f6, $f2, $f26 \n\t"
|
|
2269
|
+
"punpcklwd $f8, $f18, $f22 \n\t"
|
|
2270
|
+
"punpckhwd $f10, $f18, $f22 \n\t"
|
|
2271
|
+
"punpckhwd $f2, $f0, $f24 \n\t"
|
|
2272
|
+
"punpcklwd $f0, $f0, $f24 \n\t"
|
|
2273
|
+
"punpckhwd $f18, $f16, $f20 \n\t"
|
|
2274
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2275
|
+
"mov.d $f20, $f2 \n\t"
|
|
2276
|
+
"mov.d $f24, $f6 \n\t"
|
|
2277
|
+
"mov.d $f2, $f16 \n\t"
|
|
2278
|
+
"mov.d $f22, $f18 \n\t"
|
|
2279
|
+
"mov.d $f6, $f8 \n\t"
|
|
2280
|
+
"mov.d $f26, $f10 \n\t"
|
|
2281
|
+
"dli %[iAlpha], 0x20 \n\t"
|
|
2282
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
2283
|
+
"gsswlc1 $f0, 0x3($9) \n\t"
|
|
2284
|
+
"gsswrc1 $f0, 0x0($9) \n\t"
|
|
2285
|
+
"daddu $12, $9, %[iStride] \n\t"
|
|
2286
|
+
"gsswlc1 $f20, 0x3($12) \n\t"
|
|
2287
|
+
"gsswrc1 $f20, 0x0($12) \n\t"
|
|
2288
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2289
|
+
"gsswlc1 $f4, 0x3($12) \n\t"
|
|
2290
|
+
"gsswrc1 $f4, 0x0($12) \n\t"
|
|
2291
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2292
|
+
"gsswlc1 $f24, 0x3($12) \n\t"
|
|
2293
|
+
"gsswrc1 $f24, 0x0($12) \n\t"
|
|
2294
|
+
"dsrl $f0, $f0, $f8 \n\t"
|
|
2295
|
+
"dsrl $f20, $f20, $f8 \n\t"
|
|
2296
|
+
"dsrl $f4, $f4, $f8 \n\t"
|
|
2297
|
+
"dsrl $f24, $f24, $f8 \n\t"
|
|
2298
|
+
"gsswlc1 $f0, 0x3($10) \n\t"
|
|
2299
|
+
"gsswrc1 $f0, 0x0($10) \n\t"
|
|
2300
|
+
"daddu $13, $10, %[iStride] \n\t"
|
|
2301
|
+
"daddu $8, $13, %[iStride] \n\t"
|
|
2302
|
+
"gsswlc1 $f20, 0x3($13) \n\t"
|
|
2303
|
+
"gsswrc1 $f20, 0x0($13) \n\t"
|
|
2304
|
+
"daddu $13, $8, %[iStride] \n\t"
|
|
2305
|
+
"gsswlc1 $f4, 0x3($8) \n\t"
|
|
2306
|
+
"gsswrc1 $f4, 0x0($8) \n\t"
|
|
2307
|
+
"gsswlc1 $f24, 0x3($13) \n\t"
|
|
2308
|
+
"gsswrc1 $f24, 0x0($13) \n\t"
|
|
2309
|
+
"gsswlc1 $f2, 0x3(%[pPixCb]) \n\t"
|
|
2310
|
+
"gsswrc1 $f2, 0x0(%[pPixCb]) \n\t"
|
|
2311
|
+
"daddu $12, %[pPixCb], %[iStride] \n\t"
|
|
2312
|
+
"gsswlc1 $f22, 0x3($12) \n\t"
|
|
2313
|
+
"gsswrc1 $f22, 0x0($12) \n\t"
|
|
2314
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2315
|
+
"gsswlc1 $f6, 0x3($12) \n\t"
|
|
2316
|
+
"gsswrc1 $f6, 0x0($12) \n\t"
|
|
2317
|
+
"daddu $12, $12, %[iStride] \n\t"
|
|
2318
|
+
"gsswlc1 $f26, 0x3($12) \n\t"
|
|
2319
|
+
"gsswrc1 $f26, 0x0($12) \n\t"
|
|
2320
|
+
"dsrl $f2, $f2, $f8 \n\t"
|
|
2321
|
+
"dsrl $f22, $f22, $f8 \n\t"
|
|
2322
|
+
"dsrl $f6, $f6, $f8 \n\t"
|
|
2323
|
+
"dsrl $f26, $f26, $f8 \n\t"
|
|
2324
|
+
"gsswlc1 $f2, 0x3(%[pPixCr]) \n\t"
|
|
2325
|
+
"gsswrc1 $f2, 0x0(%[pPixCr]) \n\t"
|
|
2326
|
+
"daddu $13, %[pPixCr], %[iStride] \n\t"
|
|
2327
|
+
"daddu $8, $13, %[iStride] \n\t"
|
|
2328
|
+
"gsswlc1 $f22, 0x3($13) \n\t"
|
|
2329
|
+
"gsswrc1 $f22, 0x0($13) \n\t"
|
|
2330
|
+
"daddu $13, $8, %[iStride] \n\t"
|
|
2331
|
+
"gsswlc1 $f6, 0x3($8) \n\t"
|
|
2332
|
+
"gsswrc1 $f6, 0x0($8) \n\t"
|
|
2333
|
+
"gsswlc1 $f26, 0x3($13) \n\t"
|
|
2334
|
+
"gsswrc1 $f26, 0x0($13) \n\t"
|
|
2335
|
+
: [pPixCb]"+&r"((unsigned char *)pPixCb), [pPixCr]"+&r"((unsigned char *)pPixCr)
|
|
2336
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha),
|
|
2337
|
+
[iBeta]"r"(iBeta), [tmp]"r"((unsigned char *)tmp)
|
|
2338
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$13", "$f0", "$f2", "$f4",
|
|
2339
|
+
"$f6", "$f8", "$f10", "$f12", "$f14", "$f16", "$f18", "$f20", "$f22",
|
|
2340
|
+
"$f24", "$f26", "$f28", "$f30"
|
|
2341
|
+
);
|
|
2342
|
+
RECOVER_REG;
|
|
2343
|
+
}
|
|
2344
|
+
|
|
2345
|
+
void DeblockChromaLt4H_mmi(uint8_t *pPixCb, uint8_t *pPixCr, int32_t iStride,
|
|
2346
|
+
int32_t iAlpha, int32_t iBeta, int8_t *pTC) {
|
|
2347
|
+
unsigned char tmp[320] __attribute__((aligned(32)));
|
|
2348
|
+
BACKUP_REG;
|
|
2349
|
+
__asm__ volatile (
|
|
2350
|
+
".set arch=loongson3a \n\t"
|
|
2351
|
+
"daddiu %[pPixCb], %[pPixCb], -0x2 \n\t"
|
|
2352
|
+
"daddiu %[pPixCr], %[pPixCr], -0x2 \n\t"
|
|
2353
|
+
"daddu $8, %[pPixCb], %[iStride] \n\t"
|
|
2354
|
+
"gsldlc1 $f0, 0x7(%[pPixCb]) \n\t"
|
|
2355
|
+
"gsldlc1 $f4, 0x7($8) \n\t"
|
|
2356
|
+
"gsldrc1 $f0, 0x0(%[pPixCb]) \n\t"
|
|
2357
|
+
"gsldrc1 $f4, 0x0($8) \n\t"
|
|
2358
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2359
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2360
|
+
"gsldlc1 $f8, 0x7($9) \n\t"
|
|
2361
|
+
"gsldlc1 $f12, 0x7($8) \n\t"
|
|
2362
|
+
"gsldrc1 $f8, 0x0($9) \n\t"
|
|
2363
|
+
"gsldrc1 $f12, 0x0($8) \n\t"
|
|
2364
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2365
|
+
|
|
2366
|
+
"daddu $10, %[pPixCr], %[iStride] \n\t"
|
|
2367
|
+
"gsldlc1 $f16, 0x7(%[pPixCr]) \n\t"
|
|
2368
|
+
"gsldlc1 $f20, 0x7($10) \n\t"
|
|
2369
|
+
"gsldrc1 $f16, 0x0(%[pPixCr]) \n\t"
|
|
2370
|
+
"gsldrc1 $f20, 0x0($10) \n\t"
|
|
2371
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2372
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2373
|
+
"gsldlc1 $f24, 0x7($11) \n\t"
|
|
2374
|
+
"gsldlc1 $f28, 0x7($10) \n\t"
|
|
2375
|
+
"gsldrc1 $f24, 0x0($11) \n\t"
|
|
2376
|
+
"gsldrc1 $f28, 0x0($10) \n\t"
|
|
2377
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2378
|
+
|
|
2379
|
+
"punpcklwd $f0, $f0, $f16 \n\t"
|
|
2380
|
+
"punpcklwd $f4, $f4, $f20 \n\t"
|
|
2381
|
+
"punpcklwd $f8, $f8, $f24 \n\t"
|
|
2382
|
+
"punpcklwd $f12, $f12, $f28 \n\t"
|
|
2383
|
+
"gsldlc1 $f16, 0x7($9) \n\t"
|
|
2384
|
+
"gsldlc1 $f20, 0x7($11) \n\t"
|
|
2385
|
+
"gsldrc1 $f16, 0x0($9) \n\t"
|
|
2386
|
+
"gsldrc1 $f20, 0x0($11) \n\t"
|
|
2387
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2388
|
+
"mov.d $f2, $f16 \n\t"
|
|
2389
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2390
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2391
|
+
"gsldlc1 $f16, 0x7($8) \n\t"
|
|
2392
|
+
"gsldlc1 $f20, 0x7($10) \n\t"
|
|
2393
|
+
"gsldrc1 $f16, 0x0($8) \n\t"
|
|
2394
|
+
"gsldrc1 $f20, 0x0($10) \n\t"
|
|
2395
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2396
|
+
"mov.d $f6, $f16 \n\t"
|
|
2397
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2398
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2399
|
+
|
|
2400
|
+
"gsldlc1 $f16, 0x7($9) \n\t"
|
|
2401
|
+
"gsldlc1 $f20, 0x7($11) \n\t"
|
|
2402
|
+
"gsldrc1 $f16, 0x0($9) \n\t"
|
|
2403
|
+
"gsldrc1 $f20, 0x0($11) \n\t"
|
|
2404
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2405
|
+
"mov.d $f10, $f16 \n\t"
|
|
2406
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2407
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2408
|
+
|
|
2409
|
+
"gsldlc1 $f16, 0x7($8) \n\t"
|
|
2410
|
+
"gsldlc1 $f20, 0x7($10) \n\t"
|
|
2411
|
+
"gsldrc1 $f16, 0x0($8) \n\t"
|
|
2412
|
+
"gsldrc1 $f20, 0x0($10) \n\t"
|
|
2413
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2414
|
+
"mov.d $f14, $f16 \n\t"
|
|
2415
|
+
|
|
2416
|
+
"punpcklbh $f24, $f2, $f6 \n\t"
|
|
2417
|
+
"punpckhbh $f26, $f2, $f6 \n\t"
|
|
2418
|
+
"punpckhbh $f2, $f0, $f4 \n\t"
|
|
2419
|
+
"punpcklbh $f0, $f0, $f4 \n\t"
|
|
2420
|
+
"punpcklbh $f28, $f10, $f14 \n\t"
|
|
2421
|
+
"punpckhbh $f30, $f10, $f14 \n\t"
|
|
2422
|
+
"punpckhbh $f10, $f8, $f12 \n\t"
|
|
2423
|
+
"punpcklbh $f8, $f8, $f12 \n\t"
|
|
2424
|
+
|
|
2425
|
+
"punpcklhw $f16, $f2, $f10 \n\t"
|
|
2426
|
+
"punpckhhw $f18, $f2, $f10 \n\t"
|
|
2427
|
+
"punpckhhw $f2, $f0, $f8 \n\t"
|
|
2428
|
+
"punpcklhw $f0, $f0, $f8 \n\t"
|
|
2429
|
+
"punpcklhw $f20, $f26, $f30 \n\t"
|
|
2430
|
+
"punpckhhw $f22, $f26, $f30 \n\t"
|
|
2431
|
+
"punpckhhw $f26, $f24, $f28 \n\t"
|
|
2432
|
+
"punpcklhw $f24, $f24, $f28 \n\t"
|
|
2433
|
+
|
|
2434
|
+
"punpcklwd $f4, $f2, $f26 \n\t"
|
|
2435
|
+
"punpckhwd $f6, $f2, $f26 \n\t"
|
|
2436
|
+
"punpckhwd $f2, $f0, $f24 \n\t"
|
|
2437
|
+
"punpcklwd $f0, $f0, $f24 \n\t"
|
|
2438
|
+
"punpcklwd $f8, $f18, $f22 \n\t"
|
|
2439
|
+
"punpckhwd $f10, $f18, $f22 \n\t"
|
|
2440
|
+
"punpckhwd $f18, $f16, $f20 \n\t"
|
|
2441
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2442
|
+
|
|
2443
|
+
"mov.d $f20, $f2 \n\t"
|
|
2444
|
+
"mov.d $f22, $f18 \n\t"
|
|
2445
|
+
"mov.d $f2, $f16 \n\t"
|
|
2446
|
+
"mov.d $f24, $f6 \n\t"
|
|
2447
|
+
"mov.d $f26, $f10 \n\t"
|
|
2448
|
+
"mov.d $f6, $f8 \n\t"
|
|
2449
|
+
"daddiu $11, %[tmp], 0x70 \n\t"
|
|
2450
|
+
|
|
2451
|
+
"gssqc1 $f2, $f0, 0x0($11) \n\t"
|
|
2452
|
+
"gssqc1 $f22, $f20, 0x10($11) \n\t"
|
|
2453
|
+
"gssqc1 $f6, $f4, 0x20($11) \n\t"
|
|
2454
|
+
"gssqc1 $f26, $f24, 0x30($11) \n\t"
|
|
2455
|
+
|
|
2456
|
+
"lb $8, 0x3(%[pTC]) \n\t"
|
|
2457
|
+
"lb $9, 0x2(%[pTC]) \n\t"
|
|
2458
|
+
"lb $10, 0x1(%[pTC]) \n\t"
|
|
2459
|
+
"lb $11, 0x0(%[pTC]) \n\t"
|
|
2460
|
+
|
|
2461
|
+
"and $12, $8, 0xFFFF \n\t"
|
|
2462
|
+
"dmtc1 $12, $f8 \n\t"
|
|
2463
|
+
|
|
2464
|
+
"and $9, $9, 0xFFFF \n\t"
|
|
2465
|
+
"dmtc1 $9, $f12 \n\t"
|
|
2466
|
+
"mov.d $f16, $f12 \n\t"
|
|
2467
|
+
|
|
2468
|
+
"and $9, $10, 0xFFFF \n\t"
|
|
2469
|
+
"dmtc1 $9, $f20 \n\t"
|
|
2470
|
+
"xor $f0, $f0, $f0 \n\t"
|
|
2471
|
+
"mov.d $f24, $f20 \n\t"
|
|
2472
|
+
"and $9, $11, 0xFFFF \n\t"
|
|
2473
|
+
"punpcklhw $f24, $f24, $f8 \n\t"
|
|
2474
|
+
|
|
2475
|
+
"mov.d $f4, $f8 \n\t"
|
|
2476
|
+
"dmtc1 $9, $f28 \n\t"
|
|
2477
|
+
"mov.d $f0, $f28 \n\t"
|
|
2478
|
+
|
|
2479
|
+
"punpcklhw $f28, $f28, $f12 \n\t"
|
|
2480
|
+
"punpcklhw $f20, $f20, $f4 \n\t"
|
|
2481
|
+
"xor $f4, $f4, $f4 \n\t"
|
|
2482
|
+
"xor $f6, $f6, $f6 \n\t"
|
|
2483
|
+
"punpcklhw $f28, $f28, $f20 \n\t"
|
|
2484
|
+
"gslqc1 $f22, $f20, 0xA0(%[tmp]) \n\t"
|
|
2485
|
+
"punpcklhw $f0, $f0, $f16 \n\t"
|
|
2486
|
+
"punpcklhw $f0, $f0, $f24 \n\t"
|
|
2487
|
+
|
|
2488
|
+
"gslqc1 $f26, $f24, 0x70(%[tmp]) \n\t"
|
|
2489
|
+
"punpckhhw $f2, $f0, $f28 \n\t"
|
|
2490
|
+
"punpcklhw $f0, $f0, $f28 \n\t"
|
|
2491
|
+
"gslqc1 $f30, $f28, 0x80(%[tmp]) \n\t"
|
|
2492
|
+
"psubh $f8, $f4, $f0 \n\t"
|
|
2493
|
+
"psubh $f10, $f6, $f2 \n\t"
|
|
2494
|
+
"gssqc1 $f10, $f8, 0xD0(%[tmp]) \n\t"
|
|
2495
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
2496
|
+
"punpcklhw $f12, $f8, $f8 \n\t"
|
|
2497
|
+
"punpcklwd $f16, $f12, $f12 \n\t"
|
|
2498
|
+
"mov.d $f18, $f16 \n\t"
|
|
2499
|
+
|
|
2500
|
+
"dmtc1 %[iBeta], $f8 \n\t"
|
|
2501
|
+
"punpcklhw $f12, $f8, $f8 \n\t"
|
|
2502
|
+
"punpcklwd $f8, $f12, $f12 \n\t"
|
|
2503
|
+
"mov.d $f10, $f8 \n\t"
|
|
2504
|
+
|
|
2505
|
+
"gslqc1 $f14, $f12, 0x90(%[tmp]) \n\t"
|
|
2506
|
+
"gssqc1 $f10, $f8, 0x50(%[tmp]) \n\t"
|
|
2507
|
+
"punpckhbh $f10, $f24, $f4 \n\t"
|
|
2508
|
+
"punpcklbh $f8, $f24, $f4 \n\t"
|
|
2509
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
2510
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
2511
|
+
|
|
2512
|
+
"gssqc1 $f10, $f8, 0x40(%[tmp]) \n\t"
|
|
2513
|
+
"gssqc1 $f26, $f24, 0xB0(%[tmp]) \n\t"
|
|
2514
|
+
"gslqc1 $f26, $f24, 0x90(%[tmp]) \n\t"
|
|
2515
|
+
"punpcklbh $f8, $f28, $f4 \n\t"
|
|
2516
|
+
"punpckhbh $f10, $f28, $f4 \n\t"
|
|
2517
|
+
"punpcklbh $f28, $f30, $f6 \n\t"
|
|
2518
|
+
"punpckhbh $f30, $f30, $f6 \n\t"
|
|
2519
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
2520
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
2521
|
+
"punpckhbh $f14, $f12, $f4 \n\t"
|
|
2522
|
+
"punpcklbh $f12, $f12, $f4 \n\t"
|
|
2523
|
+
"punpckhbh $f22, $f20, $f4 \n\t"
|
|
2524
|
+
"punpcklbh $f20, $f20, $f4 \n\t"
|
|
2525
|
+
"gssqc1 $f30, $f28, 0xF0(%[tmp]) \n\t"
|
|
2526
|
+
"gssqc1 $f26, $f24, 0xC0(%[tmp]) \n\t"
|
|
2527
|
+
"gslqc1 $f26, $f24, 0xA0(%[tmp]) \n\t"
|
|
2528
|
+
"punpcklbh $f24, $f26, $f6 \n\t"
|
|
2529
|
+
"punpckhbh $f26, $f26, $f6 \n\t"
|
|
2530
|
+
|
|
2531
|
+
"dli $13, 0x4 \n\t"
|
|
2532
|
+
"gssqc1 $f26, $f24, 0xE0(%[tmp]) \n\t"
|
|
2533
|
+
"dmtc1 $13, $f24 \n\t"
|
|
2534
|
+
"punpcklhw $f28, $f24, $f24 \n\t"
|
|
2535
|
+
"punpcklwd $f24, $f28, $f28 \n\t"
|
|
2536
|
+
"mov.d $f26, $f24 \n\t"
|
|
2537
|
+
"dli $12, 0x2 \n\t"
|
|
2538
|
+
"dli $13, 0x3 \n\t"
|
|
2539
|
+
|
|
2540
|
+
"gssqc1 $f2, $f0, 0x20(%[tmp]) \n\t"
|
|
2541
|
+
"dmfc1 %[iAlpha], $f0 \n\t"
|
|
2542
|
+
"dmfc1 %[iBeta], $f2 \n\t"
|
|
2543
|
+
"gssqc1 $f26, $f24, 0x30(%[tmp]) \n\t"
|
|
2544
|
+
"gslqc1 $f30, $f28, 0x40(%[tmp]) \n\t"
|
|
2545
|
+
"psubh $f28, $f28, $f20 \n\t"
|
|
2546
|
+
"psubh $f30, $f30, $f22 \n\t"
|
|
2547
|
+
"pcmpgth $f24, $f0, $f4 \n\t"
|
|
2548
|
+
"pcmpgth $f26, $f2, $f6 \n\t"
|
|
2549
|
+
|
|
2550
|
+
"dmtc1 $12, $f0 \n\t"
|
|
2551
|
+
"dmtc1 $13, $f2 \n\t"
|
|
2552
|
+
"gssqc1 $f26, $f24, 0x60(%[tmp]) \n\t"
|
|
2553
|
+
"gslqc1 $f6, $f4, 0xD0(%[tmp]) \n\t"
|
|
2554
|
+
"psubh $f24, $f12, $f8 \n\t"
|
|
2555
|
+
"psubh $f26, $f14, $f10 \n\t"
|
|
2556
|
+
"psllh $f24, $f24, $f0 \n\t"
|
|
2557
|
+
"psllh $f26, $f26, $f0 \n\t"
|
|
2558
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
2559
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
2560
|
+
"gslqc1 $f30, $f28, 0x30(%[tmp]) \n\t"
|
|
2561
|
+
"paddh $f24, $f24, $f28 \n\t"
|
|
2562
|
+
"paddh $f26, $f26, $f30 \n\t"
|
|
2563
|
+
"psrah $f24, $f24, $f2 \n\t"
|
|
2564
|
+
"psrah $f26, $f26, $f2 \n\t"
|
|
2565
|
+
"pmaxsh $f4, $f4, $f24 \n\t"
|
|
2566
|
+
"pmaxsh $f6, $f6, $f26 \n\t"
|
|
2567
|
+
|
|
2568
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
2569
|
+
"gslqc1 $f26, $f24, 0x20(%[tmp]) \n\t"
|
|
2570
|
+
"pminsh $f24, $f24, $f4 \n\t"
|
|
2571
|
+
"pminsh $f26, $f26, $f6 \n\t"
|
|
2572
|
+
|
|
2573
|
+
"gssqc1 $f26, $f24, 0x20(%[tmp]) \n\t"
|
|
2574
|
+
"psubh $f4, $f8, $f12 \n\t"
|
|
2575
|
+
"psubh $f6, $f10, $f14 \n\t"
|
|
2576
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f0, $f2)
|
|
2577
|
+
"pcmpgth $f24, $f16, $f4 \n\t"
|
|
2578
|
+
"pcmpgth $f26, $f18, $f6 \n\t"
|
|
2579
|
+
"gslqc1 $f6, $f4, 0x40(%[tmp]) \n\t"
|
|
2580
|
+
"psubh $f4, $f4, $f8 \n\t"
|
|
2581
|
+
"psubh $f6, $f6, $f10 \n\t"
|
|
2582
|
+
WELS_AbsH($f4, $f6, $f4, $f6, $f0, $f2)
|
|
2583
|
+
"pcmpgth $f28, $f28, $f4 \n\t"
|
|
2584
|
+
"pcmpgth $f30, $f30, $f6 \n\t"
|
|
2585
|
+
|
|
2586
|
+
"gslqc1 $f6, $f4, 0x50(%[tmp]) \n\t"
|
|
2587
|
+
"and $f24, $f24, $f28 \n\t"
|
|
2588
|
+
"and $f26, $f26, $f30 \n\t"
|
|
2589
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
2590
|
+
"psubh $f20, $f20, $f12 \n\t"
|
|
2591
|
+
"psubh $f22, $f22, $f14 \n\t"
|
|
2592
|
+
WELS_AbsH($f20, $f22, $f20, $f22, $f0, $f2)
|
|
2593
|
+
"pcmpgth $f4, $f4, $f20 \n\t"
|
|
2594
|
+
"pcmpgth $f6, $f6, $f22 \n\t"
|
|
2595
|
+
|
|
2596
|
+
"gslqc1 $f22, $f20, 0xB0(%[tmp]) \n\t"
|
|
2597
|
+
"gslqc1 $f2, $f0, 0xE0(%[tmp]) \n\t"
|
|
2598
|
+
"psubh $f20, $f20, $f0 \n\t"
|
|
2599
|
+
"psubh $f22, $f22, $f2 \n\t"
|
|
2600
|
+
"and $f24, $f24, $f4 \n\t"
|
|
2601
|
+
"and $f26, $f26, $f6 \n\t"
|
|
2602
|
+
"gslqc1 $f2, $f0, 0x60(%[tmp]) \n\t"
|
|
2603
|
+
"and $f24, $f24, $f0 \n\t"
|
|
2604
|
+
"and $f26, $f26, $f2 \n\t"
|
|
2605
|
+
|
|
2606
|
+
"gslqc1 $f6, $f4, 0x20(%[tmp]) \n\t"
|
|
2607
|
+
"and $f4, $f4, $f24 \n\t"
|
|
2608
|
+
"and $f6, $f6, $f26 \n\t"
|
|
2609
|
+
"gslqc1 $f26, $f24, 0xC0(%[tmp]) \n\t"
|
|
2610
|
+
"gssqc1 $f6, $f4, 0x40(%[tmp]) \n\t"
|
|
2611
|
+
"gslqc1 $f6, $f4, 0xF0(%[tmp]) \n\t"
|
|
2612
|
+
|
|
2613
|
+
"dmtc1 $12, $f0 \n\t"
|
|
2614
|
+
"psubh $f24, $f24, $f4 \n\t"
|
|
2615
|
+
"psubh $f26, $f26, $f6 \n\t"
|
|
2616
|
+
"psllh $f24, $f24, $f0 \n\t"
|
|
2617
|
+
"psllh $f26, $f26, $f0 \n\t"
|
|
2618
|
+
"paddh $f24, $f24, $f20 \n\t"
|
|
2619
|
+
"paddh $f26, $f26, $f22 \n\t"
|
|
2620
|
+
"gslqc1 $f2, $f0, 0x30(%[tmp]) \n\t"
|
|
2621
|
+
"paddh $f24, $f24, $f0 \n\t"
|
|
2622
|
+
"paddh $f26, $f26, $f2 \n\t"
|
|
2623
|
+
"dmtc1 %[iBeta], $f2 \n\t"
|
|
2624
|
+
|
|
2625
|
+
"dmtc1 $13, $f0 \n\t"
|
|
2626
|
+
"gslqc1 $f22, $f20, 0xD0(%[tmp]) \n\t"
|
|
2627
|
+
"psrah $f24, $f24, $f0 \n\t"
|
|
2628
|
+
"psrah $f26, $f26, $f0 \n\t"
|
|
2629
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
2630
|
+
"pmaxsh $f20, $f20, $f24 \n\t"
|
|
2631
|
+
"pmaxsh $f22, $f22, $f26 \n\t"
|
|
2632
|
+
"pminsh $f0, $f0, $f20 \n\t"
|
|
2633
|
+
"pminsh $f2, $f2, $f22 \n\t"
|
|
2634
|
+
|
|
2635
|
+
"dmfc1 %[iAlpha], $f0 \n\t"
|
|
2636
|
+
"dmfc1 %[iBeta], $f2 \n\t"
|
|
2637
|
+
"gslqc1 $f22, $f20, 0xC0(%[tmp]) \n\t"
|
|
2638
|
+
"psubh $f24, $f4, $f20 \n\t"
|
|
2639
|
+
"psubh $f26, $f6, $f22 \n\t"
|
|
2640
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f0, $f2)
|
|
2641
|
+
"pcmpgth $f16, $f16, $f24 \n\t"
|
|
2642
|
+
"pcmpgth $f18, $f18, $f26 \n\t"
|
|
2643
|
+
|
|
2644
|
+
"gslqc1 $f26, $f24, 0xB0(%[tmp]) \n\t"
|
|
2645
|
+
"psubh $f24, $f24, $f4 \n\t"
|
|
2646
|
+
"psubh $f26, $f26, $f6 \n\t"
|
|
2647
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f0, $f2)
|
|
2648
|
+
"pcmpgth $f28, $f28, $f24 \n\t"
|
|
2649
|
+
"pcmpgth $f30, $f30, $f26 \n\t"
|
|
2650
|
+
|
|
2651
|
+
"gslqc1 $f26, $f24, 0xE0(%[tmp]) \n\t"
|
|
2652
|
+
"and $f16, $f16, $f28 \n\t"
|
|
2653
|
+
"and $f18, $f18, $f30 \n\t"
|
|
2654
|
+
|
|
2655
|
+
"gslqc1 $f30, $f28, 0x50(%[tmp]) \n\t"
|
|
2656
|
+
"psubh $f24, $f24, $f20 \n\t"
|
|
2657
|
+
"psubh $f26, $f26, $f22 \n\t"
|
|
2658
|
+
WELS_AbsH($f24, $f26, $f24, $f26, $f0, $f2)
|
|
2659
|
+
"pcmpgth $f28, $f28, $f24 \n\t"
|
|
2660
|
+
"pcmpgth $f30, $f30, $f26 \n\t"
|
|
2661
|
+
"and $f16, $f16, $f28 \n\t"
|
|
2662
|
+
"and $f18, $f18, $f30 \n\t"
|
|
2663
|
+
"gslqc1 $f30, $f28, 0x60(%[tmp]) \n\t"
|
|
2664
|
+
"dmtc1 %[iAlpha], $f0 \n\t"
|
|
2665
|
+
"dmtc1 %[iBeta], $f2 \n\t"
|
|
2666
|
+
"and $f16, $f16, $f28 \n\t"
|
|
2667
|
+
"and $f18, $f18, $f30 \n\t"
|
|
2668
|
+
"and $f0, $f0, $f16 \n\t"
|
|
2669
|
+
"and $f2, $f2, $f18 \n\t"
|
|
2670
|
+
|
|
2671
|
+
"gslqc1 $f18, $f16, 0x40(%[tmp]) \n\t"
|
|
2672
|
+
"paddh $f8, $f8, $f16 \n\t"
|
|
2673
|
+
"paddh $f10, $f10, $f18 \n\t"
|
|
2674
|
+
"paddh $f4, $f4, $f0 \n\t"
|
|
2675
|
+
"paddh $f6, $f6, $f2 \n\t"
|
|
2676
|
+
"psubh $f12, $f12, $f16 \n\t"
|
|
2677
|
+
"psubh $f14, $f14, $f18 \n\t"
|
|
2678
|
+
"psubh $f20, $f20, $f0 \n\t"
|
|
2679
|
+
"psubh $f22, $f22, $f2 \n\t"
|
|
2680
|
+
"packushb $f8, $f8, $f10 \n\t"
|
|
2681
|
+
"packushb $f10, $f4, $f6 \n\t"
|
|
2682
|
+
"packushb $f12, $f12, $f14 \n\t"
|
|
2683
|
+
"packushb $f14, $f20, $f22 \n\t"
|
|
2684
|
+
|
|
2685
|
+
"gssqc1 $f10, $f8, 0x80(%[tmp]) \n\t"
|
|
2686
|
+
"gssqc1 $f14, $f12, 0x90(%[tmp]) \n\t"
|
|
2687
|
+
"daddiu $11, %[tmp], 0x70 \n\t"
|
|
2688
|
+
|
|
2689
|
+
"gslqc1 $f2, $f0, 0x0($11) \n\t"
|
|
2690
|
+
"gslqc1 $f6, $f4, 0x10($11) \n\t"
|
|
2691
|
+
"gslqc1 $f10, $f8, 0x20($11) \n\t"
|
|
2692
|
+
"gslqc1 $f14, $f12, 0x30($11) \n\t"
|
|
2693
|
+
|
|
2694
|
+
"punpcklbh $f24, $f2, $f6 \n\t"
|
|
2695
|
+
"punpckhbh $f26, $f2, $f6 \n\t"
|
|
2696
|
+
"punpckhbh $f2, $f0, $f4 \n\t"
|
|
2697
|
+
"punpcklbh $f0, $f0, $f4 \n\t"
|
|
2698
|
+
|
|
2699
|
+
"punpcklbh $f28, $f10, $f14 \n\t"
|
|
2700
|
+
"punpckhbh $f30, $f10, $f14 \n\t"
|
|
2701
|
+
"punpckhbh $f10, $f8, $f12 \n\t"
|
|
2702
|
+
"punpcklbh $f8, $f8, $f12 \n\t"
|
|
2703
|
+
|
|
2704
|
+
"punpcklhw $f16, $f2, $f10 \n\t"
|
|
2705
|
+
"punpckhhw $f18, $f2, $f10 \n\t"
|
|
2706
|
+
"punpckhhw $f2, $f0, $f8 \n\t"
|
|
2707
|
+
"punpcklhw $f0, $f0, $f8 \n\t"
|
|
2708
|
+
"punpcklhw $f20, $f26, $f30 \n\t"
|
|
2709
|
+
"punpckhhw $f22, $f26, $f30 \n\t"
|
|
2710
|
+
"punpckhhw $f26, $f24, $f28 \n\t"
|
|
2711
|
+
"punpcklhw $f24, $f24, $f28 \n\t"
|
|
2712
|
+
|
|
2713
|
+
"punpcklwd $f4, $f2, $f26 \n\t"
|
|
2714
|
+
"punpckhwd $f6, $f2, $f26 \n\t"
|
|
2715
|
+
"punpckhwd $f2, $f0, $f24 \n\t"
|
|
2716
|
+
"punpcklwd $f0, $f0, $f24 \n\t"
|
|
2717
|
+
"punpcklwd $f8, $f18, $f22 \n\t"
|
|
2718
|
+
"punpckhwd $f10, $f18, $f22 \n\t"
|
|
2719
|
+
"punpckhwd $f18, $f16, $f20 \n\t"
|
|
2720
|
+
"punpcklwd $f16, $f16, $f20 \n\t"
|
|
2721
|
+
|
|
2722
|
+
"mov.d $f20, $f2 \n\t"
|
|
2723
|
+
"mov.d $f22, $f18 \n\t"
|
|
2724
|
+
"mov.d $f2, $f16 \n\t"
|
|
2725
|
+
"mov.d $f24, $f6 \n\t"
|
|
2726
|
+
"mov.d $f26, $f10 \n\t"
|
|
2727
|
+
"mov.d $f6, $f8 \n\t"
|
|
2728
|
+
|
|
2729
|
+
"dli %[iAlpha], 0x20 \n\t"
|
|
2730
|
+
"daddu $8, %[pPixCb], %[iStride] \n\t"
|
|
2731
|
+
"gsswlc1 $f0, 0x3(%[pPixCb]) \n\t"
|
|
2732
|
+
"gsswlc1 $f20, 0x3($8) \n\t"
|
|
2733
|
+
"gsswrc1 $f0, 0x0(%[pPixCb]) \n\t"
|
|
2734
|
+
"gsswrc1 $f20, 0x0($8) \n\t"
|
|
2735
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2736
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2737
|
+
"gsswlc1 $f4, 0x3($9) \n\t"
|
|
2738
|
+
"gsswlc1 $f24, 0x3($8) \n\t"
|
|
2739
|
+
"gsswrc1 $f4, 0x0($9) \n\t"
|
|
2740
|
+
"gsswrc1 $f24, 0x0($8) \n\t"
|
|
2741
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2742
|
+
"dmtc1 %[iAlpha], $f8 \n\t"
|
|
2743
|
+
|
|
2744
|
+
"dsrl $f0, $f0, $f8 \n\t"
|
|
2745
|
+
"dsrl $f20, $f20, $f8 \n\t"
|
|
2746
|
+
"dsrl $f4, $f4, $f8 \n\t"
|
|
2747
|
+
"dsrl $f24, $f24, $f8 \n\t"
|
|
2748
|
+
"daddu $10, %[pPixCr], %[iStride] \n\t"
|
|
2749
|
+
"gsswlc1 $f0, 0x3(%[pPixCr]) \n\t"
|
|
2750
|
+
"gsswlc1 $f20, 0x3($10) \n\t"
|
|
2751
|
+
"gsswrc1 $f0, 0x0(%[pPixCr]) \n\t"
|
|
2752
|
+
"gsswrc1 $f20, 0x0($10) \n\t"
|
|
2753
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2754
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2755
|
+
"gsswlc1 $f4, 0x3($11) \n\t"
|
|
2756
|
+
"gsswlc1 $f24, 0x3($10) \n\t"
|
|
2757
|
+
"gsswrc1 $f4, 0x0($11) \n\t"
|
|
2758
|
+
"gsswrc1 $f24, 0x0($10) \n\t"
|
|
2759
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2760
|
+
|
|
2761
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2762
|
+
"gsswlc1 $f2, 0x3($9) \n\t"
|
|
2763
|
+
"gsswlc1 $f22, 0x3($8) \n\t"
|
|
2764
|
+
"gsswrc1 $f2, 0x0($9) \n\t"
|
|
2765
|
+
"gsswrc1 $f22, 0x0($8) \n\t"
|
|
2766
|
+
"daddu $9, $8, %[iStride] \n\t"
|
|
2767
|
+
"daddu $8, $9, %[iStride] \n\t"
|
|
2768
|
+
"gsswlc1 $f6, 0x3($9) \n\t"
|
|
2769
|
+
"gsswlc1 $f26, 0x3($8) \n\t"
|
|
2770
|
+
"gsswrc1 $f6, 0x0($9) \n\t"
|
|
2771
|
+
"gsswrc1 $f26, 0x0($8) \n\t"
|
|
2772
|
+
|
|
2773
|
+
"dsrl $f2, $f2, $f8 \n\t"
|
|
2774
|
+
"dsrl $f22, $f22, $f8 \n\t"
|
|
2775
|
+
"dsrl $f6, $f6, $f8 \n\t"
|
|
2776
|
+
"dsrl $f26, $f26, $f8 \n\t"
|
|
2777
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2778
|
+
"gsswlc1 $f2, 0x3($11) \n\t"
|
|
2779
|
+
"gsswlc1 $f22, 0x3($10) \n\t"
|
|
2780
|
+
"gsswrc1 $f2, 0x0($11) \n\t"
|
|
2781
|
+
"gsswrc1 $f22, 0x0($10) \n\t"
|
|
2782
|
+
"daddu $11, $10, %[iStride] \n\t"
|
|
2783
|
+
"daddu $10, $11, %[iStride] \n\t"
|
|
2784
|
+
"gsswlc1 $f6, 0x3($11) \n\t"
|
|
2785
|
+
"gsswlc1 $f26, 0x3($10) \n\t"
|
|
2786
|
+
"gsswrc1 $f6, 0x0($11) \n\t"
|
|
2787
|
+
"gsswrc1 $f26, 0x0($10) \n\t"
|
|
2788
|
+
: [pPixCb]"+&r"((unsigned char *)pPixCb), [pPixCr]"+&r"((unsigned char *)pPixCr)
|
|
2789
|
+
: [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha),
|
|
2790
|
+
[iBeta]"r"(iBeta), [tmp]"r"((unsigned char *)tmp), [pTC]"r"((char *)pTC)
|
|
2791
|
+
: "memory", "$8", "$9", "$10", "$11", "$12", "$13", "$f0", "$f2", "$f4",
|
|
2792
|
+
"$f6", "$f8", "$f10", "$f12","$f14", "$f16", "$f18", "$f20", "$f22", "$f24",
|
|
2793
|
+
"$f26", "$f28", "$f30"
|
|
2794
|
+
);
|
|
2795
|
+
RECOVER_REG;
|
|
2796
|
+
}
|
|
2797
|
+
|
|
2798
|
+
void WelsNonZeroCount_mmi(int8_t *pNonZeroCount) {
|
|
2799
|
+
__asm__ volatile(
|
|
2800
|
+
".set arch=loongson3a \n\t"
|
|
2801
|
+
"gsldlc1 $f0, 0x7(%[pNonZeroCount]) \n\t"
|
|
2802
|
+
"gsldlc1 $f2, 0xF(%[pNonZeroCount]) \n\t"
|
|
2803
|
+
"gsldlc1 $f4, 0x17(%[pNonZeroCount]) \n\t"
|
|
2804
|
+
"gsldrc1 $f4, 0x10(%[pNonZeroCount]) \n\t"
|
|
2805
|
+
"gsldrc1 $f0, 0x0(%[pNonZeroCount]) \n\t"
|
|
2806
|
+
"gsldrc1 $f2, 0x8(%[pNonZeroCount]) \n\t"
|
|
2807
|
+
"pcmpeqh $f8, $f8, $f8 \n\t"
|
|
2808
|
+
"dli $8, 0xF \n\t"
|
|
2809
|
+
"dmtc1 $8, $f6 \n\t"
|
|
2810
|
+
"psrlh $f8, $f8, $f6 \n\t"
|
|
2811
|
+
"packushb $f8, $f8, $f8 \n\t"
|
|
2812
|
+
|
|
2813
|
+
"pminub $f0, $f0, $f8 \n\t"
|
|
2814
|
+
"pminub $f2, $f2, $f8 \n\t"
|
|
2815
|
+
"pminub $f4, $f4, $f8 \n\t"
|
|
2816
|
+
"gssdlc1 $f0, 0x7(%[pNonZeroCount]) \n\t"
|
|
2817
|
+
"gssdlc1 $f2, 0xF(%[pNonZeroCount]) \n\t"
|
|
2818
|
+
"gssdlc1 $f4, 0x17(%[pNonZeroCount]) \n\t"
|
|
2819
|
+
"gssdrc1 $f0, 0x0(%[pNonZeroCount]) \n\t"
|
|
2820
|
+
"gssdrc1 $f2, 0x8(%[pNonZeroCount]) \n\t"
|
|
2821
|
+
"gssdrc1 $f4, 0x10(%[pNonZeroCount]) \n\t"
|
|
2822
|
+
:
|
|
2823
|
+
: [pNonZeroCount] "r"((unsigned char *)pNonZeroCount)
|
|
2824
|
+
: "memory", "$8", "$f0", "$f2", "$f4", "$f6", "$f8"
|
|
2825
|
+
);
|
|
2826
|
+
}
|