@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.
Files changed (340) hide show
  1. package/_openh264install/include/wels/codec_api.h +592 -0
  2. package/_openh264install/include/wels/codec_app_def.h +820 -0
  3. package/_openh264install/include/wels/codec_def.h +216 -0
  4. package/_openh264install/include/wels/codec_ver.h +15 -0
  5. package/_openh264install/lib/libopenh264.a +0 -0
  6. package/_openh264install/lib/pkgconfig/openh264.pc +10 -0
  7. package/binding.gyp +4 -0
  8. package/openh264/.reviewboardrc +3 -0
  9. package/openh264/.travis.yml +42 -0
  10. package/openh264/CODING_STYLE +8 -0
  11. package/openh264/CONTRIBUTORS +55 -0
  12. package/openh264/LICENSE +23 -0
  13. package/openh264/Makefile +408 -0
  14. package/openh264/README.md +197 -0
  15. package/openh264/RELEASES +467 -0
  16. package/openh264/SECURITY.md +16 -0
  17. package/openh264/autotest/codec_comparision/run_MultiCodecComparision.sh +431 -0
  18. package/openh264/autotest/performanceTest/android/run_AutoTest_android.sh +180 -0
  19. package/openh264/autotest/performanceTest/ios/fruitstrap +0 -0
  20. package/openh264/autotest/performanceTest/ios/iFileTransfer +0 -0
  21. package/openh264/autotest/performanceTest/ios/run_AutoTest_ios.sh +193 -0
  22. package/openh264/autotest/performanceTest/ios/uiascript.js +19 -0
  23. package/openh264/autotest/performanceTest/parsePerfData.sh +107 -0
  24. package/openh264/autotest/performanceTest/run_perfTest.sh +122 -0
  25. package/openh264/autotest/unitTest/android/run_AutoTest_android.sh +112 -0
  26. package/openh264/autotest/unitTest/ios/run_AutoTest_ios.sh +167 -0
  27. package/openh264/autotest/unitTest/run_ParseUTxml.sh +70 -0
  28. package/openh264/autotest/unitTest/run_unitTest.sh +98 -0
  29. package/openh264/build.gradle +30 -0
  30. package/openh264/code-coverage.sh +6 -0
  31. package/openh264/codec/api/meson.build +1 -0
  32. package/openh264/codec/api/wels/codec_api.h +592 -0
  33. package/openh264/codec/api/wels/codec_app_def.h +820 -0
  34. package/openh264/codec/api/wels/codec_def.h +216 -0
  35. package/openh264/codec/api/wels/codec_ver.h +15 -0
  36. package/openh264/codec/api/wels/meson.build +9 -0
  37. package/openh264/codec/common/arm/arm_arch_common_macro.S +83 -0
  38. package/openh264/codec/common/arm/copy_mb_neon.S +161 -0
  39. package/openh264/codec/common/arm/deblocking_neon.S +857 -0
  40. package/openh264/codec/common/arm/expand_picture_neon.S +154 -0
  41. package/openh264/codec/common/arm/intra_pred_common_neon.S +82 -0
  42. package/openh264/codec/common/arm/mc_neon.S +2283 -0
  43. package/openh264/codec/common/arm64/arm_aarch64_common.h +85 -0
  44. package/openh264/codec/common/arm64/arm_arch64_common_macro.S +79 -0
  45. package/openh264/codec/common/arm64/copy_mb_aarch64_neon.S +202 -0
  46. package/openh264/codec/common/arm64/deblocking_aarch64_neon.S +852 -0
  47. package/openh264/codec/common/arm64/expand_picture_aarch64_neon.S +150 -0
  48. package/openh264/codec/common/arm64/intra_pred_common_aarch64_neon.S +58 -0
  49. package/openh264/codec/common/arm64/mc_aarch64_neon.S +2614 -0
  50. package/openh264/codec/common/generate_version.sh +28 -0
  51. package/openh264/codec/common/inc/WelsList.h +287 -0
  52. package/openh264/codec/common/inc/WelsLock.h +97 -0
  53. package/openh264/codec/common/inc/WelsTask.h +75 -0
  54. package/openh264/codec/common/inc/WelsTaskThread.h +83 -0
  55. package/openh264/codec/common/inc/WelsThread.h +106 -0
  56. package/openh264/codec/common/inc/WelsThreadLib.h +151 -0
  57. package/openh264/codec/common/inc/WelsThreadPool.h +124 -0
  58. package/openh264/codec/common/inc/asmdefs_mmi.h +345 -0
  59. package/openh264/codec/common/inc/copy_mb.h +103 -0
  60. package/openh264/codec/common/inc/cpu.h +81 -0
  61. package/openh264/codec/common/inc/cpu_core.h +103 -0
  62. package/openh264/codec/common/inc/crt_util_safe_x.h +104 -0
  63. package/openh264/codec/common/inc/deblocking_common.h +124 -0
  64. package/openh264/codec/common/inc/expand_pic.h +107 -0
  65. package/openh264/codec/common/inc/golomb_common.h +166 -0
  66. package/openh264/codec/common/inc/intra_pred_common.h +86 -0
  67. package/openh264/codec/common/inc/loongson_intrinsics.h +1919 -0
  68. package/openh264/codec/common/inc/ls_defines.h +130 -0
  69. package/openh264/codec/common/inc/macros.h +329 -0
  70. package/openh264/codec/common/inc/mc.h +407 -0
  71. package/openh264/codec/common/inc/measure_time.h +88 -0
  72. package/openh264/codec/common/inc/memory_align.h +116 -0
  73. package/openh264/codec/common/inc/msa_macros.h +2393 -0
  74. package/openh264/codec/common/inc/sad_common.h +139 -0
  75. package/openh264/codec/common/inc/typedefs.h +86 -0
  76. package/openh264/codec/common/inc/utils.h +95 -0
  77. package/openh264/codec/common/inc/version.h +10 -0
  78. package/openh264/codec/common/inc/version_gen.h.template +6 -0
  79. package/openh264/codec/common/inc/welsCodecTrace.h +64 -0
  80. package/openh264/codec/common/inc/wels_common_defs.h +373 -0
  81. package/openh264/codec/common/inc/wels_const_common.h +63 -0
  82. package/openh264/codec/common/loongarch/copy_mb_lsx.c +194 -0
  83. package/openh264/codec/common/loongarch/deblock_lsx.c +1489 -0
  84. package/openh264/codec/common/loongarch/intra_pred_com_lasx.c +112 -0
  85. package/openh264/codec/common/loongarch/intra_pred_com_lsx.c +91 -0
  86. package/openh264/codec/common/loongarch/mc_chroma_lsx.c +192 -0
  87. package/openh264/codec/common/loongarch/mc_horver_lsx.c +654 -0
  88. package/openh264/codec/common/loongarch/satd_sad_lasx.c +568 -0
  89. package/openh264/codec/common/meson.build +84 -0
  90. package/openh264/codec/common/mips/copy_mb_mmi.c +477 -0
  91. package/openh264/codec/common/mips/copy_mb_msa.c +80 -0
  92. package/openh264/codec/common/mips/deblock_mmi.c +2826 -0
  93. package/openh264/codec/common/mips/deblock_msa.c +1024 -0
  94. package/openh264/codec/common/mips/expand_picture_mmi.c +673 -0
  95. package/openh264/codec/common/mips/intra_pred_com_mmi.c +548 -0
  96. package/openh264/codec/common/mips/satd_sad_mmi.c +2154 -0
  97. package/openh264/codec/common/targets.mk +136 -0
  98. package/openh264/codec/common/x86/asm_inc.asm +743 -0
  99. package/openh264/codec/common/x86/cpuid.asm +263 -0
  100. package/openh264/codec/common/x86/dct.asm +1036 -0
  101. package/openh264/codec/common/x86/deblock.asm +848 -0
  102. package/openh264/codec/common/x86/expand_picture.asm +728 -0
  103. package/openh264/codec/common/x86/intra_pred_com.asm +117 -0
  104. package/openh264/codec/common/x86/mb_copy.asm +615 -0
  105. package/openh264/codec/common/x86/mc_chroma.asm +313 -0
  106. package/openh264/codec/common/x86/mc_luma.asm +4490 -0
  107. package/openh264/codec/common/x86/satd_sad.asm +2734 -0
  108. package/openh264/codec/common/x86/vaa.asm +411 -0
  109. package/openh264/codec/console/common/inc/read_config.h +68 -0
  110. package/openh264/codec/console/common/meson.build +12 -0
  111. package/openh264/codec/console/common/targets.mk +20 -0
  112. package/openh264/codec/console/dec/inc/d3d9_utils.h +148 -0
  113. package/openh264/codec/console/dec/meson.build +15 -0
  114. package/openh264/codec/console/dec/targets.mk +20 -0
  115. package/openh264/codec/console/enc/meson.build +9 -0
  116. package/openh264/codec/console/enc/targets.mk +19 -0
  117. package/openh264/codec/console/meson.build +3 -0
  118. package/openh264/codec/decoder/core/arm/block_add_neon.S +129 -0
  119. package/openh264/codec/decoder/core/arm/intra_pred_neon.S +635 -0
  120. package/openh264/codec/decoder/core/arm64/block_add_aarch64_neon.S +131 -0
  121. package/openh264/codec/decoder/core/arm64/intra_pred_aarch64_neon.S +524 -0
  122. package/openh264/codec/decoder/core/inc/au_parser.h +199 -0
  123. package/openh264/codec/decoder/core/inc/bit_stream.h +62 -0
  124. package/openh264/codec/decoder/core/inc/cabac_decoder.h +111 -0
  125. package/openh264/codec/decoder/core/inc/deblocking.h +118 -0
  126. package/openh264/codec/decoder/core/inc/dec_frame.h +156 -0
  127. package/openh264/codec/decoder/core/inc/dec_golomb.h +344 -0
  128. package/openh264/codec/decoder/core/inc/decode_mb_aux.h +83 -0
  129. package/openh264/codec/decoder/core/inc/decode_slice.h +104 -0
  130. package/openh264/codec/decoder/core/inc/decoder.h +186 -0
  131. package/openh264/codec/decoder/core/inc/decoder_context.h +590 -0
  132. package/openh264/codec/decoder/core/inc/decoder_core.h +206 -0
  133. package/openh264/codec/decoder/core/inc/error_code.h +222 -0
  134. package/openh264/codec/decoder/core/inc/error_concealment.h +68 -0
  135. package/openh264/codec/decoder/core/inc/fmo.h +114 -0
  136. package/openh264/codec/decoder/core/inc/get_intra_predictor.h +191 -0
  137. package/openh264/codec/decoder/core/inc/manage_dec_ref.h +62 -0
  138. package/openh264/codec/decoder/core/inc/mb_cache.h +80 -0
  139. package/openh264/codec/decoder/core/inc/memmgr_nal_unit.h +66 -0
  140. package/openh264/codec/decoder/core/inc/mv_pred.h +195 -0
  141. package/openh264/codec/decoder/core/inc/nal_prefix.h +56 -0
  142. package/openh264/codec/decoder/core/inc/nalu.h +79 -0
  143. package/openh264/codec/decoder/core/inc/parameter_sets.h +217 -0
  144. package/openh264/codec/decoder/core/inc/parse_mb_syn_cabac.h +89 -0
  145. package/openh264/codec/decoder/core/inc/parse_mb_syn_cavlc.h +137 -0
  146. package/openh264/codec/decoder/core/inc/pic_queue.h +62 -0
  147. package/openh264/codec/decoder/core/inc/picture.h +112 -0
  148. package/openh264/codec/decoder/core/inc/rec_mb.h +104 -0
  149. package/openh264/codec/decoder/core/inc/slice.h +208 -0
  150. package/openh264/codec/decoder/core/inc/vlc_decoder.h +177 -0
  151. package/openh264/codec/decoder/core/inc/wels_common_basis.h +317 -0
  152. package/openh264/codec/decoder/core/inc/wels_const.h +65 -0
  153. package/openh264/codec/decoder/core/inc/wels_decoder_thread.h +170 -0
  154. package/openh264/codec/decoder/core/loongarch/mb_aux_lsx.c +296 -0
  155. package/openh264/codec/decoder/core/mips/dct_mmi.c +786 -0
  156. package/openh264/codec/decoder/core/x86/dct.asm +72 -0
  157. package/openh264/codec/decoder/core/x86/intra_pred.asm +1456 -0
  158. package/openh264/codec/decoder/meson.build +69 -0
  159. package/openh264/codec/decoder/plus/inc/welsDecoderExt.h +170 -0
  160. package/openh264/codec/decoder/targets.mk +115 -0
  161. package/openh264/codec/encoder/core/arm/intra_pred_neon.S +589 -0
  162. package/openh264/codec/encoder/core/arm/intra_pred_sad_3_opt_neon.S +735 -0
  163. package/openh264/codec/encoder/core/arm/memory_neon.S +62 -0
  164. package/openh264/codec/encoder/core/arm/pixel_neon.S +882 -0
  165. package/openh264/codec/encoder/core/arm/reconstruct_neon.S +903 -0
  166. package/openh264/codec/encoder/core/arm/svc_motion_estimation.S +366 -0
  167. package/openh264/codec/encoder/core/arm64/intra_pred_aarch64_neon.S +505 -0
  168. package/openh264/codec/encoder/core/arm64/intra_pred_sad_3_opt_aarch64_neon.S +611 -0
  169. package/openh264/codec/encoder/core/arm64/memory_aarch64_neon.S +63 -0
  170. package/openh264/codec/encoder/core/arm64/pixel_aarch64_neon.S +621 -0
  171. package/openh264/codec/encoder/core/arm64/reconstruct_aarch64_neon.S +717 -0
  172. package/openh264/codec/encoder/core/arm64/svc_motion_estimation_aarch64_neon.S +350 -0
  173. package/openh264/codec/encoder/core/inc/as264_common.h +138 -0
  174. package/openh264/codec/encoder/core/inc/au_set.h +152 -0
  175. package/openh264/codec/encoder/core/inc/deblocking.h +92 -0
  176. package/openh264/codec/encoder/core/inc/decode_mb_aux.h +112 -0
  177. package/openh264/codec/encoder/core/inc/dq_map.h +58 -0
  178. package/openh264/codec/encoder/core/inc/encode_mb_aux.h +199 -0
  179. package/openh264/codec/encoder/core/inc/encoder.h +148 -0
  180. package/openh264/codec/encoder/core/inc/encoder_context.h +240 -0
  181. package/openh264/codec/encoder/core/inc/extern.h +124 -0
  182. package/openh264/codec/encoder/core/inc/get_intra_predictor.h +186 -0
  183. package/openh264/codec/encoder/core/inc/mb_cache.h +141 -0
  184. package/openh264/codec/encoder/core/inc/md.h +181 -0
  185. package/openh264/codec/encoder/core/inc/mt_defs.h +91 -0
  186. package/openh264/codec/encoder/core/inc/mv_pred.h +191 -0
  187. package/openh264/codec/encoder/core/inc/nal_encap.h +143 -0
  188. package/openh264/codec/encoder/core/inc/param_svc.h +576 -0
  189. package/openh264/codec/encoder/core/inc/parameter_sets.h +181 -0
  190. package/openh264/codec/encoder/core/inc/paraset_strategy.h +310 -0
  191. package/openh264/codec/encoder/core/inc/picture.h +137 -0
  192. package/openh264/codec/encoder/core/inc/picture_handle.h +66 -0
  193. package/openh264/codec/encoder/core/inc/rc.h +284 -0
  194. package/openh264/codec/encoder/core/inc/ref_list_mgr_svc.h +148 -0
  195. package/openh264/codec/encoder/core/inc/sample.h +154 -0
  196. package/openh264/codec/encoder/core/inc/set_mb_syn_cabac.h +127 -0
  197. package/openh264/codec/encoder/core/inc/set_mb_syn_cavlc.h +92 -0
  198. package/openh264/codec/encoder/core/inc/slice.h +212 -0
  199. package/openh264/codec/encoder/core/inc/slice_multi_threading.h +97 -0
  200. package/openh264/codec/encoder/core/inc/stat.h +102 -0
  201. package/openh264/codec/encoder/core/inc/svc_base_layer_md.h +102 -0
  202. package/openh264/codec/encoder/core/inc/svc_enc_frame.h +136 -0
  203. package/openh264/codec/encoder/core/inc/svc_enc_golomb.h +123 -0
  204. package/openh264/codec/encoder/core/inc/svc_enc_macroblock.h +82 -0
  205. package/openh264/codec/encoder/core/inc/svc_enc_slice_segment.h +212 -0
  206. package/openh264/codec/encoder/core/inc/svc_encode_mb.h +63 -0
  207. package/openh264/codec/encoder/core/inc/svc_encode_slice.h +184 -0
  208. package/openh264/codec/encoder/core/inc/svc_mode_decision.h +94 -0
  209. package/openh264/codec/encoder/core/inc/svc_motion_estimate.h +369 -0
  210. package/openh264/codec/encoder/core/inc/svc_set_mb_syn.h +70 -0
  211. package/openh264/codec/encoder/core/inc/svc_set_mb_syn_cavlc.h +67 -0
  212. package/openh264/codec/encoder/core/inc/vlc_encoder.h +92 -0
  213. package/openh264/codec/encoder/core/inc/wels_common_basis.h +160 -0
  214. package/openh264/codec/encoder/core/inc/wels_const.h +174 -0
  215. package/openh264/codec/encoder/core/inc/wels_func_ptr_def.h +300 -0
  216. package/openh264/codec/encoder/core/inc/wels_preprocess.h +251 -0
  217. package/openh264/codec/encoder/core/inc/wels_task_base.h +79 -0
  218. package/openh264/codec/encoder/core/inc/wels_task_encoder.h +142 -0
  219. package/openh264/codec/encoder/core/inc/wels_task_management.h +135 -0
  220. package/openh264/codec/encoder/core/inc/wels_transpose_matrix.h +58 -0
  221. package/openh264/codec/encoder/core/loongarch/dct_lasx.c +452 -0
  222. package/openh264/codec/encoder/core/loongarch/get_intra_predictor_lsx.c +166 -0
  223. package/openh264/codec/encoder/core/loongarch/quant_lsx.c +153 -0
  224. package/openh264/codec/encoder/core/loongarch/sample_lasx.c +87 -0
  225. package/openh264/codec/encoder/core/loongarch/svc_motion_estimate_lsx.c +104 -0
  226. package/openh264/codec/encoder/core/mips/dct_mmi.c +529 -0
  227. package/openh264/codec/encoder/core/mips/quant_mmi.c +553 -0
  228. package/openh264/codec/encoder/core/mips/score_mmi.c +324 -0
  229. package/openh264/codec/encoder/core/x86/coeff.asm +695 -0
  230. package/openh264/codec/encoder/core/x86/dct.asm +108 -0
  231. package/openh264/codec/encoder/core/x86/intra_pred.asm +1129 -0
  232. package/openh264/codec/encoder/core/x86/matrix_transpose.asm +395 -0
  233. package/openh264/codec/encoder/core/x86/memzero.asm +132 -0
  234. package/openh264/codec/encoder/core/x86/quant.asm +507 -0
  235. package/openh264/codec/encoder/core/x86/sample_sc.asm +1829 -0
  236. package/openh264/codec/encoder/core/x86/score.asm +363 -0
  237. package/openh264/codec/encoder/meson.build +98 -0
  238. package/openh264/codec/encoder/plus/inc/welsEncoderExt.h +129 -0
  239. package/openh264/codec/encoder/targets.mk +146 -0
  240. package/openh264/codec/meson.build +9 -0
  241. package/openh264/codec/processing/interface/IWelsVP.h +310 -0
  242. package/openh264/codec/processing/meson.build +71 -0
  243. package/openh264/codec/processing/targets.mk +117 -0
  244. package/openh264/docs/doxygen/Doxyfile +1865 -0
  245. package/openh264/docs/doxygen/Home.rest +25 -0
  246. package/openh264/docs/doxygen/ISVCDecoder.rest +19 -0
  247. package/openh264/docs/doxygen/ISVCEncoder.rest +16 -0
  248. package/openh264/docs/doxygen/UsageExampleForDecoder.rest +10 -0
  249. package/openh264/docs/doxygen/UsageExampleForEncoder.rest +15 -0
  250. package/openh264/docs/doxygen2rst.py +619 -0
  251. package/openh264/gmpopenh264.info +4 -0
  252. package/openh264/gradlew +172 -0
  253. package/openh264/meson.build +218 -0
  254. package/openh264/meson_options.txt +1 -0
  255. package/openh264/module/RefCounted.h +141 -0
  256. package/openh264/module/gmp-openh264.cpp +1196 -0
  257. package/openh264/module/targets.mk +14 -0
  258. package/openh264/module/task_utils.h +38 -0
  259. package/openh264/module/task_utils.py +207 -0
  260. package/openh264/module/task_utils_generated.h +2087 -0
  261. package/openh264/openh264.def +7 -0
  262. package/openh264/openh264.pc.in +10 -0
  263. package/openh264/openh264.rc +61 -0
  264. package/openh264/openh264.rc.template +61 -0
  265. package/openh264/res/Adobe_PDF_sample_a_1024x768_50Frms.264 +0 -0
  266. package/openh264/res/BA1_FT_C.264 +0 -0
  267. package/openh264/res/BA1_Sony_D.jsv +0 -0
  268. package/openh264/res/BAMQ1_JVC_C.264 +0 -0
  269. package/openh264/res/BAMQ2_JVC_C.264 +0 -0
  270. package/openh264/res/BANM_MW_D.264 +0 -0
  271. package/openh264/res/BASQP1_Sony_C.jsv +0 -0
  272. package/openh264/res/BA_MW_D.264 +0 -0
  273. package/openh264/res/BA_MW_D_IDR_LOST.264 +0 -0
  274. package/openh264/res/BA_MW_D_P_LOST.264 +0 -0
  275. package/openh264/res/CI1_FT_B.264 +0 -0
  276. package/openh264/res/CI_MW_D.264 +0 -0
  277. package/openh264/res/CVFC1_Sony_C.jsv +0 -0
  278. package/openh264/res/CVPCMNL1_SVA_C.264 +0 -0
  279. package/openh264/res/CiscoVT2people_160x96_6fps.yuv +1 -0
  280. package/openh264/res/CiscoVT2people_320x192_12fps.yuv +1 -0
  281. package/openh264/res/Cisco_Absolute_Power_1280x720_30fps.yuv +3607 -0
  282. package/openh264/res/Cisco_Adobe_PDF_sample_a_1024x768_CAVLC_Bframe_9.264 +0 -0
  283. package/openh264/res/Cisco_Men_whisper_640x320_CABAC_Bframe_9.264 +0 -0
  284. package/openh264/res/Cisco_Men_whisper_640x320_CAVLC_Bframe_9.264 +0 -0
  285. package/openh264/res/Error_I_P.264 +0 -0
  286. package/openh264/res/LS_SVA_D.264 +0 -0
  287. package/openh264/res/MIDR_MW_D.264 +0 -0
  288. package/openh264/res/MPS_MW_A.264 +0 -0
  289. package/openh264/res/MR1_BT_A.h264 +0 -0
  290. package/openh264/res/MR1_MW_A.264 +0 -0
  291. package/openh264/res/MR2_MW_A.264 +0 -0
  292. package/openh264/res/MR2_TANDBERG_E.264 +0 -0
  293. package/openh264/res/NL1_Sony_D.jsv +0 -0
  294. package/openh264/res/NLMQ1_JVC_C.264 +0 -0
  295. package/openh264/res/NLMQ2_JVC_C.264 +0 -0
  296. package/openh264/res/NRF_MW_E.264 +0 -0
  297. package/openh264/res/QCIF_2P_I_allIPCM.264 +0 -0
  298. package/openh264/res/SVA_BA1_B.264 +0 -0
  299. package/openh264/res/SVA_BA2_D.264 +0 -0
  300. package/openh264/res/SVA_Base_B.264 +0 -0
  301. package/openh264/res/SVA_CL1_E.264 +0 -0
  302. package/openh264/res/SVA_FM1_E.264 +0 -0
  303. package/openh264/res/SVA_NL1_B.264 +0 -0
  304. package/openh264/res/SVA_NL2_E.264 +0 -0
  305. package/openh264/res/SarVui.264 +0 -0
  306. package/openh264/res/Static.264 +0 -0
  307. package/openh264/res/Static_152_100.yuv +1 -0
  308. package/openh264/res/VID_1280x544_cabac_temporal_direct.264 +0 -0
  309. package/openh264/res/VID_1280x544_cavlc_temporal_direct.264 +0 -0
  310. package/openh264/res/VID_1280x720_cabac_temporal_direct.264 +0 -0
  311. package/openh264/res/VID_1280x720_cavlc_temporal_direct.264 +0 -0
  312. package/openh264/res/VID_1920x1080_cabac_temporal_direct.264 +0 -0
  313. package/openh264/res/VID_1920x1080_cavlc_temporal_direct.264 +0 -0
  314. package/openh264/res/Zhling_1280x720.264 +0 -0
  315. package/openh264/res/jm_1080p_allslice.264 +0 -0
  316. package/openh264/res/sps_subsetsps_bothVUI.264 +0 -0
  317. package/openh264/res/test_cif_I_CABAC_PCM.264 +0 -0
  318. package/openh264/res/test_cif_I_CABAC_slice.264 +0 -0
  319. package/openh264/res/test_cif_P_CABAC_slice.264 +0 -0
  320. package/openh264/res/test_qcif_cabac.264 +0 -0
  321. package/openh264/res/test_scalinglist_jm.264 +0 -0
  322. package/openh264/res/test_vd_1d.264 +0 -0
  323. package/openh264/res/test_vd_rc.264 +0 -0
  324. package/openh264/run_Test.sh +91 -0
  325. package/openh264/settings.gradle +4 -0
  326. package/openh264/subprojects/gtest.wrap +10 -0
  327. package/openh264/testbin/AutoBuild_Windows_VS2008.bat +59 -0
  328. package/openh264/testbin/CmdLineExample.sh +7 -0
  329. package/openh264/testbin/CmdLineReadMe +22 -0
  330. package/openh264/testbin/layer2.cfg +37 -0
  331. package/openh264/testbin/layer2_arbitrary_res.cfg +37 -0
  332. package/openh264/testbin/layer2_vd.cfg +38 -0
  333. package/openh264/testbin/layer2_vd_rc.cfg +38 -0
  334. package/openh264/testbin/welsenc.cfg +71 -0
  335. package/openh264/testbin/welsenc_arbitrary_res.cfg +70 -0
  336. package/openh264/testbin/welsenc_ios.cfg +71 -0
  337. package/openh264/testbin/welsenc_vd_1d.cfg +70 -0
  338. package/openh264/testbin/welsenc_vd_rc.cfg +69 -0
  339. package/openh264/ut.def +2 -0
  340. package/package.json +1 -1
package/openh264/codec/common/mips/satd_sad_mmi.c ADDED
@@ -0,0 +1,2154 @@
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 satd_sad_mmi.c
33
+ *
34
+ * \brief Loongson optimization
35
+ *
36
+ * \date 23/07/2018 Created
37
+ *
38
+ *************************************************************************************
39
+ */
40
+ #include <stdint.h>
41
+ #include "asmdefs_mmi.h"
42
+
43
+ #define MMI_SumWHorizon1(f0, f2, f4, f6, f8, f10, r0) \
44
+ "dli "#r0", 0x10 \n\t" \
45
+ "dmtc1 "#r0", "#f8" \n\t" \
46
+ "dli "#r0", 0x20 \n\t" \
47
+ "dmtc1 "#r0", "#f10" \n\t" \
48
+ "mov.d "#f4", "#f2" \n\t" \
49
+ "xor "#f6", "#f6", "#f6" \n\t" \
50
+ "paddush "#f0", "#f0", "#f4" \n\t" \
51
+ "paddush "#f2", "#f2", "#f6" \n\t" \
52
+ "dsrl "#f6", "#f2", "#f10" \n\t" \
53
+ "punpcklwd "#f4", "#f2", "#f2" \n\t" \
54
+ "punpckhwd "#f4", "#f0", "#f4" \n\t" \
55
+ "paddush "#f0", "#f0", "#f4" \n\t" \
56
+ "paddush "#f2", "#f2", "#f6" \n\t" \
57
+ "dsrl "#f4", "#f0", "#f8" \n\t" \
58
+ "pinsrh_3 "#f4", "#f4", "#f2" \n\t" \
59
+ "dsrl "#f6", "#f2", "#f8" \n\t" \
60
+ "paddush "#f0", "#f0", "#f4" \n\t" \
61
+ "paddush "#f2", "#f2", "#f6" \n\t"
62
+
63
+ #define MMI_GetSad8x4 \
64
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t" \
65
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t" \
66
+ "gsldlc1 $f4, 0x7($8) \n\t" \
67
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t" \
68
+ "gsldrc1 $f4, 0x0($8) \n\t" \
69
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t" \
70
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t" \
71
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t" \
72
+ "gsldlc1 $f6, 0x7($8) \n\t" \
73
+ "gsldlc1 $f8, 0x7(%[pSample2]) \n\t" \
74
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t" \
75
+ "gsldrc1 $f6, 0x0($8) \n\t" \
76
+ "gsldrc1 $f8, 0x0(%[pSample2]) \n\t" \
77
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t" \
78
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t" \
79
+ "gsldlc1 $f12, 0x7($9) \n\t" \
80
+ "gsldlc1 $f10, 0x7(%[pSample2]) \n\t" \
81
+ "gsldrc1 $f12, 0x0($9) \n\t" \
82
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t" \
83
+ "gsldrc1 $f10, 0x0(%[pSample2]) \n\t" \
84
+ "gsldlc1 $f14, 0x7($9) \n\t" \
85
+ "gsldrc1 $f14, 0x0($9) \n\t" \
86
+ "pasubub $f0, $f0, $f8 \n\t" \
87
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t" \
88
+ "pasubub $f2, $f2, $f10 \n\t" \
89
+ "biadd $f0, $f0 \n\t" \
90
+ "biadd $f2, $f2 \n\t" \
91
+ "pasubub $f4, $f4, $f12 \n\t" \
92
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t" \
93
+ "pasubub $f6, $f6, $f14 \n\t" \
94
+ "biadd $f4, $f4 \n\t" \
95
+ "biadd $f6, $f6 \n\t" \
96
+ "paddh $f24, $f24, $f0 \n\t" \
97
+ "paddh $f26, $f26, $f2 \n\t" \
98
+ "paddh $f24, $f24, $f4 \n\t" \
99
+ "paddh $f26, $f26, $f6 \n\t"
100
+
101
+ #define MMI_GetSad8x4_End \
102
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t" \
103
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t" \
104
+ "gsldlc1 $f4, 0x7($8) \n\t" \
105
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t" \
106
+ "gsldrc1 $f4, 0x0($8) \n\t" \
107
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t" \
108
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t" \
109
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t" \
110
+ "gsldlc1 $f6, 0x7($8) \n\t" \
111
+ "gsldlc1 $f8, 0x7(%[pSample2]) \n\t" \
112
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t" \
113
+ "gsldrc1 $f6, 0x0($8) \n\t" \
114
+ "gsldrc1 $f8, 0x0(%[pSample2]) \n\t" \
115
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t" \
116
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t" \
117
+ "gsldlc1 $f12, 0x7($9) \n\t" \
118
+ "gsldlc1 $f10, 0x7(%[pSample2]) \n\t" \
119
+ "gsldrc1 $f12, 0x0($9) \n\t" \
120
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t" \
121
+ "gsldrc1 $f10, 0x0(%[pSample2]) \n\t" \
122
+ "gsldlc1 $f14, 0x7($9) \n\t" \
123
+ "gsldrc1 $f14, 0x0($9) \n\t" \
124
+ "pasubub $f0, $f0, $f8 \n\t" \
125
+ "pasubub $f2, $f2, $f10 \n\t" \
126
+ "biadd $f0, $f0 \n\t" \
127
+ "biadd $f2, $f2 \n\t" \
128
+ "pasubub $f4, $f4, $f12 \n\t" \
129
+ "pasubub $f6, $f6, $f14 \n\t" \
130
+ "biadd $f4, $f4 \n\t" \
131
+ "biadd $f6, $f6 \n\t" \
132
+ "paddh $f24, $f24, $f0 \n\t" \
133
+ "paddh $f26, $f26, $f2 \n\t" \
134
+ "paddh $f24, $f24, $f4 \n\t" \
135
+ "paddh $f26, $f26, $f6 \n\t"
136
+
137
+ #define CACHE_SPLIT_CHECK(r0, width, cacheline) \
138
+ "and "#r0", "#r0", 0x1f \n\t" \
139
+ PTR_ADDIU ""#r0", "#r0", -0x1f \n\t"
140
+
141
+ #define MMI_GetSad2x16 \
142
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
143
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
144
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
145
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
146
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
147
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
148
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
149
+ "pasubub $f4, $f4, $f8 \n\t" \
150
+ "pasubub $f6, $f6, $f10 \n\t" \
151
+ "biadd $f4, $f4 \n\t" \
152
+ "biadd $f6, $f6 \n\t" \
153
+ "paddh $f0, $f0, $f4 \n\t" \
154
+ "paddh $f2, $f2, $f6 \n\t" \
155
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
156
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
157
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
158
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
159
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
160
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
161
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
162
+ "pasubub $f4, $f4, $f8 \n\t" \
163
+ "pasubub $f6, $f6, $f10 \n\t" \
164
+ "biadd $f4, $f4 \n\t" \
165
+ "biadd $f6, $f6 \n\t" \
166
+ "paddh $f0, $f0, $f4 \n\t" \
167
+ "paddh $f2, $f2, $f6 \n\t"
168
+
169
+ #define MMI_GetSad4x16 \
170
+ "gsldlc1 $f0, 0x7(%[pSample2]) \n\t" \
171
+ "gsldlc1 $f2, 0xF(%[pSample2]) \n\t" \
172
+ "gsldrc1 $f0, 0x0(%[pSample2]) \n\t" \
173
+ "gsldrc1 $f2, 0x8(%[pSample2]) \n\t" \
174
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
175
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
176
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
177
+ "pasubub $f0, $f0, $f8 \n\t" \
178
+ "pasubub $f2, $f2, $f10 \n\t" \
179
+ "biadd $f0, $f0 \n\t" \
180
+ "biadd $f2, $f2 \n\t" \
181
+ "paddh $f28, $f28, $f0 \n\t" \
182
+ "paddh $f30, $f30, $f2 \n\t" \
183
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
184
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
185
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
186
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
187
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
188
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
189
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
190
+ "pasubub $f4, $f4, $f8 \n\t" \
191
+ "pasubub $f6, $f6, $f10 \n\t" \
192
+ "biadd $f4, $f4 \n\t" \
193
+ "biadd $f6, $f6 \n\t" \
194
+ "paddh $f28, $f28, $f4 \n\t" \
195
+ "paddh $f30, $f30, $f6 \n\t" \
196
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
197
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
198
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
199
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
200
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
201
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
202
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
203
+ "pasubub $f4, $f4, $f8 \n\t" \
204
+ "pasubub $f6, $f6, $f10 \n\t" \
205
+ "biadd $f4, $f4 \n\t" \
206
+ "biadd $f6, $f6 \n\t" \
207
+ "paddh $f28, $f28, $f4 \n\t" \
208
+ "paddh $f30, $f30, $f6 \n\t" \
209
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
210
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
211
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
212
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
213
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
214
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
215
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
216
+ "pasubub $f4, $f4, $f8 \n\t" \
217
+ "pasubub $f6, $f6, $f10 \n\t" \
218
+ "biadd $f4, $f4 \n\t" \
219
+ "biadd $f6, $f6 \n\t" \
220
+ "paddh $f28, $f28, $f4 \n\t" \
221
+ "paddh $f30, $f30, $f6 \n\t"
222
+
223
+ #define MMI_GetSad4x16_Aligned \
224
+ "gslqc1 $f2, $f0, 0x0(%[pSample2]) \n\t" \
225
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
226
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
227
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
228
+ "pasubub $f0, $f0, $f8 \n\t" \
229
+ "pasubub $f2, $f2, $f10 \n\t" \
230
+ "biadd $f0, $f0 \n\t" \
231
+ "biadd $f2, $f2 \n\t" \
232
+ "paddh $f28, $f28, $f0 \n\t" \
233
+ "paddh $f30, $f30, $f2 \n\t" \
234
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
235
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
236
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
237
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
238
+ "pasubub $f4, $f4, $f8 \n\t" \
239
+ "pasubub $f6, $f6, $f10 \n\t" \
240
+ "biadd $f4, $f4 \n\t" \
241
+ "biadd $f6, $f6 \n\t" \
242
+ "paddh $f28, $f28, $f4 \n\t" \
243
+ "paddh $f30, $f30, $f6 \n\t" \
244
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
245
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
246
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
247
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
248
+ "pasubub $f4, $f4, $f8 \n\t" \
249
+ "pasubub $f6, $f6, $f10 \n\t" \
250
+ "biadd $f4, $f4 \n\t" \
251
+ "biadd $f6, $f6 \n\t" \
252
+ "paddh $f28, $f28, $f4 \n\t" \
253
+ "paddh $f30, $f30, $f6 \n\t" \
254
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
255
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
256
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
257
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
258
+ "pasubub $f4, $f4, $f8 \n\t" \
259
+ "pasubub $f6, $f6, $f10 \n\t" \
260
+ "biadd $f4, $f4 \n\t" \
261
+ "biadd $f6, $f6 \n\t" \
262
+ "paddh $f28, $f28, $f4 \n\t" \
263
+ "paddh $f30, $f30, $f6 \n\t"
264
+
265
+ #define MMI_GetSad4x16_End \
266
+ "gsldlc1 $f0, 0x7(%[pSample2]) \n\t" \
267
+ "gsldlc1 $f2, 0xF(%[pSample2]) \n\t" \
268
+ "gsldrc1 $f0, 0x0(%[pSample2]) \n\t" \
269
+ "gsldrc1 $f2, 0x8(%[pSample2]) \n\t" \
270
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
271
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
272
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
273
+ "pasubub $f0, $f0, $f8 \n\t" \
274
+ "pasubub $f2, $f2, $f10 \n\t" \
275
+ "biadd $f0, $f0 \n\t" \
276
+ "biadd $f2, $f2 \n\t" \
277
+ "paddh $f28, $f28, $f0 \n\t" \
278
+ "paddh $f30, $f30, $f2 \n\t" \
279
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
280
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
281
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
282
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
283
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
284
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
285
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
286
+ "pasubub $f4, $f4, $f8 \n\t" \
287
+ "pasubub $f6, $f6, $f10 \n\t" \
288
+ "biadd $f4, $f4 \n\t" \
289
+ "biadd $f6, $f6 \n\t" \
290
+ "paddh $f28, $f28, $f4 \n\t" \
291
+ "paddh $f30, $f30, $f6 \n\t" \
292
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
293
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
294
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
295
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
296
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
297
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
298
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
299
+ "pasubub $f4, $f4, $f8 \n\t" \
300
+ "pasubub $f6, $f6, $f10 \n\t" \
301
+ "biadd $f4, $f4 \n\t" \
302
+ "biadd $f6, $f6 \n\t" \
303
+ "paddh $f28, $f28, $f4 \n\t" \
304
+ "paddh $f30, $f30, $f6 \n\t" \
305
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t" \
306
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t" \
307
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t" \
308
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t" \
309
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
310
+ "pasubub $f4, $f4, $f8 \n\t" \
311
+ "pasubub $f6, $f6, $f10 \n\t" \
312
+ "biadd $f4, $f4 \n\t" \
313
+ "biadd $f6, $f6 \n\t" \
314
+ "paddh $f28, $f28, $f4 \n\t" \
315
+ "paddh $f30, $f30, $f6 \n\t"
316
+
317
+ #define MMI_GetSad4x16_Aligned_End \
318
+ "gslqc1 $f2, $f0, 0x0(%[pSample2]) \n\t" \
319
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
320
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
321
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
322
+ "pasubub $f0, $f0, $f8 \n\t" \
323
+ "pasubub $f2, $f2, $f10 \n\t" \
324
+ "biadd $f0, $f0 \n\t" \
325
+ "biadd $f2, $f2 \n\t" \
326
+ "paddh $f28, $f28, $f0 \n\t" \
327
+ "paddh $f30, $f30, $f2 \n\t" \
328
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
329
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
330
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
331
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
332
+ "pasubub $f4, $f4, $f8 \n\t" \
333
+ "pasubub $f6, $f6, $f10 \n\t" \
334
+ "biadd $f4, $f4 \n\t" \
335
+ "biadd $f6, $f6 \n\t" \
336
+ "paddh $f28, $f28, $f4 \n\t" \
337
+ "paddh $f30, $f30, $f6 \n\t" \
338
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
339
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
340
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t" \
341
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t" \
342
+ "pasubub $f4, $f4, $f8 \n\t" \
343
+ "pasubub $f6, $f6, $f10 \n\t" \
344
+ "biadd $f4, $f4 \n\t" \
345
+ "biadd $f6, $f6 \n\t" \
346
+ "paddh $f28, $f28, $f4 \n\t" \
347
+ "paddh $f30, $f30, $f6 \n\t" \
348
+ "gslqc1 $f6, $f4, 0x0(%[pSample2]) \n\t" \
349
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t" \
350
+ "pasubub $f4, $f4, $f8 \n\t" \
351
+ "pasubub $f6, $f6, $f10 \n\t" \
352
+ "biadd $f4, $f4 \n\t" \
353
+ "biadd $f6, $f6 \n\t" \
354
+ "paddh $f28, $f28, $f4 \n\t" \
355
+ "paddh $f30, $f30, $f6 \n\t"
356
+
357
+ #define MMI_Get4LW16Sad(f0, f2, f4, f6, f8, f10, f12, f14, r0) \
358
+ "pasubub "#f0", "#f0", "#f12" \n\t" \
359
+ "pasubub "#f2", "#f2", "#f14" \n\t" \
360
+ "pasubub "#f12", "#f12", "#f8" \n\t" \
361
+ "pasubub "#f14", "#f14", "#f10" \n\t" \
362
+ "biadd "#f0", "#f0" \n\t" \
363
+ "biadd "#f2", "#f2" \n\t" \
364
+ "biadd "#f12", "#f12" \n\t" \
365
+ "biadd "#f14", "#f14" \n\t" \
366
+ "paddh $f20, $f20, "#f0" \n\t" \
367
+ "paddh $f22, $f22, "#f2" \n\t" \
368
+ "paddh $f16, $f16, "#f12" \n\t" \
369
+ "paddh $f18, $f18, "#f14" \n\t" \
370
+ "gsldlc1 "#f12", 0x6("#r0") \n\t" \
371
+ "gsldlc1 "#f14", 0xE("#r0") \n\t" \
372
+ "gsldrc1 "#f12", -0x1("#r0") \n\t" \
373
+ "gsldrc1 "#f14", 0x7("#r0") \n\t" \
374
+ "pasubub "#f12", "#f12", "#f4" \n\t" \
375
+ "pasubub "#f14", "#f14", "#f6" \n\t" \
376
+ "biadd "#f12", "#f12" \n\t" \
377
+ "biadd "#f14", "#f14" \n\t" \
378
+ "paddh $f24, $f24, "#f12" \n\t" \
379
+ "paddh $f26, $f26, "#f14" \n\t" \
380
+ "gsldlc1 "#f12", 0x8("#r0") \n\t" \
381
+ "gsldlc1 "#f14", 0x10("#r0") \n\t" \
382
+ "gsldrc1 "#f12", 0x1("#r0") \n\t" \
383
+ "gsldrc1 "#f14", 0x9("#r0") \n\t" \
384
+ "pasubub "#f12", "#f12", "#f4" \n\t" \
385
+ "pasubub "#f14", "#f14", "#f6" \n\t" \
386
+ "biadd "#f12", "#f12" \n\t" \
387
+ "biadd "#f14", "#f14" \n\t" \
388
+ "paddh $f28, $f28, "#f12" \n\t" \
389
+ "paddh $f30, $f30, "#f14" \n\t"
390
+
391
+ #define MMI_HDMTwo4x4(f0, f2, f4, f6, f8, f10, f12, f14, f16, f18) \
392
+ MMI_SumSub(f0, f2, f4, f6, f16, f18) \
393
+ MMI_SumSub(f8, f10, f12, f14, f16, f18) \
394
+ MMI_SumSub(f4, f6, f12, f14, f16, f18) \
395
+ MMI_SumSub(f0, f2, f8, f10, f16, f18)
396
+
397
+ #define MMI_SumAbs4(f0, f2, f4, f6, f8, f10, f12, f14, f16, f18, f20, f22, f24, f26) \
398
+ WELS_AbsH(f0, f2, f0, f2, f8, f10) \
399
+ WELS_AbsH(f4, f6, f4, f6, f8, f10) \
400
+ WELS_AbsH(f12, f14, f12, f14, f20, f22) \
401
+ WELS_AbsH(f16, f18, f16, f18, f20, f22) \
402
+ "paddush "#f0", "#f0", "#f4" \n\t" \
403
+ "paddush "#f2", "#f2", "#f6" \n\t" \
404
+ "paddush "#f12", "#f12", "#f16" \n\t" \
405
+ "paddush "#f14", "#f14", "#f18" \n\t" \
406
+ "paddush "#f24", "#f24", "#f0" \n\t" \
407
+ "paddush "#f26", "#f26", "#f2" \n\t" \
408
+ "paddush "#f24", "#f24", "#f12" \n\t" \
409
+ "paddush "#f26", "#f26", "#f14" \n\t"
410
+
411
+ #define MMI_SumWHorizon(f0, f2, f4, f6, f8, f10) \
412
+ "paddh "#f0", "#f0", "#f2" \n\t" \
413
+ "punpckhhw "#f2", "#f0", "#f8" \n\t" \
414
+ "punpcklhw "#f0", "#f0", "#f8" \n\t" \
415
+ "paddw "#f0", "#f0", "#f2" \n\t" \
416
+ "pshufh "#f2", "#f0", "#f10" \n\t" \
417
+ "paddw "#f0", "#f0", "#f2" \n\t"
418
+
419
+ #define MMI_LoadDiff8P_Offset_Stride0(f0, f2, f4, f6, f8, r0, r1) \
420
+ "gsldlc1 "#f0", 0x7("#r0") \n\t" \
421
+ "gsldlc1 "#f4", 0x7("#r1") \n\t" \
422
+ PTR_ADDU "$11, %[pSample1], %[iStride1] \n\t" \
423
+ "gsldrc1 "#f0", 0x0("#r0") \n\t" \
424
+ "gsldrc1 "#f4", 0x0("#r1") \n\t" \
425
+ PTR_ADDU "$12, %[pSample2], %[iStride2] \n\t" \
426
+ "punpckhbh "#f2", "#f0", "#f8" \n\t" \
427
+ "punpcklbh "#f0", "#f0", "#f8" \n\t" \
428
+ "punpckhbh "#f6", "#f4", "#f8" \n\t" \
429
+ "punpcklbh "#f4", "#f4", "#f8" \n\t" \
430
+ "psubh "#f0", "#f0", "#f4" \n\t" \
431
+ "psubh "#f2", "#f2", "#f6" \n\t"
432
+
433
+ #define MMI_LoadDiff8P_Offset_Stride1(f0, f2, f4, f6, f8, r0, r1) \
434
+ "gsldlc1 "#f0", 0x7("#r0") \n\t" \
435
+ "gsldlc1 "#f4", 0x7("#r1") \n\t" \
436
+ PTR_ADDU "%[pSample1], $11, %[iStride1] \n\t" \
437
+ "gsldrc1 "#f0", 0x0("#r0") \n\t" \
438
+ "gsldrc1 "#f4", 0x0("#r1") \n\t" \
439
+ PTR_ADDU "%[pSample2], $12, %[iStride2] \n\t" \
440
+ "punpckhbh "#f2", "#f0", "#f8" \n\t" \
441
+ "punpcklbh "#f0", "#f0", "#f8" \n\t" \
442
+ "punpckhbh "#f6", "#f4", "#f8" \n\t" \
443
+ "punpcklbh "#f4", "#f4", "#f8" \n\t" \
444
+ "psubh "#f0", "#f0", "#f4" \n\t" \
445
+ "psubh "#f2", "#f2", "#f6" \n\t"
446
+
447
+ #define MMI_LoadDiff8P_Offset8(f0, f2, f4, f6, f8, r0, r1) \
448
+ "gsldlc1 "#f0", 0x7("#r0") \n\t" \
449
+ "gsldlc1 "#f4", 0x7("#r1") \n\t" \
450
+ PTR_ADDU "%[pSample1], $9, 0x8 \n\t" \
451
+ "gsldrc1 "#f0", 0x0("#r0") \n\t" \
452
+ "gsldrc1 "#f4", 0x0("#r1") \n\t" \
453
+ PTR_ADDU "%[pSample2], $10, 0x8 \n\t" \
454
+ "punpckhbh "#f2", "#f0", "#f8" \n\t" \
455
+ "punpcklbh "#f0", "#f0", "#f8" \n\t" \
456
+ "punpckhbh "#f6", "#f4", "#f8" \n\t" \
457
+ "punpcklbh "#f4", "#f4", "#f8" \n\t" \
458
+ "psubh "#f0", "#f0", "#f4" \n\t" \
459
+ "psubh "#f2", "#f2", "#f6" \n\t"
460
+
461
+ #define MMI_GetSatd8x8 \
462
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
463
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
464
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
465
+ MMI_LoadDiff8P_Offset_Stride1($f12, $f14, $f20, $f22, $f28, $11, $12) \
466
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
467
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
468
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
469
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26) \
470
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
471
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
472
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
473
+ MMI_LoadDiff8P_Offset_Stride1($f12, $f14, $f20, $f22, $f28, $11, $12) \
474
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
475
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
476
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
477
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26)
478
+
479
+ #define MMI_GetSatd8x8_Offset8 \
480
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
481
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
482
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
483
+ MMI_LoadDiff8P_Offset_Stride1($f12, $f14, $f20, $f22, $f28, $11, $12) \
484
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
485
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
486
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
487
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26) \
488
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
489
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
490
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
491
+ MMI_LoadDiff8P_Offset8($f12, $f14, $f20, $f22, $f28, $11, $12) \
492
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
493
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
494
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
495
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26)
496
+
497
+ #define MMI_GetSatd8x8_End \
498
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
499
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
500
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
501
+ MMI_LoadDiff8P_Offset_Stride1($f12, $f14, $f20, $f22, $f28, $11, $12) \
502
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
503
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
504
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
505
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26) \
506
+ MMI_LoadDiff8P_Offset_Stride0($f0, $f2, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
507
+ MMI_LoadDiff8P_Offset_Stride1($f4, $f6, $f20, $f22, $f28, $11, $12) \
508
+ MMI_LoadDiff8P_Offset_Stride0($f8, $f10, $f16, $f18, $f28, %[pSample1], %[pSample2]) \
509
+ MMI_LoadDiff8P($f12, $f14, $f20, $f22, $f28, $11, $12) \
510
+ MMI_HDMTwo4x4($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $f16, $f18) \
511
+ MMI_TransTwo4x4H($f12, $f14, $f4, $f6, $f0, $f2, $f8, $f10, $f16, $f18) \
512
+ MMI_HDMTwo4x4($f12, $f14, $f4, $f6, $f8, $f10, $f16, $f18, $f20, $f22) \
513
+ MMI_SumAbs4($f16, $f18, $f4, $f6, $f0, $f2, $f8, $f10, $f12, $f14, $f20, $f22, $f24, $f26)
514
+
515
+ int32_t WelsSampleSad16x16_mmi (uint8_t* pSample1, int32_t iStride1,
516
+ uint8_t* pSample2, int32_t iStride2) {
517
+ int32_t iSadSum = 0;
518
+ BACKUP_REG;
519
+ __asm__ volatile (
520
+ ".set arch=loongson3a \n\t"
521
+ "and $8, %[pSample2], 0xF \n\t"
522
+ "xor $f28, $f28, $f28 \n\t"
523
+ "xor $f30, $f30, $f30 \n\t"
524
+ "bnez $8, unaligned \n\t"
525
+ "aligned: \n\t"
526
+ MMI_GetSad4x16_Aligned
527
+ MMI_GetSad4x16_Aligned
528
+ MMI_GetSad4x16_Aligned
529
+ MMI_GetSad4x16_Aligned_End
530
+ "b out \n\t"
531
+
532
+ "unaligned: \n\t"
533
+ MMI_GetSad4x16
534
+ MMI_GetSad4x16
535
+ MMI_GetSad4x16
536
+ MMI_GetSad4x16_End
537
+ "out: \n\t"
538
+ "mov.d $f0, $f30 \n\t"
539
+ "paddh $f0, $f0, $f28 \n\t"
540
+ "dmfc1 %[iSadSum], $f0 \n\t"
541
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSadSum]"=r"((int)iSadSum),
542
+ [pSample2]"+&r"((unsigned char *)pSample2)
543
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
544
+ : "memory", "$8", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
545
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
546
+ );
547
+ RECOVER_REG;
548
+ return iSadSum;
549
+ }
550
+
551
+ int32_t WelsSampleSad16x8_mmi (uint8_t* pSample1, int32_t iStride1,
552
+ uint8_t* pSample2, int32_t iStride2) {
553
+ int32_t iSadSum = 0;
554
+ BACKUP_REG;
555
+ __asm__ volatile (
556
+ ".set arch=loongson3a \n\t"
557
+ "gsldlc1 $f0, 0x7(%[pSample2]) \n\t"
558
+ "gsldlc1 $f2, 0xF(%[pSample2]) \n\t"
559
+ "gsldrc1 $f0, 0x0(%[pSample2]) \n\t"
560
+ "gsldrc1 $f2, 0x8(%[pSample2]) \n\t"
561
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t"
562
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
563
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
564
+ "pasubub $f0, $f0, $f8 \n\t"
565
+ "pasubub $f2, $f2, $f10 \n\t"
566
+ "biadd $f0, $f0 \n\t"
567
+ "biadd $f2, $f2 \n\t"
568
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t"
569
+ "gsldlc1 $f6, 0xF(%[pSample2]) \n\t"
570
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t"
571
+ "gsldrc1 $f6, 0x8(%[pSample2]) \n\t"
572
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t"
573
+ "pasubub $f4, $f4, $f8 \n\t"
574
+ "pasubub $f6, $f6, $f10 \n\t"
575
+ "biadd $f4, $f4 \n\t"
576
+ "biadd $f6, $f6 \n\t"
577
+ "paddh $f0, $f0, $f4 \n\t"
578
+ "paddh $f2, $f2, $f6 \n\t"
579
+
580
+ MMI_GetSad2x16
581
+ MMI_GetSad2x16
582
+ MMI_GetSad2x16
583
+
584
+ "paddh $f0, $f0, $f2 \n\t"
585
+ "dmfc1 %[iSadSum], $f0 \n\t"
586
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSadSum]"=r"((int)iSadSum),
587
+ [pSample2]"+&r"((unsigned char *)pSample2)
588
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
589
+ : "memory", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
590
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26"
591
+ );
592
+ RECOVER_REG;
593
+ return iSadSum;
594
+ }
595
+
596
+ int32_t WelsSampleSad8x16_mmi (uint8_t* pSample1, int32_t iStride1,
597
+ uint8_t* pSample2, int32_t iStride2) {
598
+ int32_t iSadSum = 0;
599
+ BACKUP_REG;
600
+ __asm__ volatile (
601
+ ".set arch=loongson3a \n\t"
602
+ "xor $f24, $f24, $f24 \n\t"
603
+ "xor $f26, $f26, $f26 \n\t"
604
+ MMI_GetSad8x4
605
+ MMI_GetSad8x4
606
+ MMI_GetSad8x4
607
+ MMI_GetSad8x4_End
608
+ "paddh $f0, $f26, $f24 \n\t"
609
+ "dmfc1 %[iSadSum], $f0 \n\t"
610
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSadSum]"=r"((int)iSadSum),
611
+ [pSample2]"+&r"((unsigned char *)pSample2)
612
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
613
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
614
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26"
615
+ );
616
+ RECOVER_REG;
617
+ return iSadSum;
618
+ }
619
+
620
+ int32_t WelsSampleSad4x4_mmi (uint8_t* pSample1, int32_t iStride1,
621
+ uint8_t* pSample2, int32_t iStride2) {
622
+ int32_t iSadSum = 0;
623
+ __asm__ volatile (
624
+ ".set arch=loongson3a \n\t"
625
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
626
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
627
+ "gsldlc1 $f2, 0x7($8) \n\t"
628
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
629
+ "gsldrc1 $f2, 0x0($8) \n\t"
630
+ "punpcklwd $f0, $f0, $f2 \n\t"
631
+
632
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
633
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
634
+ "gsldlc1 $f8, 0x7($9) \n\t"
635
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
636
+ "gsldrc1 $f8, 0x0($9) \n\t"
637
+ "punpcklwd $f6, $f6, $f8 \n\t"
638
+ "pasubub $f0, $f0, $f6 \n\t"
639
+ "biadd $f0, $f0 \n\t"
640
+
641
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
642
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
643
+
644
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
645
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t"
646
+ "gsldlc1 $f4, 0x7($8) \n\t"
647
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t"
648
+ "gsldrc1 $f4, 0x0($8) \n\t"
649
+ "punpcklwd $f2, $f2, $f4 \n\t"
650
+
651
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
652
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
653
+ "gsldlc1 $f8, 0x7($9) \n\t"
654
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
655
+ "gsldrc1 $f8, 0x0($9) \n\t"
656
+ "punpcklwd $f6, $f6, $f8 \n\t"
657
+ "pasubub $f2, $f2, $f6 \n\t"
658
+ "biadd $f2, $f2 \n\t"
659
+ "paddh $f0, $f0, $f2 \n\t"
660
+
661
+ "dmfc1 %[iSadSum], $f0 \n\t"
662
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSadSum]"=r"((int)iSadSum),
663
+ [pSample2]"+&r"((unsigned char *)pSample2)
664
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
665
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8"
666
+ );
667
+ return iSadSum;
668
+ }
669
+
670
+ int32_t WelsSampleSad8x8_mmi (uint8_t* pSample1, int32_t iStride1,
671
+ uint8_t* pSample2, int32_t iStride2) {
672
+ int32_t iSadSum = 0;
673
+ BACKUP_REG;
674
+ __asm__ volatile (
675
+ ".set arch=loongson3a \n\t"
676
+ CACHE_SPLIT_CHECK($8, 8, 32)
677
+ "blez $8, 1f \n\t"
678
+ "nop \n\t"
679
+ "xor $f28, $f28, $f28 \n\t"
680
+ "xor $f30, $f30, $f30 \n\t"
681
+
682
+ "move $9, %[pSample2] \n\t"
683
+ "and $9, $9, 0x7 \n\t"
684
+ PTR_SUBU "%[pSample2], %[pSample2], $9 \n\t"
685
+ "dli $8, 0x8 \n\t"
686
+ PTR_SUBU "$8, $8, $9 \n\t"
687
+
688
+ "dsll $9, $9, 0x3 \n\t"
689
+ "dsll $8, $8, 0x3 \n\t"
690
+ "dmtc1 $9, $f20 \n\t"
691
+ "dmtc1 $8, $f24 \n\t"
692
+ "dli $9, 0x8 \n\t"
693
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
694
+ PTR_ADDU "$9, $9, %[pSample2] \n\t"
695
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
696
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
697
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t"
698
+
699
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t"
700
+ "gsldlc1 $f8, 0x7($9) \n\t"
701
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t"
702
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t"
703
+ "gsldrc1 $f8, 0x0($9) \n\t"
704
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
705
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
706
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
707
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
708
+ "gsldlc1 $f10, 0x7($9) \n\t"
709
+ "dsrl $f4, $f4, $f20 \n\t"
710
+ "gsldrc1 $f10, 0x0($9) \n\t"
711
+ "dsrl $f6, $f6, $f20 \n\t"
712
+ "dsll $f8, $f8, $f24 \n\t"
713
+ "dsll $f10, $f10, $f24 \n\t"
714
+ "or $f4, $f4, $f8 \n\t"
715
+ "or $f6, $f6, $f10 \n\t"
716
+
717
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
718
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
719
+ "pasubub $f0, $f0, $f4 \n\t"
720
+ "pasubub $f2, $f2, $f6 \n\t"
721
+ "biadd $f0, $f0 \n\t"
722
+ "biadd $f2, $f2 \n\t"
723
+ "paddh $f28, $f28, $f0 \n\t"
724
+ "paddh $f30, $f30, $f2 \n\t"
725
+
726
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
727
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
728
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
729
+
730
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
731
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t"
732
+
733
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t"
734
+ "gsldlc1 $f8, 0x7($9) \n\t"
735
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t"
736
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t"
737
+ "gsldrc1 $f8, 0x0($9) \n\t"
738
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
739
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
740
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
741
+ "gsldlc1 $f10, 0x7($9) \n\t"
742
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
743
+ "gsldrc1 $f10, 0x0($9) \n\t"
744
+ "dsrl $f4, $f4, $f20 \n\t"
745
+ "dsrl $f6, $f6, $f20 \n\t"
746
+ "dsll $f8, $f8, $f24 \n\t"
747
+ "dsll $f10, $f10, $f24 \n\t"
748
+ "or $f4, $f4, $f8 \n\t"
749
+ "or $f6, $f6, $f10 \n\t"
750
+
751
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
752
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
753
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
754
+
755
+ "pasubub $f0, $f0, $f4 \n\t"
756
+ "pasubub $f2, $f2, $f6 \n\t"
757
+ "biadd $f0, $f0 \n\t"
758
+ "biadd $f2, $f2 \n\t"
759
+ "paddh $f28, $f28, $f0 \n\t"
760
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
761
+ "paddh $f30, $f30, $f2 \n\t"
762
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
763
+
764
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
765
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t"
766
+
767
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t"
768
+ "gsldlc1 $f8, 0x7($9) \n\t"
769
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t"
770
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t"
771
+ "gsldrc1 $f8, 0x0($9) \n\t"
772
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
773
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
774
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
775
+ "gsldlc1 $f10, 0x7($9) \n\t"
776
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
777
+ "gsldrc1 $f10, 0x0($9) \n\t"
778
+ "dsrl $f4, $f4, $f20 \n\t"
779
+ "dsrl $f6, $f6, $f20 \n\t"
780
+ "dsll $f8, $f8, $f24 \n\t"
781
+ "dsll $f10, $f10, $f24 \n\t"
782
+ "or $f4, $f4, $f8 \n\t"
783
+ "or $f6, $f6, $f10 \n\t"
784
+
785
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
786
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
787
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
788
+
789
+ "pasubub $f0, $f0, $f4 \n\t"
790
+ "pasubub $f2, $f2, $f6 \n\t"
791
+ "biadd $f0, $f0 \n\t"
792
+ "biadd $f2, $f2 \n\t"
793
+ "paddh $f28, $f28, $f0 \n\t"
794
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
795
+ "paddh $f30, $f30, $f2 \n\t"
796
+
797
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
798
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
799
+ "gsldlc1 $f2, 0x7(%[pSample1]) \n\t"
800
+
801
+ "gsldlc1 $f4, 0x7(%[pSample2]) \n\t"
802
+ "gsldlc1 $f8, 0x7($9) \n\t"
803
+ "gsldrc1 $f2, 0x0(%[pSample1]) \n\t"
804
+ "gsldrc1 $f4, 0x0(%[pSample2]) \n\t"
805
+ "gsldrc1 $f8, 0x0($9) \n\t"
806
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
807
+ PTR_ADDU "$9, $9, %[iStride2] \n\t"
808
+ "gsldlc1 $f6, 0x7(%[pSample2]) \n\t"
809
+ "gsldlc1 $f10, 0x7($9) \n\t"
810
+ "gsldrc1 $f6, 0x0(%[pSample2]) \n\t"
811
+ "gsldrc1 $f10, 0x0($9) \n\t"
812
+ "dsrl $f4, $f4, $f20 \n\t"
813
+ "dsrl $f6, $f6, $f20 \n\t"
814
+ "dsll $f8, $f8, $f24 \n\t"
815
+ "dsll $f10, $f10, $f24 \n\t"
816
+ "or $f4, $f4, $f8 \n\t"
817
+ "or $f6, $f6, $f10 \n\t"
818
+
819
+ "pasubub $f0, $f0, $f4 \n\t"
820
+ "pasubub $f2, $f2, $f6 \n\t"
821
+ "biadd $f0, $f0 \n\t"
822
+ "biadd $f2, $f2 \n\t"
823
+ "paddh $f28, $f28, $f0 \n\t"
824
+ "paddh $f30, $f30, $f2 \n\t"
825
+
826
+ "mov.d $f0, $f30 \n\t"
827
+ "paddh $f0, $f0, $f28 \n\t"
828
+ "dmfc1 %[iSadSum], $f0 \n\t"
829
+ "j 2f \n\t"
830
+ "nop \n\t"
831
+
832
+ "1: \n\t"
833
+ "xor $f24, $f24, $f24 \n\t"
834
+ "xor $f26, $f26, $f26 \n\t"
835
+ MMI_GetSad8x4
836
+ MMI_GetSad8x4_End
837
+ "paddh $f0, $f26, $f24 \n\t"
838
+ "dmfc1 %[iSadSum], $f0 \n\t"
839
+ "2: \n\t"
840
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSadSum]"=r"((int)iSadSum),
841
+ [pSample2]"+&r"((unsigned char *)pSample2)
842
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
843
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
844
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
845
+ );
846
+ RECOVER_REG;
847
+ return iSadSum;
848
+ }
849
+
850
+ int32_t WelsSampleSatd4x4_mmi (uint8_t* pSample1, int32_t iStride1,
851
+ uint8_t* pSample2, int32_t iStride2) {
852
+ int32_t iSatdSum = 0;
853
+ BACKUP_REG;
854
+ __asm__ volatile (
855
+ ".set arch=loongson3a \n\t"
856
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
857
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
858
+ "gsldlc1 $f4, 0x7($8) \n\t"
859
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
860
+ "gsldrc1 $f4, 0x0($8) \n\t"
861
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
862
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
863
+ "gsldlc1 $f8, 0x7(%[pSample1]) \n\t"
864
+ "gsldlc1 $f12, 0x7($8) \n\t"
865
+ "gsldrc1 $f8, 0x0(%[pSample1]) \n\t"
866
+ "gsldrc1 $f12, 0x0($8) \n\t"
867
+ "punpcklwd $f0, $f0, $f8 \n\t"
868
+ "punpcklwd $f4, $f4, $f12 \n\t"
869
+
870
+ PTR_ADDU "$8, %[pSample2], %[iStride2] \n\t"
871
+ "gsldlc1 $f16, 0x7(%[pSample2]) \n\t"
872
+ "gsldlc1 $f20, 0x7($8) \n\t"
873
+ "gsldrc1 $f16, 0x0(%[pSample2]) \n\t"
874
+ "gsldrc1 $f20, 0x0($8) \n\t"
875
+ PTR_ADDU "%[pSample2], $8, %[iStride2] \n\t"
876
+ PTR_ADDU "$8, %[pSample2], %[iStride2] \n\t"
877
+ "gsldlc1 $f24, 0x7(%[pSample2]) \n\t"
878
+ "gsldlc1 $f28, 0x7($8) \n\t"
879
+ "gsldrc1 $f24, 0x0(%[pSample2]) \n\t"
880
+ "gsldrc1 $f28, 0x0($8) \n\t"
881
+ "punpcklwd $f16, $f16, $f24 \n\t"
882
+ "punpcklwd $f20, $f20, $f28 \n\t"
883
+
884
+ "xor $f24, $f24, $f24 \n\t"
885
+ "xor $f26, $f26, $f26 \n\t"
886
+ "punpckhbh $f2, $f0, $f24 \n\t"
887
+ "punpcklbh $f0, $f0, $f24 \n\t"
888
+ "punpckhbh $f6, $f4, $f24 \n\t"
889
+ "punpcklbh $f4, $f4, $f24 \n\t"
890
+ "punpckhbh $f18, $f16, $f24 \n\t"
891
+ "punpcklbh $f16, $f16, $f24 \n\t"
892
+ "punpckhbh $f22, $f20, $f24 \n\t"
893
+ "punpcklbh $f20, $f20, $f24 \n\t"
894
+
895
+ "psubh $f0, $f0, $f16 \n\t"
896
+ "psubh $f2, $f2, $f18 \n\t"
897
+ "psubh $f4, $f4, $f20 \n\t"
898
+ "psubh $f6, $f6, $f22 \n\t"
899
+
900
+ "mov.d $f8, $f0 \n\t"
901
+ "mov.d $f10, $f2 \n\t"
902
+ "paddh $f0, $f0, $f4 \n\t"
903
+ "paddh $f2, $f2, $f6 \n\t"
904
+ "psubh $f8, $f8, $f4 \n\t"
905
+ "psubh $f10, $f10, $f6 \n\t"
906
+ MMI_XSawp_DQ($f0, $f2, $f8, $f10, $f12, $f14)
907
+
908
+ "mov.d $f16, $f0 \n\t"
909
+ "mov.d $f18, $f2 \n\t"
910
+ "paddh $f0, $f0, $f12 \n\t"
911
+ "paddh $f2, $f2, $f14 \n\t"
912
+ "psubh $f16, $f16, $f12 \n\t"
913
+ "psubh $f18, $f18, $f14 \n\t"
914
+
915
+ "mov.d $f8, $f2 \n\t"
916
+ "punpckhhw $f2, $f0, $f16 \n\t"
917
+ "punpcklhw $f0, $f0, $f16 \n\t"
918
+ "punpcklhw $f16, $f18, $f8 \n\t"
919
+ "punpckhhw $f18, $f18, $f8 \n\t"
920
+
921
+ MMI_XSawp_WD($f0, $f2, $f16, $f18, $f12, $f14)
922
+ MMI_XSawp_DQ($f0, $f2, $f12, $f14, $f20, $f22)
923
+
924
+ "mov.d $f28, $f0 \n\t"
925
+ "mov.d $f30, $f2 \n\t"
926
+ "paddh $f0, $f0, $f20 \n\t"
927
+ "paddh $f2, $f2, $f22 \n\t"
928
+ "psubh $f28, $f28, $f20 \n\t"
929
+ "psubh $f30, $f30, $f22 \n\t"
930
+
931
+ MMI_XSawp_DQ($f0, $f2, $f28, $f30, $f4, $f6)
932
+
933
+ "psubh $f8, $f0, $f4 \n\t"
934
+ "psubh $f10, $f2, $f6 \n\t"
935
+ "paddh $f0, $f0, $f4 \n\t"
936
+ "paddh $f2, $f2, $f6 \n\t"
937
+
938
+ WELS_AbsH($f0, $f2, $f0, $f2, $f12, $f14)
939
+ "paddush $f24, $f24, $f0 \n\t"
940
+ "paddush $f26, $f26, $f2 \n\t"
941
+ WELS_AbsH($f8, $f10, $f8, $f10, $f16, $f18)
942
+ "paddush $f24, $f24, $f8 \n\t"
943
+ "paddush $f26, $f26, $f10 \n\t"
944
+ MMI_SumWHorizon1($f24, $f26, $f16, $f18, $f28, $f30, $8)
945
+
946
+ "dmfc1 $8, $f24 \n\t"
947
+ "dli $9, 0xffff \n\t"
948
+ "and $8, $8, $9 \n\t"
949
+ "dsrl %[iSatdSum], $8, 0x1 \n\t"
950
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSatdSum]"=r"((int)iSatdSum),
951
+ [pSample2]"+&r"((unsigned char *)pSample2)
952
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
953
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
954
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
955
+ );
956
+ RECOVER_REG;
957
+ return iSatdSum;
958
+ }
959
+
960
+ int32_t WelsSampleSatd8x8_mmi (uint8_t* pSample1, int32_t iStride1,
961
+ uint8_t* pSample2, int32_t iStride2) {
962
+ int32_t iSatdSum = 0;
963
+ BACKUP_REG;
964
+ __asm__ volatile (
965
+ ".set arch=loongson3a \n\t"
966
+ "xor $f24, $f24, $f24 \n\t"
967
+ "xor $f26, $f26, $f26 \n\t"
968
+ "dli $8, 0x1 \n\t"
969
+ "xor $f28, $f28, $f28 \n\t"
970
+ "dmtc1 $8, $f30 \n\t"
971
+ MMI_GetSatd8x8_End
972
+ "psrlh $f24, $f24, $f30 \n\t"
973
+ "dli $8, 0x4e \n\t"
974
+ "psrlh $f26, $f26, $f30 \n\t"
975
+ "dmtc1 $8, $f30 \n\t"
976
+ MMI_SumWHorizon($f24, $f26, $f16, $f18, $f28, $f30)
977
+ "mfc1 %[iSatdSum], $f24 \n\t"
978
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSatdSum]"=r"((int)iSatdSum),
979
+ [pSample2]"+&r"((unsigned char *)pSample2)
980
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
981
+ : "memory", "$8", "$11", "$12", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10",
982
+ "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
983
+ );
984
+ RECOVER_REG;
985
+ return iSatdSum;
986
+ }
987
+
988
+ int32_t WelsSampleSatd8x16_mmi (uint8_t* pSample1, int32_t iStride1,
989
+ uint8_t* pSample2, int32_t iStride2) {
990
+ int32_t iSatdSum = 0;
991
+ BACKUP_REG;
992
+ __asm__ volatile (
993
+ ".set arch=loongson3a \n\t"
994
+ "xor $f24, $f24, $f24 \n\t"
995
+ "xor $f26, $f26, $f26 \n\t"
996
+ "dli $8, 0x1 \n\t"
997
+ "xor $f28, $f28, $f28 \n\t"
998
+ "dmtc1 $8, $f30 \n\t"
999
+ MMI_GetSatd8x8
1000
+ MMI_GetSatd8x8_End
1001
+ "psrlh $f24, $f24, $f30 \n\t"
1002
+ "dli $8, 0x4e \n\t"
1003
+ "psrlh $f26, $f26, $f30 \n\t"
1004
+ "dmtc1 $8, $f30 \n\t"
1005
+ MMI_SumWHorizon($f24, $f26, $f16, $f18, $f28, $f30)
1006
+ "mfc1 %[iSatdSum], $f24 \n\t"
1007
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSatdSum]"=r"((int)iSatdSum),
1008
+ [pSample2]"+&r"((unsigned char *)pSample2)
1009
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
1010
+ : "memory", "$8", "$11", "$12", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10",
1011
+ "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
1012
+ );
1013
+ RECOVER_REG;
1014
+ return iSatdSum;
1015
+ }
1016
+
1017
+ int32_t WelsSampleSatd16x8_mmi (uint8_t* pSample1, int32_t iStride1,
1018
+ uint8_t* pSample2, int32_t iStride2) {
1019
+ int32_t iSatdSum = 0;
1020
+ BACKUP_REG;
1021
+ __asm__ volatile (
1022
+ ".set arch=loongson3a \n\t"
1023
+ "xor $f24, $f24, $f24 \n\t"
1024
+ "xor $f26, $f26, $f26 \n\t"
1025
+ "dli $8, 0x1 \n\t"
1026
+ "xor $f28, $f28, $f28 \n\t"
1027
+ "dmtc1 $8, $f30 \n\t"
1028
+ "move $9, %[pSample1] \n\t"
1029
+ "move $10, %[pSample2] \n\t"
1030
+ MMI_GetSatd8x8_Offset8
1031
+
1032
+ MMI_GetSatd8x8_End
1033
+ "psrlh $f24, $f24, $f30 \n\t"
1034
+ "dli $8, 0x4e \n\t"
1035
+ "psrlh $f26, $f26, $f30 \n\t"
1036
+ "dmtc1 $8, $f30 \n\t"
1037
+ MMI_SumWHorizon($f24, $f26, $f16, $f18, $f28, $f30)
1038
+ "mfc1 %[iSatdSum], $f24 \n\t"
1039
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSatdSum]"=r"((int)iSatdSum),
1040
+ [pSample2]"+&r"((unsigned char *)pSample2)
1041
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
1042
+ : "memory", "$8", "$9", "$10", "$11", "$12", "$f0", "$f2", "$f4", "$f6",
1043
+ "$f8", "$f10", "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24",
1044
+ "$f26", "$f28", "$f30"
1045
+ );
1046
+ RECOVER_REG;
1047
+ return iSatdSum;
1048
+ }
1049
+
1050
+ int32_t WelsSampleSatd16x16_mmi (uint8_t* pSample1, int32_t iStride1,
1051
+ uint8_t* pSample2, int32_t iStride2) {
1052
+ int32_t iSatdSum = 0;
1053
+ BACKUP_REG;
1054
+ __asm__ volatile (
1055
+ ".set arch=loongson3a \n\t"
1056
+ "xor $f24, $f24, $f24 \n\t"
1057
+ "xor $f26, $f26, $f26 \n\t"
1058
+ "dli $8, 0x1 \n\t"
1059
+ "xor $f28, $f28, $f28 \n\t"
1060
+ "dmtc1 $8, $f30 \n\t"
1061
+ "move $9, %[pSample1] \n\t"
1062
+ "move $10, %[pSample2] \n\t"
1063
+
1064
+ MMI_GetSatd8x8
1065
+ MMI_GetSatd8x8_Offset8
1066
+
1067
+ MMI_GetSatd8x8
1068
+ MMI_GetSatd8x8_End
1069
+
1070
+ "dli $8, 0x4e \n\t"
1071
+ "psrlh $f24, $f24, $f30 \n\t"
1072
+ "dmtc1 $8, $f0 \n\t"
1073
+ "psrlh $f26, $f26, $f30 \n\t"
1074
+ MMI_SumWHorizon($f24, $f26, $f16, $f18, $f28, $f0)
1075
+ "mfc1 %[iSatdSum], $f24 \n\t"
1076
+ : [pSample1]"+&r"((unsigned char *)pSample1), [iSatdSum]"=r"((int)iSatdSum),
1077
+ [pSample2]"+&r"((unsigned char *)pSample2)
1078
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2)
1079
+ : "memory", "$8", "$9", "$10", "$11", "$12", "$f0", "$f2", "$f4", "$f6",
1080
+ "$f8", "$f10", "$f12", "$f14", "$f16", "$f18", "$f20", "$f22", "$f24",
1081
+ "$f26", "$f28", "$f30"
1082
+ );
1083
+ RECOVER_REG;
1084
+ return iSatdSum;
1085
+ }
1086
+
1087
+ void WelsSampleSadFour16x16_mmi (uint8_t* pSample1, int32_t iStride1, uint8_t* pSample2,
1088
+ int32_t iStride2, int32_t* pSad) {
1089
+ BACKUP_REG;
1090
+ __asm__ volatile (
1091
+ ".set arch=loongson3a \n\t"
1092
+ "xor $f16, $f16, $f16 \n\t"
1093
+ "xor $f18, $f18, $f18 \n\t"
1094
+ "xor $f20, $f20, $f20 \n\t"
1095
+ "xor $f22, $f22, $f22 \n\t"
1096
+ PTR_SUBU "%[pSample2], %[pSample2], %[iStride2] \n\t"
1097
+ "xor $f24, $f24, $f24 \n\t"
1098
+ "xor $f26, $f26, $f26 \n\t"
1099
+ "xor $f28, $f28, $f28 \n\t"
1100
+ "xor $f30, $f30, $f30 \n\t"
1101
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1102
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1103
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1104
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1105
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1106
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
1107
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
1108
+ "pasubub $f12, $f12, $f0 \n\t"
1109
+ "pasubub $f14, $f14, $f2 \n\t"
1110
+ "biadd $f12, $f12 \n\t"
1111
+ "biadd $f14, $f14 \n\t"
1112
+ "paddh $f16, $f16, $f12 \n\t"
1113
+ "paddh $f18, $f18, $f14 \n\t"
1114
+
1115
+ "gslqc1 $f6, $f4, 0x0(%[pSample1]) \n\t"
1116
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1117
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1118
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1119
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1120
+ "pasubub $f12, $f12, $f4 \n\t"
1121
+ "pasubub $f14, $f14, $f6 \n\t"
1122
+ "biadd $f12, $f12 \n\t"
1123
+ "biadd $f14, $f14 \n\t"
1124
+ "paddh $f16, $f16, $f12 \n\t"
1125
+ "paddh $f18, $f18, $f14 \n\t"
1126
+
1127
+ "gsldlc1 $f8, 0x6(%[pSample2]) \n\t"
1128
+ "gsldlc1 $f10, 0xE(%[pSample2]) \n\t"
1129
+ "gsldrc1 $f8, -0x1(%[pSample2]) \n\t"
1130
+ "gsldrc1 $f10, 0x7(%[pSample2]) \n\t"
1131
+ "pasubub $f8, $f8, $f0 \n\t"
1132
+ "pasubub $f10, $f10, $f2 \n\t"
1133
+ "biadd $f8, $f8 \n\t"
1134
+ "biadd $f10, $f10 \n\t"
1135
+ "paddh $f24, $f24, $f8 \n\t"
1136
+ "paddh $f26, $f26, $f10 \n\t"
1137
+
1138
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1139
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1140
+ "gsldlc1 $f14, 0x10(%[pSample2]) \n\t"
1141
+ "gsldrc1 $f14, 0x9(%[pSample2]) \n\t"
1142
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1143
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1144
+ "pasubub $f12, $f12, $f0 \n\t"
1145
+ "pasubub $f14, $f14, $f2 \n\t"
1146
+ "biadd $f12, $f12 \n\t"
1147
+ "biadd $f14, $f14 \n\t"
1148
+ "paddh $f28, $f28, $f12 \n\t"
1149
+ "paddh $f30, $f30, $f14 \n\t"
1150
+
1151
+ "gslqc1 $f10, $f8, 0x0($8) \n\t"
1152
+ "gsldlc1 $f12, 0x7($9) \n\t"
1153
+ "gsldlc1 $f14, 0xF($9) \n\t"
1154
+ "gsldrc1 $f12, 0x0($9) \n\t"
1155
+ "gsldrc1 $f14, 0x8($9) \n\t"
1156
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1157
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1158
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $9)
1159
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1160
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1161
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1162
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1163
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1164
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1165
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1166
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, %[pSample2])
1167
+ "gslqc1 $f6, $f4, 0x0($8) \n\t"
1168
+ "gsldlc1 $f12, 0x7($9) \n\t"
1169
+ "gsldlc1 $f14, 0xF($9) \n\t"
1170
+ "gsldrc1 $f12, 0x0($9) \n\t"
1171
+ "gsldrc1 $f14, 0x8($9) \n\t"
1172
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1173
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1174
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, $9)
1175
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t"
1176
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1177
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1178
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1179
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1180
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1181
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1182
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, %[pSample2])
1183
+ "gslqc1 $f2, $f0, 0x0($8) \n\t"
1184
+ "gsldlc1 $f12, 0x7($9) \n\t"
1185
+ "gsldlc1 $f14, 0xF($9) \n\t"
1186
+ "gsldrc1 $f12, 0x0($9) \n\t"
1187
+ "gsldrc1 $f14, 0x8($9) \n\t"
1188
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1189
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1190
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, $9)
1191
+ "gslqc1 $f6, $f4, 0x0(%[pSample1]) \n\t"
1192
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1193
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1194
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1195
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1196
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1197
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1198
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, %[pSample2])
1199
+
1200
+ "gslqc1 $f10, $f8, 0x0($8) \n\t"
1201
+ "gsldlc1 $f12, 0x7($9) \n\t"
1202
+ "gsldlc1 $f14, 0xF($9) \n\t"
1203
+ "gsldrc1 $f12, 0x0($9) \n\t"
1204
+ "gsldrc1 $f14, 0x8($9) \n\t"
1205
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1206
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1207
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $9)
1208
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1209
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1210
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1211
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1212
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1213
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1214
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1215
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, %[pSample2])
1216
+ "gslqc1 $f6, $f4, 0x0($8) \n\t"
1217
+ "gsldlc1 $f12, 0x7($9) \n\t"
1218
+ "gsldlc1 $f14, 0xF($9) \n\t"
1219
+ "gsldrc1 $f12, 0x0($9) \n\t"
1220
+ "gsldrc1 $f14, 0x8($9) \n\t"
1221
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1222
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1223
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, $9)
1224
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t"
1225
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1226
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1227
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1228
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1229
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1230
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1231
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, %[pSample2])
1232
+ "gslqc1 $f2, $f0, 0x0($8) \n\t"
1233
+ "gsldlc1 $f12, 0x7($9) \n\t"
1234
+ "gsldlc1 $f14, 0xF($9) \n\t"
1235
+ "gsldrc1 $f12, 0x0($9) \n\t"
1236
+ "gsldrc1 $f14, 0x8($9) \n\t"
1237
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1238
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1239
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, $9)
1240
+ "gslqc1 $f6, $f4, 0x0(%[pSample1]) \n\t"
1241
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1242
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1243
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1244
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1245
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1246
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1247
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, %[pSample2])
1248
+
1249
+ "gslqc1 $f10, $f8, 0x0($8) \n\t"
1250
+ "gsldlc1 $f12, 0x7($9) \n\t"
1251
+ "gsldlc1 $f14, 0xF($9) \n\t"
1252
+ "gsldrc1 $f12, 0x0($9) \n\t"
1253
+ "gsldrc1 $f14, 0x8($9) \n\t"
1254
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1255
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1256
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $9)
1257
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1258
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1259
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1260
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1261
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1262
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1263
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, %[pSample2])
1264
+ "gsldlc1 $f12, 0x7($9) \n\t"
1265
+ "gsldlc1 $f14, 0xF($9) \n\t"
1266
+ "gsldrc1 $f12, 0x0($9) \n\t"
1267
+ "gsldrc1 $f14, 0x8($9) \n\t"
1268
+ "pasubub $f8, $f8, $f12 \n\t"
1269
+ "pasubub $f10, $f10, $f14 \n\t"
1270
+ "biadd $f8, $f8 \n\t"
1271
+ "biadd $f10, $f10 \n\t"
1272
+ "paddh $f20, $f20, $f8 \n\t"
1273
+ "paddh $f22, $f22, $f10 \n\t"
1274
+
1275
+ "gsldlc1 $f8, 0x6($9) \n\t"
1276
+ "gsldlc1 $f10, 0xE($9) \n\t"
1277
+ "gsldrc1 $f8, -0x1($9) \n\t"
1278
+ "gsldrc1 $f10, 0x7($9) \n\t"
1279
+ "pasubub $f8, $f8, $f0 \n\t"
1280
+ "pasubub $f10, $f10, $f2 \n\t"
1281
+ "biadd $f8, $f8 \n\t"
1282
+ "biadd $f10, $f10 \n\t"
1283
+ "paddh $f24, $f24, $f8 \n\t"
1284
+ "paddh $f26, $f26, $f10 \n\t"
1285
+
1286
+ "gsldlc1 $f12, 0x8($9) \n\t"
1287
+ "gsldlc1 $f14, 0x10($9) \n\t"
1288
+ "gsldrc1 $f12, 0x1($9) \n\t"
1289
+ "gsldrc1 $f14, 0x9($9) \n\t"
1290
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1291
+ "pasubub $f12, $f12, $f0 \n\t"
1292
+ "pasubub $f14, $f14, $f2 \n\t"
1293
+ "biadd $f12, $f12 \n\t"
1294
+ "biadd $f14, $f14 \n\t"
1295
+ "paddh $f28, $f28, $f12 \n\t"
1296
+ "paddh $f30, $f30, $f14 \n\t"
1297
+
1298
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1299
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1300
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1301
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1302
+ "pasubub $f0, $f0, $f12 \n\t"
1303
+ "pasubub $f2, $f2, $f14 \n\t"
1304
+ "biadd $f0, $f0 \n\t"
1305
+ "biadd $f2, $f2 \n\t"
1306
+ "paddh $f20, $f20, $f0 \n\t"
1307
+ "paddh $f22, $f22, $f2 \n\t"
1308
+
1309
+ "paddh $f16, $f16, $f18 \n\t"
1310
+ "paddh $f20, $f20, $f22 \n\t"
1311
+ "paddh $f24, $f24, $f26 \n\t"
1312
+ "paddh $f28, $f28, $f30 \n\t"
1313
+ "punpcklwd $f16, $f16, $f20 \n\t"
1314
+ "punpcklwd $f24, $f24, $f28 \n\t"
1315
+ "gssqc1 $f24, $f16, 0x0(%[pSad]) \n\t"
1316
+ : [pSample1]"+&r"((unsigned char *)pSample1),
1317
+ [pSample2]"+&r"((unsigned char *)pSample2)
1318
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2),
1319
+ [pSad]"r"((int *)pSad)
1320
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
1321
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
1322
+ );
1323
+ RECOVER_REG;
1324
+ }
1325
+
1326
+ void WelsSampleSadFour16x8_mmi (uint8_t* pSample1, int32_t iStride1, uint8_t* pSample2,
1327
+ int32_t iStride2, int32_t* pSad) {
1328
+ BACKUP_REG;
1329
+ __asm__ volatile (
1330
+ ".set arch=loongson3a \n\t"
1331
+ "xor $f16, $f16, $f16 \n\t"
1332
+ "xor $f18, $f18, $f18 \n\t"
1333
+ "xor $f20, $f20, $f20 \n\t"
1334
+ "xor $f22, $f22, $f22 \n\t"
1335
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1336
+ PTR_SUBU "%[pSample2], %[pSample2], %[iStride2] \n\t"
1337
+ "xor $f24, $f24, $f24 \n\t"
1338
+ "xor $f26, $f26, $f26 \n\t"
1339
+ "xor $f28, $f28, $f28 \n\t"
1340
+ "xor $f30, $f30, $f30 \n\t"
1341
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1342
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1343
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1344
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1345
+ PTR_ADDU "%[pSample1], %[pSample1], %[iStride1] \n\t"
1346
+ "pasubub $f12, $f12, $f0 \n\t"
1347
+ "pasubub $f14, $f14, $f2 \n\t"
1348
+ PTR_ADDU "%[pSample2], %[pSample2], %[iStride2] \n\t"
1349
+ "biadd $f12, $f12 \n\t"
1350
+ "biadd $f14, $f14 \n\t"
1351
+ "paddh $f16, $f16, $f12 \n\t"
1352
+ "paddh $f18, $f18, $f14 \n\t"
1353
+
1354
+ "gslqc1 $f6, $f4, 0x0(%[pSample1]) \n\t"
1355
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1356
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1357
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1358
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1359
+ "pasubub $f12, $f12, $f4 \n\t"
1360
+ "pasubub $f14, $f14, $f6 \n\t"
1361
+ "biadd $f12, $f12 \n\t"
1362
+ "biadd $f14, $f14 \n\t"
1363
+ "paddh $f16, $f16, $f12 \n\t"
1364
+ "paddh $f18, $f18, $f14 \n\t"
1365
+
1366
+ "gsldlc1 $f8, 0x6(%[pSample2]) \n\t"
1367
+ "gsldlc1 $f10, 0xE(%[pSample2]) \n\t"
1368
+ "gsldrc1 $f8, -0x1(%[pSample2]) \n\t"
1369
+ "gsldrc1 $f10, 0x7(%[pSample2]) \n\t"
1370
+ "pasubub $f8, $f8, $f0 \n\t"
1371
+ "pasubub $f10, $f10, $f2 \n\t"
1372
+ "biadd $f8, $f8 \n\t"
1373
+ "biadd $f10, $f10 \n\t"
1374
+ "paddh $f24, $f24, $f8 \n\t"
1375
+ "paddh $f26, $f26, $f10 \n\t"
1376
+
1377
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1378
+ "gsldlc1 $f14, 0x10(%[pSample2]) \n\t"
1379
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1380
+ "gsldrc1 $f14, 0x9(%[pSample2]) \n\t"
1381
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1382
+ "pasubub $f12, $f12, $f0 \n\t"
1383
+ "pasubub $f14, $f14, $f2 \n\t"
1384
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1385
+ "biadd $f12, $f12 \n\t"
1386
+ "biadd $f14, $f14 \n\t"
1387
+ "paddh $f28, $f28, $f12 \n\t"
1388
+ "paddh $f30, $f30, $f14 \n\t"
1389
+
1390
+ "gslqc1 $f10, $f8, 0x0($8) \n\t"
1391
+ "gsldlc1 $f12, 0x7($9) \n\t"
1392
+ "gsldlc1 $f14, 0xF($9) \n\t"
1393
+ "gsldrc1 $f12, 0x0($9) \n\t"
1394
+ "gsldrc1 $f14, 0x8($9) \n\t"
1395
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1396
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1397
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, $9)
1398
+ "gslqc1 $f2, $f0, 0x0(%[pSample1]) \n\t"
1399
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1400
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1401
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1402
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1403
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1404
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1405
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, %[pSample2])
1406
+ "gslqc1 $f6, $f4, 0x0($8) \n\t"
1407
+ "gsldlc1 $f12, 0x7($9) \n\t"
1408
+ "gsldlc1 $f14, 0xF($9) \n\t"
1409
+ "gsldrc1 $f12, 0x0($9) \n\t"
1410
+ "gsldrc1 $f14, 0x8($9) \n\t"
1411
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1412
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1413
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, $9)
1414
+ "gslqc1 $f10, $f8, 0x0(%[pSample1]) \n\t"
1415
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1416
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1417
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1418
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1419
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1420
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1421
+ MMI_Get4LW16Sad($f0, $f2, $f4, $f6, $f8, $f10, $f12, $f14, %[pSample2])
1422
+ "gslqc1 $f2, $f0, 0x0($8) \n\t"
1423
+ "gsldlc1 $f12, 0x7($9) \n\t"
1424
+ "gsldlc1 $f14, 0xF($9) \n\t"
1425
+ "gsldrc1 $f12, 0x0($9) \n\t"
1426
+ "gsldrc1 $f14, 0x8($9) \n\t"
1427
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1428
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1429
+ MMI_Get4LW16Sad($f4, $f6, $f8, $f10, $f0, $f2, $f12, $f14, $9)
1430
+ "gslqc1 $f6, $f4, 0x0(%[pSample1]) \n\t"
1431
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1432
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1433
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1434
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1435
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1436
+ MMI_Get4LW16Sad($f8, $f10, $f0, $f2, $f4, $f6, $f12, $f14, %[pSample2])
1437
+ "gsldlc1 $f12, 0x7($9) \n\t"
1438
+ "gsldlc1 $f14, 0xF($9) \n\t"
1439
+ "gsldrc1 $f12, 0x0($9) \n\t"
1440
+ "gsldrc1 $f14, 0x8($9) \n\t"
1441
+ "pasubub $f0, $f0, $f12 \n\t"
1442
+ "pasubub $f2, $f2, $f14 \n\t"
1443
+ "biadd $f0, $f0 \n\t"
1444
+ "biadd $f2, $f2 \n\t"
1445
+ "paddh $f20, $f20, $f0 \n\t"
1446
+ "paddh $f22, $f22, $f2 \n\t"
1447
+
1448
+ "gsldlc1 $f0, 0x6($9) \n\t"
1449
+ "gsldlc1 $f2, 0xE($9) \n\t"
1450
+ "gsldrc1 $f0, -0x1($9) \n\t"
1451
+ "gsldrc1 $f2, 0x7($9) \n\t"
1452
+ "pasubub $f0, $f0, $f4 \n\t"
1453
+ "pasubub $f2, $f2, $f6 \n\t"
1454
+ "biadd $f0, $f0 \n\t"
1455
+ "biadd $f2, $f2 \n\t"
1456
+ "paddh $f24, $f24, $f0 \n\t"
1457
+ "paddh $f26, $f26, $f2 \n\t"
1458
+
1459
+ "gsldlc1 $f12, 0x8($9) \n\t"
1460
+ "gsldlc1 $f14, 0x10($9) \n\t"
1461
+ "gsldrc1 $f12, 0x1($9) \n\t"
1462
+ "gsldrc1 $f14, 0x9($9) \n\t"
1463
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1464
+ "pasubub $f12, $f12, $f4 \n\t"
1465
+ "pasubub $f14, $f14, $f6 \n\t"
1466
+ "biadd $f12, $f12 \n\t"
1467
+ "biadd $f14, $f14 \n\t"
1468
+ "paddh $f28, $f28, $f12 \n\t"
1469
+ "paddh $f30, $f30, $f14 \n\t"
1470
+
1471
+ "gsldlc1 $f12, 0x7(%[pSample2]) \n\t"
1472
+ "gsldlc1 $f14, 0xF(%[pSample2]) \n\t"
1473
+ "gsldrc1 $f12, 0x0(%[pSample2]) \n\t"
1474
+ "gsldrc1 $f14, 0x8(%[pSample2]) \n\t"
1475
+ "pasubub $f4, $f4, $f12 \n\t"
1476
+ "pasubub $f6, $f6, $f14 \n\t"
1477
+ "biadd $f4, $f4 \n\t"
1478
+ "biadd $f6, $f6 \n\t"
1479
+ "paddh $f20, $f20, $f4 \n\t"
1480
+ "paddh $f22, $f22, $f6 \n\t"
1481
+
1482
+ "paddh $f16, $f16, $f18 \n\t"
1483
+ "paddh $f20, $f20, $f22 \n\t"
1484
+ "paddh $f24, $f24, $f26 \n\t"
1485
+ "paddh $f28, $f28, $f30 \n\t"
1486
+ "punpcklwd $f16, $f16, $f20 \n\t"
1487
+ "punpcklwd $f24, $f24, $f28 \n\t"
1488
+ "gssqc1 $f24, $f16, 0x0(%[pSad]) \n\t"
1489
+ : [pSample1]"+&r"((unsigned char *)pSample1),
1490
+ [pSample2]"+&r"((unsigned char *)pSample2)
1491
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2),
1492
+ [pSad]"r"((int *)pSad)
1493
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
1494
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28","$f30"
1495
+ );
1496
+ RECOVER_REG;
1497
+ }
1498
+
1499
+ void WelsSampleSadFour8x16_mmi (uint8_t* pSample1, int32_t iStride1, uint8_t* pSample2,
1500
+ int32_t iStride2, int32_t* pSad) {
1501
+ BACKUP_REG;
1502
+ __asm__ volatile (
1503
+ ".set arch=loongson3a \n\t"
1504
+ "xor $f16, $f16, $f16 \n\t"
1505
+ "xor $f18, $f18, $f18 \n\t"
1506
+ "xor $f20, $f20, $f20 \n\t"
1507
+ "xor $f22, $f22, $f22 \n\t"
1508
+ "xor $f24, $f24, $f24 \n\t"
1509
+ "xor $f26, $f26, $f26 \n\t"
1510
+ "xor $f28, $f28, $f28 \n\t"
1511
+ "xor $f30, $f30, $f30 \n\t"
1512
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1513
+ PTR_SUBU "$9, %[pSample2], %[iStride2] \n\t"
1514
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1515
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1516
+ "gsldlc1 $f2, 0x7($8) \n\t"
1517
+ "gsldlc1 $f12, 0x7($9) \n\t"
1518
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1519
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1520
+ "gsldrc1 $f2, 0x0($8) \n\t"
1521
+ "gsldrc1 $f12, 0x0($9) \n\t"
1522
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1523
+ "pasubub $f12, $f12, $f0 \n\t"
1524
+ "pasubub $f14, $f14, $f2 \n\t"
1525
+ "biadd $f12, $f12 \n\t"
1526
+ "biadd $f14, $f14 \n\t"
1527
+ "paddh $f16, $f16, $f12 \n\t"
1528
+ "paddh $f18, $f18, $f14 \n\t"
1529
+
1530
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1531
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1532
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1533
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1534
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1535
+
1536
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1537
+ "gsldlc1 $f6, 0x6($9) \n\t"
1538
+ "gsldlc1 $f14, 0x8($9) \n\t"
1539
+ "gsldrc1 $f6, -0x1($9) \n\t"
1540
+ "gsldrc1 $f14, 0x1($9) \n\t"
1541
+ "pasubub $f4, $f4, $f0 \n\t"
1542
+ "pasubub $f6, $f6, $f2 \n\t"
1543
+ "biadd $f4, $f4 \n\t"
1544
+ "biadd $f6, $f6 \n\t"
1545
+ "paddh $f24, $f24, $f4 \n\t"
1546
+ "paddh $f26, $f26, $f6 \n\t"
1547
+ "pasubub $f12, $f12, $f0 \n\t"
1548
+ "pasubub $f14, $f14, $f2 \n\t"
1549
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1550
+ "biadd $f12, $f12 \n\t"
1551
+ "biadd $f14, $f14 \n\t"
1552
+ "paddh $f28, $f28, $f12 \n\t"
1553
+ "paddh $f30, $f30, $f14 \n\t"
1554
+
1555
+ "gsldlc1 $f12, 0x7($9) \n\t"
1556
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1557
+ "gsldrc1 $f12, 0x0($9) \n\t"
1558
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1559
+ "pasubub $f0, $f0, $f12 \n\t"
1560
+ "pasubub $f2, $f2, $f14 \n\t"
1561
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1562
+ "biadd $f0, $f0 \n\t"
1563
+ "biadd $f2, $f2 \n\t"
1564
+ "paddh $f20, $f20, $f0 \n\t"
1565
+ "paddh $f22, $f22, $f2 \n\t"
1566
+
1567
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1568
+ "gsldlc1 $f2, 0x7($8) \n\t"
1569
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1570
+ "gsldrc1 $f2, 0x0($8) \n\t"
1571
+ "pasubub $f12, $f12, $f0 \n\t"
1572
+ "pasubub $f14, $f14, $f2 \n\t"
1573
+ "biadd $f12, $f12 \n\t"
1574
+ "biadd $f14, $f14 \n\t"
1575
+ "paddh $f16, $f16, $f12 \n\t"
1576
+ "paddh $f18, $f18, $f14 \n\t"
1577
+
1578
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1579
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1580
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1581
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1582
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1583
+
1584
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1585
+ "gsldlc1 $f6, 0x6($9) \n\t"
1586
+ "gsldlc1 $f14, 0x8($9) \n\t"
1587
+ "gsldrc1 $f6, -0x1($9) \n\t"
1588
+ "gsldrc1 $f14, 0x1($9) \n\t"
1589
+
1590
+ "pasubub $f4, $f4, $f0 \n\t"
1591
+ "pasubub $f6, $f6, $f2 \n\t"
1592
+ "biadd $f4, $f4 \n\t"
1593
+ "biadd $f6, $f6 \n\t"
1594
+ "paddh $f24, $f24, $f4 \n\t"
1595
+ "paddh $f26, $f26, $f6 \n\t"
1596
+ "pasubub $f12, $f12, $f0 \n\t"
1597
+ "pasubub $f14, $f14, $f2 \n\t"
1598
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1599
+ "biadd $f12, $f12 \n\t"
1600
+ "biadd $f14, $f14 \n\t"
1601
+ "paddh $f28, $f28, $f12 \n\t"
1602
+ "paddh $f30, $f30, $f14 \n\t"
1603
+
1604
+ "gsldlc1 $f12, 0x7($9) \n\t"
1605
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1606
+ "gsldrc1 $f12, 0x0($9) \n\t"
1607
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1608
+ "pasubub $f0, $f0, $f12 \n\t"
1609
+ "pasubub $f2, $f2, $f14 \n\t"
1610
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1611
+ "biadd $f0, $f0 \n\t"
1612
+ "biadd $f2, $f2 \n\t"
1613
+ "paddh $f20, $f20, $f0 \n\t"
1614
+ "paddh $f22, $f22, $f2 \n\t"
1615
+
1616
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1617
+ "gsldlc1 $f2, 0x7($8) \n\t"
1618
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1619
+ "gsldrc1 $f2, 0x0($8) \n\t"
1620
+ "pasubub $f12, $f12, $f0 \n\t"
1621
+ "pasubub $f14, $f14, $f2 \n\t"
1622
+ "biadd $f12, $f12 \n\t"
1623
+ "biadd $f14, $f14 \n\t"
1624
+ "paddh $f16, $f16, $f12 \n\t"
1625
+ "paddh $f18, $f18, $f14 \n\t"
1626
+
1627
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1628
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1629
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1630
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1631
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1632
+
1633
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1634
+ "gsldlc1 $f6, 0x6($9) \n\t"
1635
+ "gsldlc1 $f14, 0x8($9) \n\t"
1636
+ "gsldrc1 $f6, -0x1($9) \n\t"
1637
+ "gsldrc1 $f14, 0x1($9) \n\t"
1638
+
1639
+ "pasubub $f4, $f4, $f0 \n\t"
1640
+ "pasubub $f6, $f6, $f2 \n\t"
1641
+ "biadd $f4, $f4 \n\t"
1642
+ "biadd $f6, $f6 \n\t"
1643
+ "paddh $f24, $f24, $f4 \n\t"
1644
+ "paddh $f26, $f26, $f6 \n\t"
1645
+ "pasubub $f12, $f12, $f0 \n\t"
1646
+ "pasubub $f14, $f14, $f2 \n\t"
1647
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1648
+ "biadd $f12, $f12 \n\t"
1649
+ "biadd $f14, $f14 \n\t"
1650
+ "paddh $f28, $f28, $f12 \n\t"
1651
+ "paddh $f30, $f30, $f14 \n\t"
1652
+
1653
+ "gsldlc1 $f12, 0x7($9) \n\t"
1654
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1655
+ "gsldrc1 $f12, 0x0($9) \n\t"
1656
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1657
+ "pasubub $f0, $f0, $f12 \n\t"
1658
+ "pasubub $f2, $f2, $f14 \n\t"
1659
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1660
+ "biadd $f0, $f0 \n\t"
1661
+ "biadd $f2, $f2 \n\t"
1662
+ "paddh $f20, $f20, $f0 \n\t"
1663
+ "paddh $f22, $f22, $f2 \n\t"
1664
+
1665
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1666
+ "gsldlc1 $f2, 0x7($8) \n\t"
1667
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1668
+ "gsldrc1 $f2, 0x0($8) \n\t"
1669
+ "pasubub $f12, $f12, $f0 \n\t"
1670
+ "pasubub $f14, $f14, $f2 \n\t"
1671
+ "biadd $f12, $f12 \n\t"
1672
+ "biadd $f14, $f14 \n\t"
1673
+ "paddh $f16, $f16, $f12 \n\t"
1674
+ "paddh $f18, $f18, $f14 \n\t"
1675
+
1676
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1677
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1678
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1679
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1680
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1681
+
1682
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1683
+ "gsldlc1 $f6, 0x6($9) \n\t"
1684
+ "gsldlc1 $f14, 0x8($9) \n\t"
1685
+ "gsldrc1 $f6, -0x1($9) \n\t"
1686
+ "gsldrc1 $f14, 0x1($9) \n\t"
1687
+ "pasubub $f4, $f4, $f0 \n\t"
1688
+ "pasubub $f6, $f6, $f2 \n\t"
1689
+ "biadd $f4, $f4 \n\t"
1690
+ "biadd $f6, $f6 \n\t"
1691
+ "paddh $f24, $f24, $f4 \n\t"
1692
+ "paddh $f26, $f26, $f6 \n\t"
1693
+ "pasubub $f12, $f12, $f0 \n\t"
1694
+ "pasubub $f14, $f14, $f2 \n\t"
1695
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1696
+ "biadd $f12, $f12 \n\t"
1697
+ "biadd $f14, $f14 \n\t"
1698
+ "paddh $f28, $f28, $f12 \n\t"
1699
+ "paddh $f30, $f30, $f14 \n\t"
1700
+
1701
+ "gsldlc1 $f12, 0x7($9) \n\t"
1702
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1703
+ "gsldrc1 $f12, 0x0($9) \n\t"
1704
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1705
+ "pasubub $f0, $f0, $f12 \n\t"
1706
+ "pasubub $f2, $f2, $f14 \n\t"
1707
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1708
+ "biadd $f0, $f0 \n\t"
1709
+ "biadd $f2, $f2 \n\t"
1710
+ "paddh $f20, $f20, $f0 \n\t"
1711
+ "paddh $f22, $f22, $f2 \n\t"
1712
+
1713
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1714
+ "gsldlc1 $f2, 0x7($8) \n\t"
1715
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1716
+ "gsldrc1 $f2, 0x0($8) \n\t"
1717
+ "pasubub $f12, $f12, $f0 \n\t"
1718
+ "pasubub $f14, $f14, $f2 \n\t"
1719
+ "biadd $f12, $f12 \n\t"
1720
+ "biadd $f14, $f14 \n\t"
1721
+ "paddh $f16, $f16, $f12 \n\t"
1722
+ "paddh $f18, $f18, $f14 \n\t"
1723
+
1724
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1725
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1726
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1727
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1728
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1729
+
1730
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1731
+ "gsldlc1 $f6, 0x6($9) \n\t"
1732
+ "gsldlc1 $f14, 0x8($9) \n\t"
1733
+ "gsldrc1 $f6, -0x1($9) \n\t"
1734
+ "gsldrc1 $f14, 0x1($9) \n\t"
1735
+
1736
+ "pasubub $f4, $f4, $f0 \n\t"
1737
+ "pasubub $f6, $f6, $f2 \n\t"
1738
+ "biadd $f4, $f4 \n\t"
1739
+ "biadd $f6, $f6 \n\t"
1740
+ "paddh $f24, $f24, $f4 \n\t"
1741
+ "paddh $f26, $f26, $f6 \n\t"
1742
+ "pasubub $f12, $f12, $f0 \n\t"
1743
+ "pasubub $f14, $f14, $f2 \n\t"
1744
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1745
+ "biadd $f12, $f12 \n\t"
1746
+ "biadd $f14, $f14 \n\t"
1747
+ "paddh $f28, $f28, $f12 \n\t"
1748
+ "paddh $f30, $f30, $f14 \n\t"
1749
+
1750
+ "gsldlc1 $f12, 0x7($9) \n\t"
1751
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1752
+ "gsldrc1 $f12, 0x0($9) \n\t"
1753
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1754
+ "pasubub $f0, $f0, $f12 \n\t"
1755
+ "pasubub $f2, $f2, $f14 \n\t"
1756
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1757
+ "biadd $f0, $f0 \n\t"
1758
+ "biadd $f2, $f2 \n\t"
1759
+ "paddh $f20, $f20, $f0 \n\t"
1760
+ "paddh $f22, $f22, $f2 \n\t"
1761
+
1762
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1763
+ "gsldlc1 $f2, 0x7($8) \n\t"
1764
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1765
+ "gsldrc1 $f2, 0x0($8) \n\t"
1766
+ "pasubub $f12, $f12, $f0 \n\t"
1767
+ "pasubub $f14, $f14, $f2 \n\t"
1768
+ "biadd $f12, $f12 \n\t"
1769
+ "biadd $f14, $f14 \n\t"
1770
+ "paddh $f16, $f16, $f12 \n\t"
1771
+ "paddh $f18, $f18, $f14 \n\t"
1772
+
1773
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1774
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1775
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1776
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1777
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1778
+
1779
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1780
+ "gsldlc1 $f6, 0x6($9) \n\t"
1781
+ "gsldlc1 $f14, 0x8($9) \n\t"
1782
+ "gsldrc1 $f6, -0x1($9) \n\t"
1783
+ "gsldrc1 $f14, 0x1($9) \n\t"
1784
+
1785
+ "pasubub $f4, $f4, $f0 \n\t"
1786
+ "pasubub $f6, $f6, $f2 \n\t"
1787
+ "biadd $f4, $f4 \n\t"
1788
+ "biadd $f6, $f6 \n\t"
1789
+ "paddh $f24, $f24, $f4 \n\t"
1790
+ "paddh $f26, $f26, $f6 \n\t"
1791
+ "pasubub $f12, $f12, $f0 \n\t"
1792
+ "pasubub $f14, $f14, $f2 \n\t"
1793
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1794
+ "biadd $f12, $f12 \n\t"
1795
+ "biadd $f14, $f14 \n\t"
1796
+ "paddh $f28, $f28, $f12 \n\t"
1797
+ "paddh $f30, $f30, $f14 \n\t"
1798
+
1799
+ "gsldlc1 $f12, 0x7($9) \n\t"
1800
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1801
+ "gsldrc1 $f12, 0x0($9) \n\t"
1802
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1803
+ "pasubub $f0, $f0, $f12 \n\t"
1804
+ "pasubub $f2, $f2, $f14 \n\t"
1805
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1806
+ "biadd $f0, $f0 \n\t"
1807
+ "biadd $f2, $f2 \n\t"
1808
+ "paddh $f20, $f20, $f0 \n\t"
1809
+ "paddh $f22, $f22, $f2 \n\t"
1810
+
1811
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1812
+ "gsldlc1 $f2, 0x7($8) \n\t"
1813
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1814
+ "gsldrc1 $f2, 0x0($8) \n\t"
1815
+ "pasubub $f12, $f12, $f0 \n\t"
1816
+ "pasubub $f14, $f14, $f2 \n\t"
1817
+ "biadd $f12, $f12 \n\t"
1818
+ "biadd $f14, $f14 \n\t"
1819
+ "paddh $f16, $f16, $f12 \n\t"
1820
+ "paddh $f18, $f18, $f14 \n\t"
1821
+
1822
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1823
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1824
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1825
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1826
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1827
+
1828
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1829
+ "gsldlc1 $f6, 0x6($9) \n\t"
1830
+ "gsldlc1 $f14, 0x8($9) \n\t"
1831
+ "gsldrc1 $f6, -0x1($9) \n\t"
1832
+ "gsldrc1 $f14, 0x1($9) \n\t"
1833
+
1834
+ "pasubub $f4, $f4, $f0 \n\t"
1835
+ "pasubub $f6, $f6, $f2 \n\t"
1836
+ "biadd $f4, $f4 \n\t"
1837
+ "biadd $f6, $f6 \n\t"
1838
+ "paddh $f24, $f24, $f4 \n\t"
1839
+ "paddh $f26, $f26, $f6 \n\t"
1840
+ "pasubub $f12, $f12, $f0 \n\t"
1841
+ "pasubub $f14, $f14, $f2 \n\t"
1842
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1843
+ "biadd $f12, $f12 \n\t"
1844
+ "biadd $f14, $f14 \n\t"
1845
+ "paddh $f28, $f28, $f12 \n\t"
1846
+ "paddh $f30, $f30, $f14 \n\t"
1847
+
1848
+ "gsldlc1 $f12, 0x7($9) \n\t"
1849
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1850
+ "gsldrc1 $f12, 0x0($9) \n\t"
1851
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1852
+ "pasubub $f0, $f0, $f12 \n\t"
1853
+ "pasubub $f2, $f2, $f14 \n\t"
1854
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1855
+ "biadd $f0, $f0 \n\t"
1856
+ "biadd $f2, $f2 \n\t"
1857
+ "paddh $f20, $f20, $f0 \n\t"
1858
+ "paddh $f22, $f22, $f2 \n\t"
1859
+
1860
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1861
+ "gsldlc1 $f2, 0x7($8) \n\t"
1862
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1863
+ "gsldrc1 $f2, 0x0($8) \n\t"
1864
+ "pasubub $f12, $f12, $f0 \n\t"
1865
+ "pasubub $f14, $f14, $f2 \n\t"
1866
+ "biadd $f12, $f12 \n\t"
1867
+ "biadd $f14, $f14 \n\t"
1868
+ "paddh $f16, $f16, $f12 \n\t"
1869
+ "paddh $f18, $f18, $f14 \n\t"
1870
+
1871
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1872
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1873
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1874
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1875
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1876
+
1877
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1878
+ "gsldlc1 $f6, 0x6($9) \n\t"
1879
+ "gsldlc1 $f14, 0x8($9) \n\t"
1880
+ "gsldrc1 $f6, -0x1($9) \n\t"
1881
+ "gsldrc1 $f14, 0x1($9) \n\t"
1882
+
1883
+ "pasubub $f4, $f4, $f0 \n\t"
1884
+ "pasubub $f6, $f6, $f2 \n\t"
1885
+ "biadd $f4, $f4 \n\t"
1886
+ "biadd $f6, $f6 \n\t"
1887
+ "paddh $f24, $f24, $f4 \n\t"
1888
+ "paddh $f26, $f26, $f6 \n\t"
1889
+ "pasubub $f12, $f12, $f0 \n\t"
1890
+ "pasubub $f14, $f14, $f2 \n\t"
1891
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1892
+ "biadd $f12, $f12 \n\t"
1893
+ "biadd $f14, $f14 \n\t"
1894
+ "paddh $f28, $f28, $f12 \n\t"
1895
+ "paddh $f30, $f30, $f14 \n\t"
1896
+
1897
+ "gsldlc1 $f12, 0x7($9) \n\t"
1898
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1899
+ "gsldrc1 $f12, 0x0($9) \n\t"
1900
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1901
+ "pasubub $f0, $f0, $f12 \n\t"
1902
+ "pasubub $f2, $f2, $f14 \n\t"
1903
+ "biadd $f0, $f0 \n\t"
1904
+ "biadd $f2, $f2 \n\t"
1905
+ "paddh $f20, $f20, $f0 \n\t"
1906
+ "paddh $f22, $f22, $f2 \n\t"
1907
+
1908
+ "paddh $f16, $f16, $f18 \n\t"
1909
+ "paddh $f20, $f20, $f22 \n\t"
1910
+ "paddh $f24, $f24, $f26 \n\t"
1911
+ "paddh $f28, $f28, $f30 \n\t"
1912
+ "punpcklwd $f16, $f16, $f20 \n\t"
1913
+ "punpcklwd $f24, $f24, $f28 \n\t"
1914
+ "gssqc1 $f24, $f16, 0x0(%[pSad]) \n\t"
1915
+ : [pSample1]"+&r"((unsigned char *)pSample1),
1916
+ [pSample2]"+&r"((unsigned char *)pSample2)
1917
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2),
1918
+ [pSad]"r"((int *)pSad)
1919
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
1920
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
1921
+ );
1922
+ RECOVER_REG;
1923
+ }
1924
+
1925
+ void WelsSampleSadFour8x8_mmi (uint8_t* pSample1, int32_t iStride1, uint8_t* pSample2,
1926
+ int32_t iStride2, int32_t* pSad) {
1927
+ BACKUP_REG;
1928
+ __asm__ volatile (
1929
+ ".set arch=loongson3a \n\t"
1930
+ "xor $f16, $f16, $f16 \n\t"
1931
+ "xor $f18, $f18, $f18 \n\t"
1932
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1933
+ "xor $f20, $f20, $f20 \n\t"
1934
+ "xor $f22, $f22, $f22 \n\t"
1935
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1936
+ "xor $f24, $f24, $f24 \n\t"
1937
+ "xor $f26, $f26, $f26 \n\t"
1938
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1939
+ PTR_SUBU "$9, %[pSample2], %[iStride2] \n\t"
1940
+ "xor $f28, $f28, $f28 \n\t"
1941
+ "xor $f30, $f30, $f30 \n\t"
1942
+ "gsldlc1 $f2, 0x7($8) \n\t"
1943
+ "gsldlc1 $f12, 0x7($9) \n\t"
1944
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1945
+ "gsldrc1 $f2, 0x0($8) \n\t"
1946
+ "gsldrc1 $f12, 0x0($9) \n\t"
1947
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1948
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1949
+ "pasubub $f12, $f12, $f0 \n\t"
1950
+ "pasubub $f14, $f14, $f2 \n\t"
1951
+ "biadd $f12, $f12 \n\t"
1952
+ "biadd $f14, $f14 \n\t"
1953
+ "paddh $f16, $f16, $f12 \n\t"
1954
+ "paddh $f18, $f18, $f14 \n\t"
1955
+
1956
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
1957
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
1958
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
1959
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
1960
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
1961
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
1962
+
1963
+ "gsldlc1 $f6, 0x6($9) \n\t"
1964
+ "gsldlc1 $f14, 0x8($9) \n\t"
1965
+ "gsldrc1 $f6, -0x1($9) \n\t"
1966
+ "gsldrc1 $f14, 0x1($9) \n\t"
1967
+ "pasubub $f4, $f4, $f0 \n\t"
1968
+ "pasubub $f6, $f6, $f2 \n\t"
1969
+ "biadd $f4, $f4 \n\t"
1970
+ "biadd $f6, $f6 \n\t"
1971
+ "paddh $f24, $f24, $f4 \n\t"
1972
+ "paddh $f26, $f26, $f6 \n\t"
1973
+ "pasubub $f12, $f12, $f0 \n\t"
1974
+ "pasubub $f14, $f14, $f2 \n\t"
1975
+ "biadd $f12, $f12 \n\t"
1976
+ "biadd $f14, $f14 \n\t"
1977
+ "paddh $f28, $f28, $f12 \n\t"
1978
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
1979
+ "paddh $f30, $f30, $f14 \n\t"
1980
+
1981
+ "gsldlc1 $f12, 0x7($9) \n\t"
1982
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
1983
+ "gsldrc1 $f12, 0x0($9) \n\t"
1984
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
1985
+ "pasubub $f0, $f0, $f12 \n\t"
1986
+ "pasubub $f2, $f2, $f14 \n\t"
1987
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
1988
+ "biadd $f0, $f0 \n\t"
1989
+ "biadd $f2, $f2 \n\t"
1990
+ "paddh $f20, $f20, $f0 \n\t"
1991
+ "paddh $f22, $f22, $f2 \n\t"
1992
+
1993
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
1994
+ "gsldlc1 $f2, 0x7($8) \n\t"
1995
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
1996
+ "gsldrc1 $f2, 0x0($8) \n\t"
1997
+ "pasubub $f12, $f12, $f0 \n\t"
1998
+ "pasubub $f14, $f14, $f2 \n\t"
1999
+ "biadd $f12, $f12 \n\t"
2000
+ "biadd $f14, $f14 \n\t"
2001
+ "paddh $f16, $f16, $f12 \n\t"
2002
+ "paddh $f18, $f18, $f14 \n\t"
2003
+
2004
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
2005
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
2006
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
2007
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
2008
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
2009
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
2010
+
2011
+ "gsldlc1 $f6, 0x6($9) \n\t"
2012
+ "gsldlc1 $f14, 0x8($9) \n\t"
2013
+ "gsldrc1 $f6, -0x1($9) \n\t"
2014
+ "gsldrc1 $f14, 0x1($9) \n\t"
2015
+
2016
+ "pasubub $f4, $f4, $f0 \n\t"
2017
+ "pasubub $f6, $f6, $f2 \n\t"
2018
+ "biadd $f4, $f4 \n\t"
2019
+ "biadd $f6, $f6 \n\t"
2020
+ "paddh $f24, $f24, $f4 \n\t"
2021
+ "paddh $f26, $f26, $f6 \n\t"
2022
+ "pasubub $f12, $f12, $f0 \n\t"
2023
+ "pasubub $f14, $f14, $f2 \n\t"
2024
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
2025
+ "biadd $f12, $f12 \n\t"
2026
+ "biadd $f14, $f14 \n\t"
2027
+ "paddh $f28, $f28, $f12 \n\t"
2028
+ "paddh $f30, $f30, $f14 \n\t"
2029
+
2030
+ "gsldlc1 $f12, 0x7($9) \n\t"
2031
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
2032
+ "gsldrc1 $f12, 0x0($9) \n\t"
2033
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
2034
+ "pasubub $f0, $f0, $f12 \n\t"
2035
+ "pasubub $f2, $f2, $f14 \n\t"
2036
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
2037
+ "biadd $f0, $f0 \n\t"
2038
+ "biadd $f2, $f2 \n\t"
2039
+ "paddh $f20, $f20, $f0 \n\t"
2040
+ "paddh $f22, $f22, $f2 \n\t"
2041
+
2042
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
2043
+ "gsldlc1 $f2, 0x7($8) \n\t"
2044
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
2045
+ "gsldrc1 $f2, 0x0($8) \n\t"
2046
+ "pasubub $f12, $f12, $f0 \n\t"
2047
+ "pasubub $f14, $f14, $f2 \n\t"
2048
+ "biadd $f12, $f12 \n\t"
2049
+ "biadd $f14, $f14 \n\t"
2050
+ "paddh $f16, $f16, $f12 \n\t"
2051
+ "paddh $f18, $f18, $f14 \n\t"
2052
+
2053
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
2054
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
2055
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
2056
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
2057
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
2058
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
2059
+
2060
+ "gsldlc1 $f6, 0x6($9) \n\t"
2061
+ "gsldlc1 $f14, 0x8($9) \n\t"
2062
+ "gsldrc1 $f6, -0x1($9) \n\t"
2063
+ "gsldrc1 $f14, 0x1($9) \n\t"
2064
+
2065
+ "pasubub $f4, $f4, $f0 \n\t"
2066
+ "pasubub $f6, $f6, $f2 \n\t"
2067
+ "biadd $f4, $f4 \n\t"
2068
+ "biadd $f6, $f6 \n\t"
2069
+ "paddh $f24, $f24, $f4 \n\t"
2070
+ "paddh $f26, $f26, $f6 \n\t"
2071
+ "pasubub $f12, $f12, $f0 \n\t"
2072
+ "pasubub $f14, $f14, $f2 \n\t"
2073
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
2074
+ "biadd $f12, $f12 \n\t"
2075
+ "biadd $f14, $f14 \n\t"
2076
+ "paddh $f28, $f28, $f12 \n\t"
2077
+ "paddh $f30, $f30, $f14 \n\t"
2078
+
2079
+ "gsldlc1 $f12, 0x7($9) \n\t"
2080
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
2081
+ "gsldrc1 $f12, 0x0($9) \n\t"
2082
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
2083
+ "pasubub $f0, $f0, $f12 \n\t"
2084
+ "pasubub $f2, $f2, $f14 \n\t"
2085
+ PTR_ADDU "$8, %[pSample1], %[iStride1] \n\t"
2086
+ "biadd $f0, $f0 \n\t"
2087
+ "biadd $f2, $f2 \n\t"
2088
+ "paddh $f20, $f20, $f0 \n\t"
2089
+ "paddh $f22, $f22, $f2 \n\t"
2090
+
2091
+ "gsldlc1 $f0, 0x7(%[pSample1]) \n\t"
2092
+ "gsldlc1 $f2, 0x7($8) \n\t"
2093
+ "gsldrc1 $f0, 0x0(%[pSample1]) \n\t"
2094
+ "gsldrc1 $f2, 0x0($8) \n\t"
2095
+ "pasubub $f12, $f12, $f0 \n\t"
2096
+ "pasubub $f14, $f14, $f2 \n\t"
2097
+ "biadd $f12, $f12 \n\t"
2098
+ "biadd $f14, $f14 \n\t"
2099
+ "paddh $f16, $f16, $f12 \n\t"
2100
+ "paddh $f18, $f18, $f14 \n\t"
2101
+
2102
+ "gsldlc1 $f4, 0x6(%[pSample2]) \n\t"
2103
+ "gsldlc1 $f12, 0x8(%[pSample2]) \n\t"
2104
+ PTR_ADDU "$9, %[pSample2], %[iStride2] \n\t"
2105
+ PTR_ADDU "%[pSample1], $8, %[iStride1] \n\t"
2106
+ "gsldrc1 $f4, -0x1(%[pSample2]) \n\t"
2107
+ "gsldrc1 $f12, 0x1(%[pSample2]) \n\t"
2108
+
2109
+ "gsldlc1 $f6, 0x6($9) \n\t"
2110
+ "gsldlc1 $f14, 0x8($9) \n\t"
2111
+ "gsldrc1 $f6, -0x1($9) \n\t"
2112
+ "gsldrc1 $f14, 0x1($9) \n\t"
2113
+
2114
+ "pasubub $f4, $f4, $f0 \n\t"
2115
+ "pasubub $f6, $f6, $f2 \n\t"
2116
+ "biadd $f4, $f4 \n\t"
2117
+ "biadd $f6, $f6 \n\t"
2118
+ "paddh $f24, $f24, $f4 \n\t"
2119
+ "paddh $f26, $f26, $f6 \n\t"
2120
+ "pasubub $f12, $f12, $f0 \n\t"
2121
+ "pasubub $f14, $f14, $f2 \n\t"
2122
+ PTR_ADDU "%[pSample2], $9, %[iStride2] \n\t"
2123
+ "biadd $f12, $f12 \n\t"
2124
+ "biadd $f14, $f14 \n\t"
2125
+ "paddh $f28, $f28, $f12 \n\t"
2126
+ "paddh $f30, $f30, $f14 \n\t"
2127
+
2128
+ "gsldlc1 $f12, 0x7($9) \n\t"
2129
+ "gsldlc1 $f14, 0x7(%[pSample2]) \n\t"
2130
+ "gsldrc1 $f12, 0x0($9) \n\t"
2131
+ "gsldrc1 $f14, 0x0(%[pSample2]) \n\t"
2132
+ "pasubub $f0, $f0, $f12 \n\t"
2133
+ "pasubub $f2, $f2, $f14 \n\t"
2134
+ "biadd $f0, $f0 \n\t"
2135
+ "biadd $f2, $f2 \n\t"
2136
+ "paddh $f20, $f20, $f0 \n\t"
2137
+ "paddh $f22, $f22, $f2 \n\t"
2138
+
2139
+ "paddh $f16, $f16, $f18 \n\t"
2140
+ "paddh $f20, $f20, $f22 \n\t"
2141
+ "paddh $f24, $f24, $f26 \n\t"
2142
+ "paddh $f28, $f28, $f30 \n\t"
2143
+ "punpcklwd $f16, $f16, $f20 \n\t"
2144
+ "punpcklwd $f24, $f24, $f28 \n\t"
2145
+ "gssqc1 $f24, $f16, 0x0(%[pSad]) \n\t"
2146
+ : [pSample1]"+&r"((unsigned char *)pSample1),
2147
+ [pSample2]"+&r"((unsigned char *)pSample2)
2148
+ : [iStride1]"r"((int)iStride1), [iStride2]"r"((int)iStride2),
2149
+ [pSad]"r"((int *)pSad)
2150
+ : "memory", "$8", "$9", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
2151
+ "$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28","$f30"
2152
+ );
2153
+ RECOVER_REG;
2154
+ }