avcodec/x86: remove 3 sub in pred4x4_tm_vp8_8.

before:
  411 decicycles in ff_pred4x4_tm_vp8_8_ssse3, 8388289 runs, 319 skips

after:
  389 decicycles in ff_pred4x4_tm_vp8_8_ssse3, 8388308 runs, 300 skips

Tested on i7 920.

libavcodec/x86/h264_intrapred.asm

@@ -2486,10 +2486,7 @@ cglobal pred4x4_tm_vp8_8, 3,3
     pshufb     mm3, mm6
     pshufb     mm4, mm6
     pshufb     mm5, mm6
-    psubw      mm2, mm7
-    psubw      mm3, mm7
-    psubw      mm4, mm7
-    psubw      mm5, mm7
+    psubw      mm0, mm7
     paddw      mm2, mm0
     paddw      mm3, mm0
     paddw      mm4, mm0