Rank chroma intra modes without the CCTX search at speed >= 4

[YixinDu1]

At speed >= 4, rank the chroma intra modes without the cross-chroma-component transform search and run that search only for the winning mode, which reduces the chroma intra search time.

Anchor: commit 64c66e1
CTC: AV2 CTC v8 (33 frames, classes A1-B2, RA)

Speed 4 (cpu-used=4)
  +--------------+-------+-------+-------+-------+-------+--------+
  | Class        |     Y |    Cb |    Cr |  wAvg |  Enc% |  Ratio |
  +--------------+-------+-------+-------+-------+-------+--------+
  | A1           | -0.03 |  0.94 |  0.71 |  0.04 |  98.6 |   37.4 |
  | A2           | -0.06 |  0.61 |  0.94 |  0.01 |  97.8 |  211.5 |
  | Avg w/o B2   | -0.07 |  1.02 |  1.02 |  0.01 |  98.2 |  122.4 |
  +--------------+-------+-------+-------+-------+-------+--------+

Speed 5 (cpu-used=5)
  +--------------+-------+-------+-------+-------+-------+--------+
  | Class        |     Y |    Cb |    Cr |  wAvg |  Enc% |  Ratio |
  +--------------+-------+-------+-------+-------+-------+--------+
  | A1           | -0.00 |  0.97 |  0.84 |  0.07 |  97.8 |   30.6 |
  | A2           |  0.02 |  0.89 |  0.42 |  0.07 |  97.5 |   35.2 |
  | Avg w/o B2   | -0.00 |  0.83 |  0.87 |  0.07 |  98.0 |   30.4 |
  +--------------+-------+-------+-------+-------+-------+--------+

Speed 6 (cpu-used=6)
  +--------------+-------+-------+-------+-------+-------+--------+
  | Class        |     Y |    Cb |    Cr |  wAvg |  Enc% |  Ratio |
  +--------------+-------+-------+-------+-------+-------+--------+
  | A1           | -0.12 |  1.01 |  1.09 | -0.03 |  99.4 |  -20.2 |
  | A2           | -0.10 |  0.52 |  0.40 | -0.05 |  97.3 |  -51.6 |
  | Avg w/o B2   | -0.08 |  0.87 |  0.54 | -0.02 |  98.0 | -128.2 |
  +--------------+-------+-------+-------+-------+-------+--------+

STATS_CHANGED

[yunqingwang1]

Looks good.