perf(metax): reduce RMSNorm blocks for small head dimensions

[JoeZhang-0x000]

MetaX's device-wide RMSNorm block size leaves idle reduction threads for small Q/K head dimensions. Select 128 threads when the normalized dimension is <=128 on MetaX. Preserve the current block-size selection for larger rows and all other CUDA backends; the computation kernel is unchanged.

Add packed-QKV-stride cases for Q/K head dimension 128 and a dimension-129 case for the unchanged branch.

Validation on MetaX C550, MACA 3.8.0.23, PyTorch 2.10.0+metax3.8.0.7:

• Built master 4c014ca with this change and the two separately proposed MetaX Embedding/FusedAddRmsNorm changes; 301 focused operator tests passed, plus the smoke selection (2 tests).
• All three patches also apply without fuzz to the plugin's locked InfiniOps 8c2f70a.
• Prior BF16 Graph measurements of the same launch change on 8c2f70a: [64,16,128], 8.81 -> 6.06 us; [8192,16,128], 741.41 -> 369.40 us. These are per-call device times, not model throughput.

NVIDIA hardware was not available for testing; the MetaX constexpr guard preserves its current selection.