Feature: add EP kernels with inter-node deduplication

[TianDi101]

No description provided.

[TianDi101]

This PR introduce MORI-EP V1 dispatch kernel, we've seen some early results showing significant performance improvement. Below is tested with DeepSeekV3 config:

FP8 Dispatch: RDMA best 39GB/s, avg 35.8GB/s ; XGMI best 127GB/s, avg 118GB/s
BF16 Dispatch: RDMA best 43.5GB/s, avg 40.5GB/s; XGMI best 140GB/s, avg 133GB/s

RDMA bandwidth is calculated in the same way DeepEP does

FP8 Dispatch

BF16 Dispatch

[TianDi101]

Perf now boost further with channel design.

FP8 Dispatch: RDMA best 56 GB/s, avg 49 GB/s ; XGMI best 183 GB/s, avg 164 GB/s
BF16 Dispatch: RDMA best 65 GB/s, avg 56 GB/s; XGMI best 212 GB/s, avg 194GB/s

FP8 Dispatch

BF16 Dispatch

[TianDi101]

With a demo ShmemPutWithImm APIs, perf now comes
fp8: dispatch: best/avg RDMA bandwidth 59.94 / 55.47 XGMI bandwidth 196.50 / 182.00 GB/s | best/avg latency 973.12 / 1068.59 µs
bf16: dispatch: best/avg RDMA bandwidth 68.50 / 65.64 XGMI bandwidth 224.44 / 215.16 GB/s | best/avg latency 1703.00 / 1784.19 µs