[1/N] Add optional NCCL m2n weight sync backend

[princepride]

Summary

This PR is the first step toward the NCCL m2n rollout weight sync proposal in #52.

It adds an opt-in non-colocated weight sync backend:

--non-colocate-weight-sync-backend {broadcast,nccl-m2n}

The existing gather-broadcast path remains the default. Colocated rollout engines continue using the existing CUDA IPC path.

Changes

• Add UpdateWeightFromNcclm2n as an optional non-colocated updater.
• Add initial NCCL m2n layout analysis for common dense/MoE/replicated tensor patterns.
• Wire backend selection through the actor update-weight setup.
• Keep explicit fallback to the existing broadcast updater for unsupported layouts or unavailable runtime support.
• Add focused unit coverage for layout decisions and fallback/native selection behavior.

Stack

This is [1/N].

Follow-up PRs will cover the native ncclm2nReshardWithWindow runtime path, vLLM worker receive integration, staging/window management, and performance comparison against the legacy gather-broadcast path.

Test Plan

python -m pytest tests/unit/backends/megatron_utils/update_weight/test_nccl_m2n_layout.py tests/unit/backends/megatron_utils/update_weight/test_update_weight_from_nccl_m2n.py

[gemini-code-assist]

Code Review

This pull request introduces an experimental NCCL Xfer weight synchronization backend (UpdateWeightFromNcclXfer) for non-colocated rollout engines, complete with command-line options, validation, layout analysis, and unit tests. The review feedback highlights two important improvements: caching the fallback reason during initialization to prevent a performance bottleneck from traversing model parameters on every weight update, and defensively handling quantization_config to avoid potential AttributeError crashes when it is passed as a configuration object rather than a dictionary.

[princepride]

Superseded by #335, re-opened from the fork branch princepride:feature/nccl-xfer-weight-sync-mvp (same commits). Closing this one.