Qinghua Zhou c351b871a1 ep: fix internode combine in multirank test ...

Two issues prevented internode HT combine from completing on 2x8 H100:

1. Wrong prefix matrices passed to internode_combine. Combine runs in the
   reverse direction of dispatch, so it must consume the receiver-side
   matrices returned by dispatch (recv_rdma_channel_prefix_matrix,
   recv_rdma_rank_prefix_sum, recv_gbl_channel_prefix_matrix), not the
   sender-side rdma_channel_prefix_matrix / gbl_channel_prefix_matrix.
   This matches DeepEP's deep_ep/buffer.py::internode_combine handle
   unpacking. Without the fix the NVL forwarder's 'NVL check' timed out
   because token_start_idx/token_end_idx were computed against the wrong
   per-channel layout.

2. Cross-rank race between dispatch and combine. Even with the correct
   matrices, launching combine immediately after dispatch deadlocked the
   forwarder NVL check (tail stuck one short of expected_head) because
   peers still had in-flight dispatch proxy traffic while fast ranks had
   already started combine. A torch.cuda.synchronize() + dist.barrier()
   between the two calls makes the test pass deterministically on 16
   ranks (combine diff == 0, max|expected| up to 60.0).

The barrier in the test is a workaround; the real fix belongs in
Buffer::internode_dispatch / Buffer::internode_combine so the
dispatch->combine handoff fully fences outstanding proxy work across
ranks. Marked with an XXX comment in the test.

2026-04-22 02:21:29 +00:00

..

ep

ep: fix internode combine in multirank test

2026-04-22 02:21:29 +00:00