Qinghua Zhou 4ed6f229f2 tests/ep: align intranode HT bench with NCCL-EP accounting ...

- Add MSCCLPP_EP_BENCH_EXPERTS / _TOPK env knobs so the bench phase can
  match NCCL-EP's `ep_bench -a ht` defaults (256 experts, top-8). The
  functional check above continues to use the smaller (num_ranks*4
  experts, topk=4) configuration.

- Switch BW accounting from recv_tokens*hidden to bench_tokens*hidden,
  matching NCCL-EP's `RDMA_send` per-rank byte count. The previous
  formula counted DeepEP's expanded recv layout (one row per
  (token,src_rank) pair), inflating reported GB/s ~5x and making
  cross-stack comparisons misleading.

2026-04-27 17:14:42 +00:00

..

ext/ep

tests/ep: align intranode HT bench with NCCL-EP accounting

2026-04-27 17:14:42 +00:00