Files
xc-llm-ascend/vllm_ascend
Chen Chen aa02a85e4d [bugfix] Fix dummy-run and multi-node issues in MoE routing and MTP (#4947)
### What this PR does / why we need it?

- Fix a premature `return` in `moe_init_routing_quant_v2.cpp` so the
routing kernel completes correctly instead of exiting early in certain
paths.
- Switch `FusedAlltoAllCommImpl` to use the MC2-based token dispatcher
and prepare/finalize routines, aligning MoE communication with the MC2
algorithm optimized for Ascend devices.
- Add a temporary override in `MtpProposer` to map `FUSED_ALLTOALL` back
to `ALLTOALL` until the MoE communication type selection logic is fully
finalized, avoiding incorrect behavior in dummy-run flows.
- Simplify the MoE communication selection for Ascend 910-93 in
`NPUModelRunner` by removing the EP-size guard on `FUSED_ALLTOALL`,
which fixes failures in multi-node / larger-EP configurations while
keeping MC2 routing under the configured token capacity.

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: mojave2 <chenchen145@huawei.com>
2025-12-15 14:18:23 +08:00
..
2025-12-02 22:10:52 +08:00
2025-12-11 18:45:43 +08:00
2025-12-14 11:09:56 +08:00
2025-12-02 17:35:47 +08:00