Unverified Commit 59bd5f6a authored by Wentao Ye's avatar Wentao Ye Committed by GitHub
Browse files

[Feat] Enable eplb with default all2all backend (#30559)


Signed-off-by: default avataryewentao256 <zhyanwentao@126.com>
parent 00a8d762
......@@ -29,14 +29,14 @@ class SharedFusedMoE(FusedMoE):
self._shared_experts = shared_experts
# Disable shared expert overlap if:
# - we are using eplb, because of correctness issues
# - we are using flashinfer with DP, since there nothing to gain
# - we are using eplb with non-default backend, because of correctness issues
# - we are using flashinfer with DP, since there nothint to gain
# - we are using marlin kernels
backend = self.moe_parallel_config.all2all_backend
self.use_overlapped = (
use_overlapped
and not (
# TODO(wentao): find the root cause and remove this condition
self.enable_eplb
(self.enable_eplb and backend != "allgather_reducescatter")
or (self.moe_config.use_flashinfer_cutlass_kernels and self.dp_size > 1)
)
and self._shared_experts is not None
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment