implementation detail · filed under inference & serving
DeepEP
An MoE all-to-all communication backend used for expert parallelism.
Also called --moe-a2a-backend deepep, DeepEP expert parallelism across nodes, DeepEP for expert parallelism.
- sources
- 4
- models
- 4
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
optional 2used 2
Documented in
Further reading
Picked by hand, not extracted: where to read more, not evidence for anything on this page.
- deepseek-ai/DeepEP code github.comofficial expert-parallel all-to-all communication library
Evidence
4 spans quoted from the sources, strongest treatment first.
its SGLang example spans two nodes with --tp 16 --ep 16 and DeepEP for expert parallelism
usedunclearin MiMo-V2.6-ProXiaomi
--moe-a2a-backend deepep
usedsoftware implementationin MiMo-V2.5Xiaomi
--moe-a2a-backend deepep
optionalsoftware implementationin MiMo-V2.5-ProXiaomi
--moe-a2a-backend deepep
optionalinference servingin MiMo-V2.5Xiaomi
Filed alongside
Other methods under inference & serving :: serving parallelism.
Attention data parallelismPrefill-decode disaggregationEncoder-Prefill-Decode disaggregationTensor parallelism (degree 4)Tensor parallelism (degree 8)Topology-aware NVLink domain placementZero-copy fused token permutation and unpermutationData-parallel vision encodingExpert parallelismFused reduce-scatter/all-gather collectivesIdentical cache-layout pinning across prefill and decode poolsLow-precision MoE combineRound-robin routing for prefill-decode disaggregationTensor parallelism for MoE layersTensorRT-LLM all-reduce backendToken migration for balanced expert placement