HarMoEny uses dynamic token redistribution and asynchronous expert prefetching to achieve near-perfect GPU load balance in multi-GPU MoE inference.
A neural probabilistic language model
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.DC 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
HarMoEny: Efficient Multi-GPU Inference of MoE Models
HarMoEny uses dynamic token redistribution and asynchronous expert prefetching to achieve near-perfect GPU load balance in multi-GPU MoE inference.