ClusterFusion fuses QKV projection, attention, and output projection into one kernel using Hopper cluster collectives, cutting average decoding latency by about 1.6x on an H100.
Deepspeed- inference: Enabling efficient inference of transformer models at unprecedented scale
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.DC 1years
2025 1verdicts
CONDITIONAL 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
ClusterFusion: Expanding Operator Fusion Scope for LLM Inference via Cluster-Level Collective Primitive
ClusterFusion fuses QKV projection, attention, and output projection into one kernel using Hopper cluster collectives, cutting average decoding latency by about 1.6x on an H100.