Litcius/Paper detail

Liger: Interleaving Intra- and Inter-Operator Parallelism for Distributed Large Model Inference

Jiangsu Du, Jinhui Wei, Jiazhi Jiang, Shenggan Cheng, Dan Huang, Zhiguang Chen, Yutong Lu

202410 citationsDOI

Abstract

Distributed large model inference is still in a dilemma where balancing cost and effect. The online scenarios demand intraoperator parallelism to achieve low latency and intensive communications makes it costly. Conversely, the inter-operator parallelism can achieve high throughput with much fewer communications, but it fails to enhance the effectiveness.

Topics & Concepts

Computer scienceInterleavingParallelism (grammar)Latency (audio)InferenceTask parallelismData parallelismParallel computingOperator (biology)ThroughputDilemmaDistributed computingWirelessOperating systemArtificial intelligenceTelecommunicationsRepressorEpistemologyPhilosophyBiochemistryGeneChemistryTranscription factorAdvanced Neural Network ApplicationsDomain Adaptation and Few-Shot LearningCOVID-19 diagnosis using AI