Communication Compression for Tensor Parallel LLM Inference

Large Language Models (LLMs) have pushed the frontier of artificial intelligence but are comprised of hundreds of billions of parameters and operations. For faster inference latency, LLMs are deployed on multiple hardware accelerators through various Model Parallelism strategies. Our paper looks int...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2024-11
Hauptverfasser: Hansen-Palmus, Jan, Michael Truong Le, Hausdörfer, Oliver, Verma, Alok
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Large Language Models (LLMs) have pushed the frontier of artificial intelligence but are comprised of hundreds of billions of parameters and operations. For faster inference latency, LLMs are deployed on multiple hardware accelerators through various Model Parallelism strategies. Our paper looks into the details on one such strategy - Tensor Parallel - and proposes to reduce latency by compressing inter-accelerator communication. We leverage fine grained quantization techniques to compress selected activations by 3.5 - 4.5x. Our proposed method leads up to 2x reduction of time-to-first-token (TTFT) with negligible model performance degradation.
ISSN:2331-8422