PipeInfer: Accelerating LLM Inference using Asynchronous Pipelined Speculation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Butler, Branden, Yu, Sixing, Mazaheri, Arya, Jannesari, Ali
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910702020591616
author Butler, Branden
Yu, Sixing
Mazaheri, Arya
Jannesari, Ali
author_facet Butler, Branden
Yu, Sixing
Mazaheri, Arya
Jannesari, Ali
contents Inference of Large Language Models (LLMs) across computer clusters has become a focal point of research in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce bottlenecks associated with memory bandwidth, but also increase end-to-end latency per inference run, requiring high speculation acceptance rates to improve performance. Combined with a variable rate of acceptance across tasks, speculative inference techniques can result in reduced performance. Additionally, pipeline-parallel designs require many user requests to maintain maximum utilization. As a remedy, we propose PipeInfer, a pipelined speculative acceleration technique to reduce inter-token latency and improve system utilization for single-request scenarios while also improving tolerance to low speculation acceptance rates and low-bandwidth interconnects. PipeInfer exhibits up to a 2.15$\times$ improvement in generation speed over standard speculative inference. PipeInfer achieves its improvement through Continuous Asynchronous Speculation and Early Inference Cancellation, the former improving latency and generation speed by running single-token inference simultaneously with several speculative runs, while the latter improves speed and latency by skipping the computation of invalidated runs, even in the middle of inference.
format Preprint
id arxiv_https___arxiv_org_abs_2407_11798
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle PipeInfer: Accelerating LLM Inference using Asynchronous Pipelined Speculation
Butler, Branden
Yu, Sixing
Mazaheri, Arya
Jannesari, Ali
Computation and Language
Distributed, Parallel, and Cluster Computing
Machine Learning
Inference of Large Language Models (LLMs) across computer clusters has become a focal point of research in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce bottlenecks associated with memory bandwidth, but also increase end-to-end latency per inference run, requiring high speculation acceptance rates to improve performance. Combined with a variable rate of acceptance across tasks, speculative inference techniques can result in reduced performance. Additionally, pipeline-parallel designs require many user requests to maintain maximum utilization. As a remedy, we propose PipeInfer, a pipelined speculative acceleration technique to reduce inter-token latency and improve system utilization for single-request scenarios while also improving tolerance to low speculation acceptance rates and low-bandwidth interconnects. PipeInfer exhibits up to a 2.15$\times$ improvement in generation speed over standard speculative inference. PipeInfer achieves its improvement through Continuous Asynchronous Speculation and Early Inference Cancellation, the former improving latency and generation speed by running single-token inference simultaneously with several speculative runs, while the latter improves speed and latency by skipping the computation of invalidated runs, even in the middle of inference.
title PipeInfer: Accelerating LLM Inference using Asynchronous Pipelined Speculation
topic Computation and Language
Distributed, Parallel, and Cluster Computing
Machine Learning
url https://arxiv.org/abs/2407.11798