Vidur: A Large-Scale Simulation Framework For LLM Inference

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Agrawal, Amey, Kedia, Nitin, Mohan, Jayashree, Panwar, Ashish, Kwatra, Nipun, Gulavani, Bhargav, Ramjee, Ramachandran, Tumanov, Alexey
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917671822426112
author Agrawal, Amey
Kedia, Nitin
Mohan, Jayashree
Panwar, Ashish
Kwatra, Nipun
Gulavani, Bhargav
Ramjee, Ramachandran
Tumanov, Alexey
author_facet Agrawal, Amey
Kedia, Nitin
Mohan, Jayashree
Panwar, Ashish
Kwatra, Nipun
Gulavani, Bhargav
Ramjee, Ramachandran
Tumanov, Alexey
contents Optimizing the deployment of Large language models (LLMs) is expensive today since it requires experimentally running an application workload against an LLM implementation while exploring large configuration space formed by system knobs such as parallelization strategies, batching techniques, and scheduling policies. To address this challenge, we present Vidur - a large-scale, high-fidelity, easily-extensible simulation framework for LLM inference performance. Vidur models the performance of LLM operators using a combination of experimental profiling and predictive modeling, and evaluates the end-to-end inference performance for different workloads by estimating several metrics of interest such as latency and throughput. We validate the fidelity of Vidur on several LLMs and show that it estimates inference latency with less than 9% error across the range. Further, we present Vidur-Search, a configuration search tool that helps optimize LLM deployment. Vidur-Search uses Vidur to automatically identify the most cost-effective deployment configuration that meets application performance constraints. For example, Vidur-Search finds the best deployment configuration for LLaMA2-70B in one hour on a CPU machine, in contrast to a deployment-based exploration which would require 42K GPU hours - costing ~218K dollars. Source code for Vidur is available at https://github.com/microsoft/vidur.
format Preprint
id arxiv_https___arxiv_org_abs_2405_05465
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Vidur: A Large-Scale Simulation Framework For LLM Inference
Agrawal, Amey
Kedia, Nitin
Mohan, Jayashree
Panwar, Ashish
Kwatra, Nipun
Gulavani, Bhargav
Ramjee, Ramachandran
Tumanov, Alexey
Machine Learning
Artificial Intelligence
Computation and Language
Optimizing the deployment of Large language models (LLMs) is expensive today since it requires experimentally running an application workload against an LLM implementation while exploring large configuration space formed by system knobs such as parallelization strategies, batching techniques, and scheduling policies. To address this challenge, we present Vidur - a large-scale, high-fidelity, easily-extensible simulation framework for LLM inference performance. Vidur models the performance of LLM operators using a combination of experimental profiling and predictive modeling, and evaluates the end-to-end inference performance for different workloads by estimating several metrics of interest such as latency and throughput. We validate the fidelity of Vidur on several LLMs and show that it estimates inference latency with less than 9% error across the range. Further, we present Vidur-Search, a configuration search tool that helps optimize LLM deployment. Vidur-Search uses Vidur to automatically identify the most cost-effective deployment configuration that meets application performance constraints. For example, Vidur-Search finds the best deployment configuration for LLaMA2-70B in one hour on a CPU machine, in contrast to a deployment-based exploration which would require 42K GPU hours - costing ~218K dollars. Source code for Vidur is available at https://github.com/microsoft/vidur.
title Vidur: A Large-Scale Simulation Framework For LLM Inference
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2405.05465