Related papers: Predictable LLM Serving on GPU Clusters

Training LLMs in distributed environments presents significant challenges due to the complexity of model execution, deployment systems, and the vast space of configurable strategies. Although various optimization techniques exist, achieving…

Modern deployment of large language models (LLMs) frequently involves both inference serving and continuous retraining to stay aligned with evolving data and user feedback. Common practices separate these workloads onto distinct servers in…

Distributed prefix caching has become a core technique for efficient LLM serving. However, for long-context requests with high cache hit ratios, retrieving reusable KVCache blocks from remote servers has emerged as a new performance…

Generative Recommender (GR) inference places embedding hot caches (EMB) and KV caches in direct competition for limited GPU HBM: allocating more memory to one improves its efficiency but degrades the other. Existing systems optimize them in…

LLM serving is increasingly multi-tenant: the same deployment must handle latency-critical interactive requests and more relaxed background workloads under a fixed GPU budget. This creates a tiered-SLO setting where maximizing overall…

Modern user-facing latency-sensitive web services include numerous distributed, intercommunicating microservices that promise to simplify software development and operation. However, multiplexing of compute resources across microservices is…

The rise of Generative AI introduces a new class of HPC workloads that integrates lightweight LLMs with traditional high-throughput applications to accelerate scientific discovery. The current design of HPC clusters is inadequate to support…

Large-scale machine learning workloads increasingly rely on multi-GPU systems, yet their performance is often limited by an overlooked component: the CPU. Through a detailed study of modern large language model (LLM) inference and serving…

NVIDIA's Multi-Instance GPU (MIG) technology enables partitioning GPU computing power and memory into separate hardware instances, providing complete isolation including compute resources, caches, and memory. However, prior work identifies…

This work examines latency, throughput, and other metrics when performing inference on confidential GPUs. We explore different traffic patterns and scheduling strategies using a single Virtual Machine with one NVIDIA H100 GPU, to perform…

LLM inference is essential for applications like text summarization, translation, and data analysis, but the high cost of GPU instances from Cloud Service Providers (CSPs) like AWS is a major burden. This paper proposes InferSave, a…

Large language model (LLM) applications are blindfolded to the infrastructure underneath and generate tokens autoregressively, indifferent to the system load, thus risking inferencing latency inflation and poor user experience. Our…

Large Language Models have revolutionized natural language processing, yet serving them efficiently in data centers remains challenging due to mixed workloads comprising latency-sensitive (LS) and best-effort (BE) jobs. Existing inference…

This study presents a benchmarking analysis of the Qualcomm Cloud AI 100 Ultra (QAic) accelerator for large language model (LLM) inference, evaluating its energy efficiency (throughput per watt), performance, and hardware scalability…

In the context of Machine Learning as a Service (MLaaS) clouds, the extensive use of Large Language Models (LLMs) often requires efficient management of significant query loads. When providing real-time inference services, several…

In cloud machine learning (ML) inference systems, providing low latency to end-users is of utmost importance. However, maximizing server utilization and system throughput is also crucial for ML service providers as it helps lower the…

This paper presents a SYCL implementation of Multi-Layer Perceptrons (MLPs), which targets and is optimized for the Intel Data Center GPU Max 1550. To increase the performance, our implementation minimizes the slow global memory accesses by…

SMEs increasingly seek alternatives to cloud LLM APIs, which raise data privacy concerns. Dedicated cloud GPU instances offer improved privacy but with limited guarantees and ongoing costs, while professional on-premise hardware (A100,…

LLM inference latency critically determines user experience and operational costs, directly impacting throughput under SLO constraints. Even brief latency spikes degrade service quality despite acceptable average performance. However,…

This paper tackles the challenge of running multiple ML inference jobs (models) under time-varying workloads, on a constrained on-premises production cluster. Our system Faro takes in latency Service Level Objectives (SLOs) for each job,…