Related papers: DuetServe: Harmonizing Prefill and Decode for LLM …

Efficient execution of deep learning workloads on dataflow architectures is crucial for overcoming memory bottlenecks and maximizing performance. While streaming intermediate results between computation kernels can significantly improve…

In multi-model LLM serving, decode execution remains inefficient due to model-specific resource partitioning: since cross-model batching is not possible, memory-bound decoding often suffers from severe GPU underutilization, especially under…

Large Language Models (LLMs) demonstrate substantial potential across a diverse array of domains via request serving. However, as trends continue to push for expanding context sizes, the autoregressive nature of LLMs results in highly…

Diffusion pipelines, renowned for their powerful visual generation capabilities, have seen widespread adoption in generative vision tasks (e.g., text-to-image/video). These pipelines typically follow an encode--diffuse--decode three-stage…

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…

Transformer-based large language models (LLMs) exhibit impressive performance in generative tasks but also introduce significant challenges in real-world serving due to inefficient use of the expensive, computation-optimized accelerators.…

Autoregressive inference in large transformer-based language models (LLMs) presents significant challenges for runtime efficiency, particularly during the decode phase where load imbalance across GPU shards can cause throughput degradation…

Lossless model compression holds tremendous promise for alleviating the memory and bandwidth bottlenecks in bit-exact Large Language Model (LLM) serving. However, existing approaches often result in substantial inference slowdowns due to…

Large language model (LLM) serving is fundamentally limited by inefficient hardware utilization. Autoregressive (AR) decoding underutilizes GPUs due to its strictly sequential execution, while diffusion LLMs (DLLMs) improve throughput by…

Speculative decoding has emerged as a promising technique for large language model (LLM) inference by accelerating autoregressive decoding via draft-then-verify. This paper studies a new edge scenario with multi-user inference, where draft…

Large Language Models (LLMs) are increasingly deployed in both latency-sensitive online services and cost-sensitive offline workloads. Co-locating these workloads on shared serving instances can improve resource utilization, but directly…

This paper introduces Helix, a distributed system for high-throughput, low-latency large language model (LLM) serving in heterogeneous GPU clusters. The key idea behind Helix is to formulate inference computation of LLMs over heterogeneous…

As the model size continuously increases, pipeline parallelism shows great promise in throughput-oriented LLM inference due to its low demand on communications. However, imbalanced pipeline workloads and complex data dependencies in the…

Real-time LLM interactions demand streamed token generations, where text tokens are progressively generated and delivered to users while balancing two objectives: responsiveness (i.e., low time-to-first-token) and steady generation…

LLM serving is increasingly dominated by decode attention, which is a memory-bound operation due to massive KV cache loading from global memory. Meanwhile, real-world workloads exhibit substantial, hierarchical shared prefixes across…

Large language models (LLMs) are increasingly deployed under the Model-as-a-Service (MaaS) paradigm. To meet stringent quality-of-service (QoS) requirements, existing LLM serving systems disaggregate the prefill and decode phases of…

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…

Serving LLMs with a cluster of GPUs is common nowadays, where the serving system must meet strict latency SLOs required by applications. However, the stateful nature of LLM serving requires maintaining huge states (i.e., KVCache) in limited…

Modern LLM serving is no longer homogeneous or monolithic. Production systems now combine disaggregated execution, complex parallelism, runtime optimizations, and stateful workloads such as reasoning, agents, and RL rollouts. Simulation is…

Large Language Models (LLMs) have become the new foundation for many applications, reshaping human society like a storm. Disaggregated inference, which separates prefill and decode stages, is a promising approach to improving hardware…