Demystifying Cost-Efficiency in LLM Serving over Heterogeneous GPUs
read the original abstract
Recent advancements in Large Language Models (LLMs) have led to increasingly diverse requests, accompanied with varying resource (compute and memory) demands to serve them. However, this in turn degrades the cost-efficiency of LLM serving as common practices primarily rely on homogeneous GPU resources. In response to this problem, this work conducts a thorough study about serving LLMs over heterogeneous GPU resources on cloud platforms. The rationale is that different GPU types exhibit distinct compute and memory characteristics, aligning well with the divergent resource demands of diverse requests. Particularly, through comprehensive benchmarking, we discover that the cost-efficiency of LLM serving can be substantially optimized by meticulously determining GPU composition, deployment configurations, and workload assignments. Subsequently, we design a scheduling algorithm via mixed-integer linear programming, aiming at deducing the most cost-efficient serving plan under the constraints of price budget and real-time GPU availability. Remarkably, our approach effectively outperforms homogeneous and heterogeneous baselines under a wide array of scenarios, covering diverse workload traces, varying GPU availablilities, and multi-model serving. This casts new light on more accessible and efficient LLM serving over heterogeneous cloud resources.
This paper has not been read by Pith yet.
Forward citations
Cited by 10 Pith papers
-
TurboServe: Serving Streaming Video Generation Efficiently and Economically
TurboServe introduces the first serving system for streaming video generation workloads, using migration-aware placement and load-driven autoscaling to cut worst-case latency by 37.5% and GPU cost by 37.2%.
-
HexAGenT: Efficient Agentic LLM Serving via Workflow- and Heterogeneity-Aware Scheduling
HexAGenT reduces the SLO scale required for timely agentic LLM workflow completion by an average of 20.1% at 95% attainment and 33.0% at 99% attainment on heterogeneous A100/H100/H200 clusters.
-
Coral: Cost-Efficient Multi-LLM Serving over Heterogeneous Cloud GPUs
Coral cuts multi-LLM serving costs by up to 2.79x and raises goodput by up to 2.39x on heterogeneous GPUs through adaptive joint optimization and a lossless two-stage decomposition that solves quickly.
-
Autopoiesis: A Self-Evolving System Paradigm for LLM Serving Under Runtime Dynamics
Autopoiesis uses LLM-driven program synthesis to evolve serving policies online during deployment, delivering up to 53% and average 34% gains over prior LLM serving systems under runtime dynamics.
-
GhostServe: A Lightweight Checkpointing System in the Shadow for Fault-Tolerant LLM Serving
GhostServe applies erasure coding to KV cache in host memory for fast recovery from failures in LLM serving, cutting checkpointing latency up to 2.7x and recovery latency 2.1x versus prior methods.
-
ChunkFlow: Communication-Aware Chunked Prefetching for Layerwise Offloading in Distributed Diffusion Transformer Inference
ChunkFlow achieves up to 1.28x step-time speedup and up to 49% lower peak GPU memory for DiT inference by using a first-order model to guide communication-aware chunked prefetching.
-
HexiSeq: Accommodating Long Context Training of LLMs over Heterogeneous Hardware
HexiSeq optimizes sequence and head partitioning across mixed GPUs to improve long-context LLM training throughput by up to 1.72x in simulations.
-
Rethinking Data Curation in LLM Training: Online Reweighting Offers Better Generalization than Offline Methods
ADAPT is an online reweighting framework for LLM training that outperforms offline data selection and mixing methods in cross-benchmark generalization under equal compute.
-
HexiScale: Facilitating Large Language Model Training over Heterogeneous Hardware
HexiScale enables LLM training on heterogeneous GPUs via asymmetric parallelism and graph partitioning, matching homogeneous performance at equal FLOPS and delivering 1.5-2.4x higher throughput than prior heterogeneou...
-
GoodServe: Towards High-Goodput Serving of Agentic LLM Inferences over Heterogeneous Resources
GoodServe proposes a predict-and-rectify routing system for agentic LLM inferences on heterogeneous GPUs that improves goodput by up to 27.4%.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.