# Auto-Scaling Heterogeneous Neural Processing Units for Energy and Cost-Efficient LLM Serving

arXiv: https://arxiv.org/abs/2607.16488 · HTML: https://mlsanity.com/paper/2607.16488/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Serving systems and request scheduling, one of 12 directions in Efficient LLM inference and serving.
- Builds on 31 of its 88 references that are in the map.
- First posted on arXiv 17 Jul 2026.

## Placed in

- Efficient LLM inference and serving / [Serving systems and request scheduling](https://mlsanity.com/inference/serving-systems-and-request-scheduling/)

## Builds on (most cited first, 30 of 31)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding](https://mlsanity.com/paper/2006.16668/) (2020)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [OpenThoughts: Data Recipes for Reasoning Models](https://mlsanity.com/paper/2506.04178/) (2025)
- [Reducing Activation Recomputation in Large Transformer Models](https://mlsanity.com/paper/2205.05198/) (2022)
- [TPU v4: An Optically Reconfigurable Supercomputer for Machine Learning with Hardware Support for Embeddings](https://mlsanity.com/paper/2304.01433/) (2023)
- [Llumnix: Dynamic Scheduling for Large Language Model Serving](https://mlsanity.com/paper/2406.03243/) (2024)
- [Sequence Parallelism: Long Sequence Training from System Perspective](https://mlsanity.com/paper/2105.13120/) (2021)
- [PipeDream: Fast and Efficient Pipeline Parallel DNN Training](https://mlsanity.com/paper/1806.03377/) (2018)
- [DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency](https://mlsanity.com/paper/2408.00741/) (2024)
- [Large Multimodal Agents: A Survey](https://mlsanity.com/paper/2402.15116/) (2024)
- [SpotServe: Serving Generative Large Language Models on Preemptible Instances](https://mlsanity.com/paper/2311.15566/) (2023)
- [LV-Eval: A Balanced Long-Context Benchmark with 5 Length Levels Up to 256K](https://mlsanity.com/paper/2402.05136/) (2024)
- [ServerlessLLM: Low-Latency Serverless Inference for Large Language Models](https://mlsanity.com/paper/2401.14351/) (2024)
- [Infinite-LLM: Efficient LLM Service for Long Context with DistAttention and Distributed KVCache](https://mlsanity.com/paper/2401.02669/) (2024)
- [Efficient Interactive LLM Serving with Proxy Model-based Sequence Length Prediction](https://mlsanity.com/paper/2404.08509/) (2024)
- [Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads](https://mlsanity.com/paper/2008.09213/) (2020)
- [Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow](https://mlsanity.com/paper/2406.01566/) (2024)
- [FIRM: An Intelligent Fine-Grained Resource Management Framework for SLO-Oriented Microservices](https://mlsanity.com/paper/2008.08509/) (2020)
- [M'elange: Cost Efficient Large Language Model Serving by Exploiting GPU Heterogeneity](https://mlsanity.com/paper/2404.14527/) (2024)
- [BLITZSCALE: Fast and Live Large Model Autoscaling with O(1) Host Caching](https://mlsanity.com/paper/2412.17246/) (2024)
- [Capacity-Aware Inference: Mitigating the Straggler Effect in Mixture of Experts](https://mlsanity.com/paper/2503.05066/) (2025)
- [Sailor: Automating Distributed Training over Dynamic, Heterogeneous, and Geo-distributed Clusters](https://mlsanity.com/paper/2504.17096/) (2025)
- [Life-Cycle Emissions of AI Hardware: A Cradle-To-Grave Approach and Generational Trends](https://mlsanity.com/paper/2502.01671/) (2025)
- [Hetis: Serving LLMs in Heterogeneous GPU Clusters with Fine-grained and Dynamic Parallelism](https://mlsanity.com/paper/2509.08309/) (2025)
- [Is the GPU Half-Empty or Half-Full? Practical Scheduling Techniques for LLMs](https://mlsanity.com/paper/2410.17840/) (2024)

## Abstract (arXiv)

To meet the ever-increasing computing demands of large language model (LLM) services, modern cloud platforms have widely deployed neural processing units (NPUs). These NPU chips have been developed and evolved at an incredibly fast pace, this inevitably produces heterogeneous compute pools backed by different versions of NPU chips. Unfortunately, due to the lack of system and architecture support for managing NPU heterogeneity in the cloud, it is unclear how to best utilize heterogeneous NPUs to maximize the energy and cost efficiency for LLM services. In this paper, we first conduct a characterization study of various generations of real NPU chips to demonstrate the potential benefits on energy/cost efficiency and performance by utilizing heterogeneous NPU chips. To realize these benefits, we present NeuScale, an auto-scaling framework to automatically exploit heterogeneous NPUs for cloud platforms. NeuScale manages heterogeneous NPU resources with a new vPod abstraction, which abstracts the core hardware parameters of different NPU versions and provides compatibility with existing ML frameworks. It makes the best-fit vPod allocations for different LLM inference requests using an intuitive and lightweight roofline-based analysis. It supports fine-grained dynamic NPU resource provisioning by adjusting both the vPod configuration (i.e., scaling up/down) and the number of vPods (e.g., scaling in/out). To validate the benefits of NeuScale at scale, we implement it with a production-level NPU simulator. Our evaluation with popular LLMs shows that NeuScale can significantly improve cost efficiency and service-level objective (SLO) satisfaction rate by best utilizing heterogeneous NPU resources.