Automated Tensor Scheduling for Hybrid CPU-GPU LLM Inference on Consumer Devices

• URL: http://arxiv.org/abs/2607.10183v2
• Date: Tue, 14 Jul 2026 15:02:22 GMT
• Title: Automated Tensor Scheduling for Hybrid CPU-GPU LLM Inference on Consumer Devices
• Abstract summary: ATSInfer is a hybrid CPU-GPU inference system for consumer devices that performs offloading at tensor granularity.<n>We implement ATSInfer and evaluate it on representative consumer platforms using both dense and MoE models.<n>Compared with existing systems, ATSInfer improves prefill throughput by up to 1.94$times$ and decode throughput by up to 3.29$times$.
• Score: 23.066831820262394
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Running large language models on consumer devices such as laptops and desktops is challenging because model weights often exceed GPU memory capacity, making offloading inference necessary to extend effective model capacity with CPU memory. Existing offloading systems, however, typically rely on coarse layer-level or expert-level scheduling, which overlooks substantial heterogeneity among tensors within the same layer and adapts poorly to changing hardware load conditions on such devices. This paper presents ATSInfer, a hybrid CPU-GPU inference system for consumer devices that performs offloading at tensor granularity. ATSInfer combines static tensor placement with load-aware dynamic transfer, and introduces asynchronous CPU-GPU coordination to efficiently schedule hardware storage, data movement, and computation across heterogeneous backends. We implement ATSInfer and evaluate it on representative consumer platforms using both dense and MoE models. Compared with existing systems, ATSInfer improves prefill throughput by up to 1.94$\times$ and decode throughput by up to 3.29$\times$, while also increasing GPU utilization and making more effective use of PCIe bandwidth. These results show that ATSInfer can substantially improve the user experience of local LLM deployment on personal consumer devices.

Related papers

• LLM Inference in a Flash! [66.43896490006965]
  Large Language Models (LLMs) have shown impressive capabilities across a range of natural language processing tasks.<n>We present an end-to-end integer-only quantization approach to eliminate expensive floating-point computations.
  arXiv  Detail & Related papers  (2026-09-14T18:04:30Z)
• Mixture-of-Parallelisms: Towards Memory-Efficient Training Stack for Mixture-of-Experts Models [67.17268530365189]
  This paper showcases a memory-efficient training stack for Mixture-of-Experts (MoE) models.<n>It combines and specializes various existing and novel parallelism techniques at different layers and stages of the MoE model training pipeline.
  arXiv  Detail & Related papers  (2026-07-02T08:06:57Z)
• UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms [68.40808444348735]
  In simulation-dominated robot control, the essential question is not which processor runs physics, but whether simulation throughput, policy learning, and runtime synchronization form an efficient end-to-end loop.<n>We present UniLab, a heterogeneous CPU-resident / GPU-learning architecture that decouples CPU-parallel simulation from GPU policy updates through a unified runtime for data movement, buffering, and synchronization.
  arXiv  Detail & Related papers  (2026-05-28T17:53:50Z)
• Profiling-Driven Adaptive Distributed Transformer Inference on Embedded Edge Deployment [60.442064966340524]
  We present a prototype study on NVIDIA Jetson Orin Nano devices connected over WiFi.<n>Key finding is that the dominant bottleneck is not just network bandwidth but also the CPU- GPU staging during communication.<n>Experiments show that this strategy reduces latency by 65%-77% and energy consumption by 34%-52% relative to full-tensor exchange.
  arXiv  Detail & Related papers  (2026-05-25T10:39:28Z)
• Eliminating Multi-GPU Performance Taxes: A Systems Approach to Efficient Distributed LLMs [61.953548065938385]
  We introduce the ''Three Taxes'' (Bulk Synchronous, Inter- Kernel Data Locality, and Kernel Launch Overhead) as an analytical framework.<n>We propose moving beyond the rigid BSP model to address key inefficiencies in distributed GPU execution.<n>We observe a 10-20% speedup in end-to-end latency over BSP-based approaches.
  arXiv  Detail & Related papers  (2025-11-04T01:15:44Z)
• Scalable GPU-Based Integrity Verification for Large Machine Learning Models [4.301162531343759]
  We present a security framework that strengthens distributed machine learning by standardizing integrity protections across CPU and GPU platforms.<n>Our approach co-locates integrity verification directly with large ML model execution on GPU accelerators.<n>We provide a hardware-agnostic foundation that enterprise teams can deploy regardless of their underlying CPU and GPU infrastructures.
  arXiv  Detail & Related papers  (2025-10-27T23:45:21Z)
• Characterizing the Efficiency of Distributed Training: A Power, Performance, and Thermal Perspective [6.51239603014107]
  Large Language Models (LLMs) have pushed training workloads beyond the limits of single-node analysis.<n>We present a comprehensive characterization of LLM training across diverse real-world workloads and hardware platforms.
  arXiv  Detail & Related papers  (2025-09-12T16:05:07Z)
• Forecasting LLM Inference Performance via Hardware-Agnostic Analytical Modeling [0.02091806248191979]
  We introduce LIFE, a lightweight and modular analytical framework that is comprised of modular analytical model of operators.<n>LIFE characterizes the influence of software and model optimizations, such as quantization, KV cache compression, LoRA adapters, chunked prefill, different attentions, and operator fusion.<n>We validate LIFE's forecasting with inference on AMD CPUs, NPUs, iGPUs and NVIDIA V100 GPUs, with Llama2-7B variants.
  arXiv  Detail & Related papers  (2025-07-29T03:08:31Z)
• MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism [26.923312725688735]
  Mixture-of-Experts (MoE) showcases tremendous potential to scale large language models (LLMs) with enhanced performance and reduced computational complexity.<n>We present MegaScale-Infer, an efficient and cost-effective system for serving large-scale MoE models.
  arXiv  Detail & Related papers  (2025-04-03T04:20:44Z)
• MoE-Lightning: High-Throughput MoE Inference on Memory-constrained GPUs [55.95879347182669]
  MoE architecture is renowned for its ability to increase model capacity without a proportional increase in inference cost. MoE-Lightning introduces a novel CPU-GPU-I/O pipelining schedule, CGOPipe, with paged weights to achieve high resource utilization. MoE-Lightning can achieve up to 10.3x higher throughput than state-of-the-art offloading-enabled LLM inference systems for Mixtral 8x7B on a single T4 GPU (16GB)
  arXiv  Detail & Related papers  (2024-11-18T01:06:12Z)
• FusionAI: Decentralized Training and Deploying LLMs with Massive Consumer-Level GPUs [57.12856172329322]
  We envision a decentralized system unlocking the potential vast untapped consumer-level GPU. This system faces critical challenges, including limited CPU and GPU memory, low network bandwidth, the variability of peer and device heterogeneity.
  arXiv  Detail & Related papers  (2023-09-03T13:27:56Z)
• FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU [89.2451963569343]
  FlexGen is a generation engine for running large language model (LLM) inference on a single commodity GPU. When running OPT-175B on a single 16GB GPU, FlexGen achieves significantly higher throughput compared to state-of-the-art offloading systems. On the HELM benchmark, FlexGen can benchmark a 30B model with a 16GB GPU on 7 representative sub-scenarios in 21 hours.
  arXiv  Detail & Related papers  (2023-03-13T05:19:28Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.