SC2021Top-tier venue
Enable simultaneous DNN services based on deterministic operator overlap and precise latency prediction
Weihao Cui, Han Zhao, Quan Chen, Ningxin Zheng, Jingwen Leng, Jieru Zhao, Zhuo Song, Tao Ma, Yong Yang, Chao Li, Minyi Guo
Abstract
While user-facing services experience diurnal load patterns, co-locating services improve hardware utilization. Prior work on co-locating services on GPUs run queries sequentially, as the latencies of the queries are neither stable nor predictable when running simultaneously. The input sensitiveness and the non-deterministic operator overlap are two primary factors of the latency unpredictability. Hence, We propose Abacus, a runtime system that runs multiple services simultaneously. Abacus enables deterministic operator overlap to enforce latency predictability. Abacus composes of an overlap-aware latency predictor, a headroom-based query controller, and segmental model executors. The predictor predicts the latencies of the deterministic operator overlap. The controller determines the appropriate operator overlap for the QoS guarantee of all the services. The executors run the operators as needed to support the deterministic operator overlap. Our evaluation shows that Abacus reduces 51.3% of the QoS violation and improves the throughput by 29.8% on average compared with state-of-the-art solutions.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext d2ceb769-c5fa-494a-a15e-bed070149f3bCited by top-tier papers15
- Microsecond-scale Preemption for Concurrent GPU-accelerated DNN InferencesMingcong Han, Hanze Zhang, Rong Chen, Haibo ChenOSDI 2022 · 153 citations
- Clover: Toward Sustainable AI with Carbon-Aware Machine Learning Inference ServiceBaolin Li, Siddharth Samsi, Vijay Gadepally, Devesh TiwariSC 2023 · 63 citations
- VELTAIR: towards high-performance multi-tenant deep learning services via adaptive compilation and schedulingZihan Liu, Jingwen Leng, Zhihui Zhang, Quan Chen et al.ASPLOS 2022 · 52 citations
- Improving GPU Sharing Performance through Adaptive Bubbleless Spatial-Temporal SharingShulai Zhang, Quan Chen, Weihao Cui, Han Zhao et al.EuroSys 2025 · 19 citations
- Efficient Performance-Aware GPU Sharing with Compatibility and Isolation through Kernel Space InterceptionShulai Zhang, Ao Xu, Quan Chen, Han Zhao et al.USENIX ATC 2025 · 16 citations
Builds on11
- Serverless in the Wild: Characterizing and Optimizing the Serverless Workload at a Large Cloud ProviderMohammad Shahrad, Rodrigo Fonseca, Iñigo Goiri, Gohar Irfan Chaudhry et al.USENIX ATC 2020 · 946 citations
- MLPerf Inference BenchmarkVijay Janapa Reddi, Christine Cheng, David Kanter, Peter Mattson et al.ISCA 2020 · 517 citations
- Serving DNNs like Clockwork: Performance Predictability from the Bottom UpArpan Gujarati, Reza Karimi, Safya Alzayat, Wei Hao et al.OSDI 2020 · 392 citations
- Sinan: ML-based and QoS-aware resource management for cloud microservicesYanqi Zhang, Weizhe Hua, Zhuangzhuang Zhou, G. Edward Suh et al.ASPLOS 2021 · 226 citations
- Rammer: Enabling Holistic Deep Learning Compiler Optimizations with rTasksLingxiao Ma, Zhiqiang Xie, Zhi Yang, Jilong Xue et al.OSDI 2020 · 192 citations
Related papers
- Tacker: Tensor-CUDA Core Kernel Fusion for Improving the GPU Utilization while Ensuring QoSHan Zhao, Weihao Cui, Quan Chen, Youtao Zhang et al.HPCA 2022 · 42 citations
- Orion: Interference-aware, Fine-grained GPU Sharing for ML ApplicationsFoteini Strati, Xianzhe Ma, Ana KlimovicEuroSys 2024 · 96 citations
- Astraea: towards QoS-aware and resource-efficient multi-stage GPU servicesWei Zhang, Quan Chen, Kaihua Fu, Ningxin Zheng et al.ASPLOS 2022 · 28 citations
- HSM: A Hybrid Slowdown Model for Multitasking GPUsXia Zhao, Magnus Jahre, Lieven EeckhoutASPLOS 2020 · 36 citations
- Serving Heterogeneous Machine Learning Models on Multi-GPU Servers with Spatio-Temporal SharingSeungbeom Choi, Sunho Lee, Yeonjae Kim, Jongse Park et al.USENIX ATC 2022 · 200 citations
