Georgios Michas
ABSTRACT
Stream Processing Engines (SPEs) have recently begun utilizing heterogeneous coprocessors (e.g., GPUs) to meet the velocity requirements of modern real-time applications. The massive parallelism and high memory bandwidth of GPUs can significantly increase processing throughput in data-intensive streaming scenarios, such as windowed aggregations. However, previous research only focused on the overall architecture of hybrid CPU-GPU streaming systems and the need for efficient in-GPU window operators was overshadowed by the limited interconnect bandwidth.
With aggregation taking up a significant portion of streaming workloads, in this work, we analyze and optimize the performance of sliding window aggregates over GPUs. Current implementations under-utilize the hardware, and for a range of query parameters they cannot even saturate the bandwidth of the interconnect. To optimize execution, we first evaluate the fundamental building blocks of streaming aggregation for GPUs and identify the performance bottlenecks. Then, we build Slider: an adaptive algorithm that selects the most appropriate primitives and kernel configurations based on the query parameters. Our evaluation shows that Slider outperforms previous approaches by 3×-1250×, and saturates both the interconnect and the memory bandwidth for a wide range of examined input workloads.
- Paris Carbone, Asterios Katsifodimos, Stephan Ewen, Volker Markl, Seif Haridi, and Kostas Tzoumas. 2015. Apache flink: Stream and batch processing in a single engine. Bulletin of the IEEE Computer Society Technical Committee on Data Engineering 36, 4 (2015).Google Scholar
- Periklis Chrysogelos, Manos Karpathiotakis, Raja Appuswamy, and Anastasia Ailamaki. 2019. HetExchange: Encapsulating heterogeneous CPU-GPU parallelism in JIT compiled engines. Proc. VLDB Endow. 12, 5 (2019), 544--556. https://doi.org/10.14778/3303753.3303760Google ScholarDigital Library
- Jim Gray, Surajit Chaudhuri, Adam Bosworth, Andrew Layman, Don Reichart, Murali Venkatrao, Frank Pellow, and Hamid Pirahesh. 1997. Data cube: A relational aggregation operator generalizing group-by, cross-tab, and sub-totals. Data mining and knowledge discovery 1, 1 (1997), 29--53.Google Scholar
- Alexandros Koliousis, Matthias Weidlich, Raul Castro Fernandez, Alexander L Wolf, Paolo Costa, and Peter Pietzuch. 2016. Saber: Window-based hybrid stream processing for heterogeneous architectures. In Proceedings of the 2016 International Conference on Management of Data. 555--569.Google ScholarDigital Library
- Jin Li, David Maier, Kristin Tufte, Vassilis Papadimos, and Peter A Tucker. 2005. No pane, no gain: efficient evaluation of sliding-window aggregates over data streams. Acm Sigmod Record 34, 1 (2005), 39--44.Google ScholarDigital Library
- Duane Merrill and Michael Garland. 2016. Single-pass parallel prefix scan with decoupled look-back. NVIDIA, Tech. Rep. NVR-2016-002 (2016).Google Scholar
- Kanat Tangwongsan, Martin Hirzel, Scott Schneider, and Kun-Lung Wu. 2015. General incremental sliding-window aggregation. Proceedings of the VLDB Endowment 8, 7 (2015), 702--713.Google ScholarDigital Library
- Ankit Toshniwal, Siddarth Taneja, Amit Shukla, Karthik Ramasamy, Jignesh M Patel, Sanjeev Kulkarni, Jason Jackson, Krishna Gade, Maosong Fu, Jake Donham, et al. 2014. Storm@ twitter. In Proceedings of the 2014 ACM SIGMOD international conference on Management of data. 147--156.Google ScholarDigital Library
- Matei Zaharia, Mosharaf Chowdhury, Tathagata Das, Ankur Dave, Justin Ma, Murphy McCauly, Michael J. Franklin, Scott Shenker, and Ion Stoica. 2012. Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing. In Proceedings of the 9th USENIX Symposium on Networked Systems Design and Implementation, NSDI 2012, San Jose, CA, USA, April 25-27, 2012, Steven D. Gribble and Dina Katabi (Eds.). USENIX Association, 15--28. https://www.usenix.org/conference/nsdi12/technical-sessions/presentation/zahariaGoogle ScholarDigital Library
- Steffen Zeuch, Bonaventura Del Monte, Jeyhun Karimov, Clemens Lutz, Manuel Renz, Jonas Traub, Sebastian Breß, Tilmann Rabl, and Volker Markl. 2019. Analyzing efficient stream processing on modern hardware. Proceedings of the VLDB Endowment 12, 5 (2019), 516--530.Google ScholarDigital Library
- Feng Zhang, Lin Yang, Shuhao Zhang, Bingsheng He, Wei Lu, and Xiaoyong Du. 2020. FineStream: Fine-Grained Window-Based Stream Processing on CPU-GPU Integrated Architectures. In 2020 USENIX Annual Technical Conference, USENIX ATC 2020, July 15-17, 2020. 633--647.Google Scholar
Recommendations
A performance and energy comparison of FPGAs, GPUs, and multicores for sliding-window applications
FPGA '12: Proceedings of the ACM/SIGDA international symposium on Field Programmable Gate ArraysWith the emergence of accelerator devices such as multicores, graphics-processing units (GPUs), and field-programmable gate arrays (FPGAs), application designers are confronted with the problem of searching a huge design space that has been shown to ...
A Tradeoff Analysis of FPGAs, GPUs, and Multicores for Sliding-Window Applications
The increasing usage of hardware accelerators such as Field-Programmable Gate Arrays (FPGAs) and Graphics Processing Units (GPUs) has significantly increased application design complexity. Such complexity results from a larger design space created by ...
Stream Aggregation with Compressed Sliding Windows
High performance stream aggregation is critical for many emerging applications that analyze massive volumes of data. Incoming data needs to be stored in a sliding window during processing, in case the aggregation functions cannot be computed ...
Comments