Accelerating OpenVX through Halide and MLIR
Pages 571 - 584
Abstract
In recent years, as many social media and AI-enabled applications have become increasingly ubiquitous, camera-centric applications have emerged as the most popular category of apps on mobile phones. A programmer can develop a camera application in an hour or less without any knowledge related to this domain by using different API provided by frameworks. This allows for the rapid promotion of this technology. OpenVX is a computer vision framework with vital considerations for performance and portability. This paper proposes a new framework that effectively accelerates OpenVX with Halide and MLIR. Our framework possesses Halide’s properties of decoupling algorithms and also has schedules such as an auto-scheduler. It also has MLIR’s multi-level dialects that structure the operations and the data accesses. To generate more efficient programs, we propose a bridge that can transform the programs written in OpenVX into Halide and then translate from Halide to MLIR. In the process, Our framework attains both Halide’s scheduling and MLIR’s dialect to generate more efficient binary code for execution speed.
References
[1]
Krizhevsky A, Sutskever I, and Hinton GE Imagenet classification with deep convolutional neural networks Communications of the ACM 2017 60 6 84-90
[2]
Baghdadi, R., Ray, J., Romdhane, M. B., Sozzo, E. D., Akkas, A., Zhang, Y., Suriana, P., Kamil, S., & Amarasinghe, S. (2018). Tiramisu: A polyhedral compiler for expressing fast and portable code.
[3]
Benabderrahmane M-W, Pouchet L-N, Cohen A, and Bastoul C Gupta R The polyhedral model is more widely applicable than you think Compiler Construction 2010 Berlin, Heidelberg Springer 283-303
[4]
Hartono, A., Baskaran, M. M., Bastoul, C., Cohen, A., Krishnamoorthy, S., Norris, B., Ramanujam, J., & Sadayappan, P. (2009). Parametric multi-level tiling of imperfectly nested loops. In Proceedings of the 23rd International Conference on Supercomputing. ICS ’09 (pp. 147–157). Association for Computing Machinery, New York, NY, USA.
[5]
Tavarageri, S., Hartono, A., Baskaran, M., Pouchet, L.-N., Ramanujam, J., & Sadayappan, P. (2010). Parametric tiling of affine loop nests. In Proceedings of the 15th Workshop on Compilers for Parallel Computers, Vienna, Austria.
[6]
Maleki, S., Gao, Y., Garzar’n, M. J., Wong, T., & Padua, D. A. (2011). An evaluation of vectorizing compilers. In 2011 International Conference on Parallel Architectures and Compilation Techniques (pp. 372–382).
[7]
Giduthuri, R., & Pulli, K. (2016). OpenVX: A framework for accelerating computer vision. In SIGGRAPH ASIA 2016 Courses. https://doi.org/10.1145/2988458.2988513
[8]
Ragan-Kelley, J., Barnes, C., Adams, A., Paris, S., Durand, F., & Amarasinghe, S. (2013). Halide: A language and compiler for optimizing parallelism, locality, and recomputation in image processing pipelines. In Proceedings of the 34th ACM SIGPLAN Conference on Programming Language Design and Implementation (pp. 519–530). PLDI ’13. Association for Computing Machinery, New York, NY, USA.
[9]
Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., Devin, M., Ghemawat, S., Irving, G., Isard, M., Kudlur, M., Levenberg, J., Monga, R., Moore, S., Murray, D. G., Steiner, B., Tucker, P., Vasudevan, V., Warden, P., Wicke, M., Yu, Y., & Zheng, X. (2016). Tensorflow: A system for large-scale machine learning. In 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16) (pp. 265–283). Retrieved September 20, 2021, from https://www.usenix.org/conference/osdi16/technical-sessions/presentation/abadi
[10]
Lattner, C., Amini, M., Bondhugula, U., Cohen, A., Davis, A., Pienaar, J., Riddle, R., Shpeisman, T., Vasilache, N., & Zinenko, O. (2021). MLIR: Scaling compiler infrastructure for domain specific computation. In 2021 IEEE/ACM International Symposium on Code Generation and Optimization (CGO) (pp. 2–14).
[11]
Bondhugula, U. (2020). High performance code generation in MLIR: An early case study with GEMM. CoRR abs/2003.00532. arXiv:2003.00532
[12]
Wang, E., et al. (2014). Intel Math Kernel Library. In High-performance computing on the Intel® Xeon Phi™. Springer, Cham. Retrieved September 1, 2021, from https://doi.org/10.1007/978-3-319-06486-4_7
[13]
Wang, Q., Zhang, X., Zhang, Y., & Yi, Q. (2013). AUGEM: Automatically generate high performance dense linear algebra Kernels on x86 CPUS. In SC ’13: Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis (pp. 1–12).
[14]
Xianyi, Z., Qian, W., & Yunquan, Z. (2012). Model-driven level 3 BLAS performance optimization on Loongson 3A processor. In 2012 IEEE 18th International Conference on Parallel and Distributed Systems (pp. 684–691).
[15]
Verdoolaege, S. (2010). ISL: An integer set library for the polyhedral model. In Proceedings of the Third International Congress Conference on Mathematical Software (pp. 299–302). ICMS’10. Springer, Berlin, Heidelberg.
[16]
Bastoul, C. (2004). Code generation in the polyhedral model is easier than you think. In PACT’13 IEEE International Conference on Parallel Architecture and Compilation Techniques, Juan-les-Pins, France (pp. 7–16).
[17]
Goto, K., & Geijn, R. A. V. D. (2008). Anatomy of high-performance matrix multiplication. ACM Transactions on Mathematical Software, 34(3).
Recommendations
Accelerating OpenVX Application Kernels Using Halide Scheduling
AbstractIn this study, we investigate how to use a Domain-Specific Language—Halide to accelerate and optimize OpenVX graphs. Halide is a new high-level image processing pipeline language. It offers developers to separate the program into algorithms and ...
A Halide-based Synergistic Computing Framework for Heterogeneous Systems
New programming models have been developed to embrace contemporary heterogeneous machines, each of which may contain several types of processors, e.g., CPUs, GPUs, FPGAs and ASICs. Unlike the conventional ones, which use separate programming schemes for ...
Optimizing memory bandwidth exploitation for OpenVX applications on embedded many-core accelerators
In recent years, image processing has been a key application area for mobile and embedded computing platforms. In this context, many-core accelerators are a viable solution to efficiently execute highly parallel kernels. However, architectural ...
Comments
Please enable JavaScript to view thecomments powered by Disqus.Information & Contributors
Information
Published In
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Publisher
Kluwer Academic Publishers
United States
Publication History
Published: 01 February 2023
Accepted: 30 November 2022
Revision received: 11 November 2022
Received: 08 July 2022
Author Tags
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 23 Nov 2024
Other Metrics
Citations
View Options
View options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in