DETECTING DEPENDENCIES IN DATAFLOW KERNELS
Keywords:
dataflow, dependency graph creation, translating the code automatically, artificial intelligenceAbstract
Purpose
The importance of artificial intelligence in detecting crime is constantly rising. Artificial intelligence algorithms are known to be computation demanding. It is common that the execution of artificial intelligence algorithms in modern cloud infrastructures is performed using dedicated hardware. Computer hardware that includes dataflow hardware along with conventional processors is capable of accelerating artificial intelligence algorithms. They also exist in a form of nodes suitable for cluster or cloud infrastructure.
Design/Methods/Approach
This article describes a method to automatically detect dependencies in dataflow kernels. The tool for constructing dependency graph is presented. The example code is chosen to depict the process of detecting dependencies. The proposed method is also suitable for hybrid processors consisting of both control-flow and dataflow hardware on the same chip die.
Findings
The dependency graph creation enable automation of translating algorithms from control-flow to dataflow hardware. This can lead to even higher acceleration factor when using dataflow hardware for the execution of artificial intelligence algorithms.
Originality/Value
While there are frameworks that can serve for designing dataflow hardware and using the hardware for executing algorithms, not much work has been performed to automatically transform control-flow into dataflow algorithms by detecting dependencies among kernels and instructions. The dependency graph was designed for this purpose.
References
Brkić, M. (2023). A system architecture for preventing social engineering attacks via e-mail. Journal of Computer and Forensic Sciences, 2(1), 43-52.
Perić, D., Bogdanoski, M., & Maček, N. (2022). Application of convolutional neural networks to spoken words evaluation based on lip movements without accompanying sound signal. Journal of Computer and Forensic Sciences, 1(1), 7-16.
Popović, B. (2023). Speech enhancement by CycleGAN using feature map regularization. Journal of Computer and Forensic Sciences, 2(1), 19-28.
Hangaragi, S., Singh, T., & Neelima, N. (2023). Face detection and Recognition using Face Mesh and deep neural network. Procedia Computer Science, 218, 741-749.
Naik, M. N., Kaur, A., Yathiraju, N., Das, S., & Pant, K. (2023, May). Improved and Accurate Face Mask Detection Using Machine Learning in the Crowded Places. In 2023 3rd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) (pp. 572-576). IEEE.
Han, X., You, Q., Wang, C., Zhang, Z., Chu, P., Hu, H., ... & Liu, Z. (2023). Mmptrack: Large-scale densely annotated multi-camera multiple people tracking benchmark. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 4860-4869).
William, P., Shrivastava, A., Shunmuga Karpagam, N., Mohanaprakash, T. A., Tongkachok, K., & Kumar, K. (2023). Crime analysis using computer vision approach with machine learning. In Mobile Radio Communications and 5G Networks: Proceedings of Third MRCN 2022 (pp. 297-315). Singapore: Springer Nature Singapore.
Bharadiya, J. P., Thomas, R. K., & Ahmed, F. (2023). Rise of Artificial Intelligence in Business and Industry. Journal of Engineering Research and Reports, 25(3), 85-103.
Zhang, B., Zhu, J., & Su, H. (2023). Toward the third generation artificial intelligence. Science China Information Sciences, 66(2), 121101.
Lam, M. S., & Wilson, R. P. (1992). Limits of control flow on parallelism. ACM SIGARCH Computer Architecture News, 20(2), 46-57.
Milutinovic, V., Salom, J., Veljovic, D., Korolija, N., Markovic, D., & Petrovic, L. (2017). DataFlow supercomputing essentials: research, development and education. Springer.
Babović, Z., Bajat, B., Đokić, V., Đorđević, F., Drašković, D., Filipović, N., ... & Zak, S. (2023). Research in computing-intensive simulations for nature-oriented civil-engineering and related scientific fields, using machine learning and big data: an overview of open problems. Journal of Big Data, 10(1), 73. Springer.
Kos, A., Tomažič, S., Salom, J., Trifunovic, N., Valero, M., & Milutinovic, V. (2015). New benchmarking methodology and programming model for big data processing. International Journal of Distributed Sensor Networks, 11(8), 271752.
Milutinovic, V. (1986). Guest editor's introduction gaas microprocessor technology. Computer, 19(10), 10-13.
Fan, Z., Li, W., Tang, S., An, X., Ye, X., & Fan, D. (2023, August). Improving utilization of dataflow architectures through software and hardware co-design. In European Conference on Parallel Processing (pp. 245-259). Cham: Springer Nature Switzerland.
Trifunovic, N., Milutinovic, V., Salom, J., & Kos, A. (2015). Paradigm shift in big data supercomputing: dataflow vs. controlflow. Journal of Big Data, 2(1), 1–9. Springer.
Milutinović, V., Furht, B., Obradović, Z., & Korolija, N. (2016). Advances in high performance computing and related issues. Mathematical problems in engineering, 2016.
Korolija, N., Djukic, T., Milutinovic, V., & Filipovic, N. (2013). Accelerating Lattice-Boltzman method using Maxeler dataflow approach. The IPSI BgD Transactions on Internet Research, 34.
Cvijetić, D., Korolija, N., & Vojinović, M. (2022). Possibilities for Parallelizing Simplicial Complexes Simulation.
Kos, A., Ranković, V., & Tomažič, S. (2015). Sorting networks on Maxeler dataflow supercomputing systems. In Advances in computers (Vol. 96, pp. 139-186). Elsevier.
Umek, A., & Kos, A. (2016). The role of high performance computing and communication for real-time biofeedback in sport. Mathematical problems in engineering, 2016.
Popovic, J., Bojic, D., & Korolija, N. (2015). Analysis of task effort estimation accuracy based on use case point size. IET Software, 9(6), 166-173.
Popović, J., Korolija, N., Marković, Ž., & Bojić, D. (2017, November). The influence of non-functional requirements in UCP method on the accuracy of effort estimates. In 2017 25th Telecommunication Forum (TELFOR) (pp. 1-4). IEEE.
Korolija, N. (2023, May). Drawbacks of Programming Dataflow Architectures and Methods to Overcome Them. In Serbian International Conference on Applied Artificial Intelligence (pp. 57-70). Cham: Springer Nature Switzerland.
Korolija, N., Bojic, D., Hurson, A. R., & Milutinovic, V. (2022). A runtime job scheduling algorithm for cluster architectures with dataflow accelerators. Advances in Computers, 201.
Huang, K., Liu, Y., Korolija, N., Carulli, J. M., & Makris, Y. (2015). Recycled IC detection based on statistical methods. IEEE transactions on computer-aided design of integrated circuits and systems, 34(6), 947–960.
Trifunovic, N., Milutinovic, V., Korolija, N., & Gaydadjiev, G. (2016). An AppGallery for dataflow computing. Journal of Big Data, 3(1), 1–30. Springer.
Milutinovic, V., Salom, J., Veljovic, D., Korolija, N., Markovic, D., Petrovic, L., ... & Petrovic, L. (2017). Maxeler AppGallery Revisited. DataFlow Supercomputing Essentials: Research, Development and Education, 3-18.
Babović, Z., Bajat, B., Barac, D., Bengin, V., Đokić, V., Đorđević, F., ... & Zak, S. (2023). Teaching computing for complex problems in civil engineering and geosciences using big data and machine learning: synergizing four different computing paradigms and four different management domains. Journal of Big Data, 10(1), 89. Springer.
Milutinovic, V. (1996). The best method for presentation of research results. IEEE TCCA Newsletter, (9), 1–6.
Banković, M., Filipović, V., Graovac, J., Hadži-Purić, J., Hurson, A. R., Kartelj, A., ... & Živković, M. (2020). Teaching graduate students how to review research articles and respond to reviewer comments. In Advances in Computers (Vol. 116, No. 1, pp. 1–63). Elsevier.
Korolija, N., Jelisavčić, V., Minchev, Z., & Milutinović, V. (2022). Towards hybrid control-flow and dataflow architectures. Archibald Reiss Days, 12.
Milutinović, V., Azer, E. S., Yoshimoto, K., Klimeck, G., Djordjevic, M., Kotlar, M., ... & Ratkovic, I. (2021, June). The ultimate dataflow for ultimate supercomputers-on-a-chip, for scientific computing, geo physics, complex mathematics, and information processing. In 2021 10th Mediterranean Conference on Embedded Computing (MECO) (pp. 1–6). IEEE.
Korolija, N., Popović, J., Cvetanović, M., & Bojović, M. (2017). Dataflow-based parallelization of control-flow algorithms. In Advances in computers (Vol. 104, pp. 73–124). Elsevier.
Miladinović, D., Bojović, M., Jelisavčić, V., & Korolija, N. (2022, June). Hybrid manycore dataflow processor. In Proceedings, IX International Conference IcETRAN, Novi Pazar, Serbia (pp. 6-9).
Korolija, N., & Štrbac-Savić, S. (2024). Merging control-flow and dataflow architectures on a single chip. Journal of Computer and Forensic Sciences.
Milfeld, K. & Korolija, N. (2022). Towards hybrid supercomputing architectures. Journal of Computer and Forensic Sciences, 1(1), 47-54.
Korolija, N., Furht, B., & Milutinović, V. (2023). Fine Grain Algorithm Parallelization on a Hybrid Control-flow and Dataflow Processor.
Milutinović, V., Trifunović, N., Korolija, N., Popović, J., & Bojić, D. (2017, November). Accelerating program execution using hybrid control flow and dataflow architectures. In 2017 25th Telecommunication Forum (TELFOR) (pp. 1–4). IEEE.
Milutinovic, V., Salom, J., Veljovic, D., Korolija, N., Markovic, D., & Petrovic, L. (2017). Transforming applications from the control flow to the dataflow paradigm. In Dataflow supercomputing essentials (pp. 107–129). Springer, Cham.
Filippas, D., Peltekis, C., Titopoulos, V., Kansizoglou, I., Sirakoulis, G., Gasteratos, A., & Dimitrakopoulos, G. (2024). A High-Level Synthesis Library for Synthesizing Efficient and Functional-Safe CNN Dataflow Accelerators. IEEE Access.
Yu, Z., Sreeram, S., Agrawal, K., Wu, J., Montgomerie-Corcoran, A., Zhang, C., ... & Zhao, Y. (2024). HASS: Hardware-Aware Sparsity Search for Dataflow DNN Accelerator. arXiv preprint arXiv:2406.03088.
