Custom Built of Smart Computing Platform for Supporting Optimization Methods and Artificial Intelligence Research
PakCarik: GPU-Accelerated Platform for AI Researches
DOI:
https://doi.org/10.53560/PPASA(58-sp1)733Keywords:
Artificial Intelligence, Machine Learning, Multi-objective Optimization, Graphics Processing Unit Accelerator, High Throughput ComputingAbstract
This paper describes a prototype of a computing platform dedicated to artificial intelligence explorations. The platform, dubbed as PakCarik, is essentially a high throughput computing platform with GPU (graphics processing units) acceleration. PakCarik is an Indonesian acronym for Platform Komputasi Cerdas Ramah Industri Kreatif, which can be translated as “Creative Industry friendly Intelligence Computing Platform”. This platform aims to provide complete development and production environment for AI-based projects, especially to those that rely on machine learning and multiobjective optimization paradigms. The method for constructing PakCarik was based on a computer hardware assembling technique that uses commercial off-the-shelf hardware and was tested on several AI-related application scenarios. The testing methods in this experiment include: high-performance lapack (HPL) benchmarking, message passing interface (MPI) benchmarking, and TensorFlow (TF) benchmarking. From the experiment, the authors can observe that PakCarik's performance is quite similar to the commonly used cloud computing services such as Google Compute Engine and Amazon EC2, even though falls a bit behind the dedicated AI platform such as Nvidia DGX-1 used in the benchmarking experiment. Its maximum computing performance was measured at 326 Gflops. The authors conclude that PakCarik is ready to be deployed in real-world applications and it can be made even more powerful by adding more GPU cards in it.
References
A. Ng, Baidu's Chief Scientist on Intersection of Supercomputing, Machine Learning. [Online] from www.nextplatform.com/2016/04/01/baidus-chiefscientist-intersection-supercomputing-machinelearning/(2016). [Accessed on July 27th 2019].
H. Nasser, Y. Hafeer and S. Ali. Towards software testing as a service for software as a service based on cloud computing model. Proceedings of the Pakistan Academy of Sciences A. Physical and Computational Sciences 55 (4): 1–8 (2018)
S.L. Graham., M. Snir, and C.A. Patterson. Getting Up to Speed: The Future of Supercomputing. National Academies Press (2005). DOI: 10.17226/11148
U. Khan and U. Naeem. Practices for clients in the adoption of hybrid cloud. Proceedings of the Pakistan Academy of Sciences A Physical and Computational Sciences. 54 (1): 13–32 (2017)
V.V. Kindratenko., J.J. Enos., G. Shi., M.T. Showerman., G.W. Arnold, J.E. Stone., J.C. Phillips, and W.M. Hwu, GPU clusters for highperformance computing. IEEE International Conference on Cluster Computing and Workshops (CLUSTER'09), (2009) pp. 1–8. DOI: 10.1109/CLUSTR.2009.5289128
NVIDIA. NVIDIA DGX-1 With Tesla V100 System Architecture. [Online] from http://images.nvidia.com/content/pdf/dgx1-v100-system-architecturewhitepaper.pdf (2018). [Accessed on July 27th 2019].
N.A. Gawande., J.A. Daily., C. Siegel., N.R. Tallent, and A. Vishnum, Scaling deep learning workloads: Nvidia dgx-1/pascal and intel knights landing. Future Generation Computer Systems 108:1162–1172 (2020). DOI:10.1016/j.future.2018.04.073
X. Glorot, and Y. Bengio. Understanding the difficulty of training deep feedforward neural networks. Thirteenth International Conference on Artificial Intelligence and Statistics (AISTATS), (Sardinia, Italy, 2010).
C. Berry., G. Hall., B. Matuszewski, and L.K. Shark. A comparison of architectures and evaluation of metrics for in-stream machine learning algorithms in industry 4.0 applications. 30th International Conference on Condition Monitoring and Diagnostic Engineering Management. (Preston and Grange-Over-Sands, UK, 2017).
L.Y. Joo., T.S. Yin., E. Xu., E. Thia., P.F. Chia., C.W.K. Kuah, and K.K. He, A feasibility study using interactive commercial off-the-shelf computer gaming in upper limb rehabilitation in patients after stroke. Journal of Rehabilitation Medicine, 42(5):437–441 (2010). DOI: 10.2340/16501977-0528
C. Szegedy., V. Vanhoucke., S. Ioffe, J. Shlens, and Z. Wojna. Rethinking the inception architecture for computer vision. The IEEE conference on computer vision and pattern recognition, pp. 2818–2826.(2016).
H. Kaiming., X. Zhang., R. Shaoqing, and J. Sun. Deep residual learning for image recognition. The IEEE conference on computer vision and pattern recognition, pp. 770–778. (2016).
Tensorflowx.org. Benchmarks. [Online] from https://www.tensorflow.org/guide/performance/benchmarks [Accessed on August 9th 2019].
A.Lee, C. Yau, M.B. Giles, A. Doucet and C.C. Holmes. On the utility of graphics cards to perform massively parallel simulation of advanced Monte Carlo methods. Journal of Computational and Graphical Statistics, 19(4): 769–789 (2010).
S. Shi, Q. Wang, P. Xu and X. Chu. Benchmarking state-of-the-art deep learning software tools. 7th International Conference on Cloud Computing and Big Data (CCBD), 2016, pp. 99–104. (2016). DOI:10.1109/CCBD.2016.029.
R. Makhlouf,. Cloudy transaction costs: A dive into cloud computing economics. Journal of Cloud Computing, 9 (1):1–11. (2020). DOI: 10.1186/s13677-019-0149-4
O. Adjoua L. Lagardère, L.H. Jolly, A. Durocher, T.Very , I. Dupays, Z. Wang, T.J. Inizan , F. Célerse, P. Ren , J.W Ponder, J.P. Piquemal Tinker-HP: Accelerating molecular dynamics simulations of large complex systems with advanced point dipole polarizable force fields using GPUs and multi- GPU systems. Journal of Chemical Theory and Computation, 17 (4), 2034–2053 (2021). DOI:10.1021/acs.jctc.0c01164
