Accelerating Solar Panel Snow Detection: CPU vs. GPU Training Performance of VGG-19 and ResNet-50

Önder Eyecioglu, Batuhan Hangun, Abdulkadir Turkan

Abstract


Solar panel snow detection is critical for maintaining photovoltaic system efficiency in cold climates. While deep learning models such as VGG-19 and ResNet-50 achieve near-perfect classification accuracy for this task, their training demands substantial computational resources. This study presents a systematic performance analysis of CNN training on the TRUBA national high-performance computing infrastructure, comparing multi-threaded CPU parallelization (1–112 threads on Intel Xeon Platinum 8480+) with GPU acceleration (NVIDIA V100). Experiments reveal that the model architecture fundamentally determines CPU scaling behavior: VGG-19, dominated by large fully connected layers, achieves 28× speedup at 56 threads (serial fraction f ? 1.8%), whereas ResNet-50, composed of many small residual operations, scales to only 1.56× (f ? 63.4%). GPU training provides 163–1067× acceleration over single-threaded CPU execution for both architectures. Classification quality (accuracy and F1-score ? 0.995) remains invariant across all hardware configurations, confirming that training infrastructure selection is purely a performance optimization decision.

Keywords


solar panel snow detection, CNN training, GPU acceleration, CPU parallelization, Amdahl's law

Full Text:

PDF

References


International Energy Agency, “World energy outlook 2023.” https://www.iea.org/reports/world-energy-outlook-2023, 2023. Last accessed: March 2026.

International Renewable Energy Agency, “Renewable capacity statistics 2024.” https://www.irena.org/publications/2024/Mar/Renewable-capacity-statistics-2024, 2024. Last accessed: March 2026.

Z. Quan, H. Lu, C. Zheng, W. Zhao, Y. Xu, J. Qin, and F. An, “Experimental measurement and numerical simulation on the snow-cover process of solar photovoltaic modules and its impact on photoelectric conversion efficiency,” Coatings, vol. 13, no. 2, p. 427, 2023, doi: 10.3390/coatings13020427

R. W. Andrews, A. Pollard, and J. M. Pearce, “The effects of snowfall on solar photovoltaic performance,” Solar Energy, vol. 92, pp. 84–97, 2013, doi: 10.1016/j.solener.2013.02.014

R. E. Pawluk, Y. Chen, and Y. She, “Photovoltaic electricity generation loss due to snow – a literature review on influence factors, estimation, and mitigation,” Renewable and Sustainable Energy Reviews, vol. 107, pp. 171–182, 2019, doi: 10.1016/j.rser.2018.12.031

S. Deitsch, V. Christlein, S. Berger, C. Buerhop-Lutz, A. Maier, F. Gallwitz, and C. Riess, “Automatic classification of defective photovoltaic module cells in electroluminescence images,” Solar Energy, vol. 185, pp. 455–468, 2019, doi: 10.1016/j.solener.2019.02.067

M. W. Akram, G. Li, Y. Jin, X. Chen, C. Zhu, X. Zhao, A. Khaliq, M. Faheem, and A. Ahmad, “CNN based automatic detection of photovoltaic cell defects in electroluminescence images,” Energy, vol. 189, p. 116319, 2019, doi: 10.1016/j.energy.2019.116319

X. Li, Q. Yang, Z. Lou, and W. Yan, “Deep learning based module defect analysis for large-scale photovoltaic farms,” IEEE Transactions on Energy Conversion, vol. 34, no. 1, pp. 520–529, 2019, doi: 10.1109/TEC.2018.2873358

I. Kishor, U. Mamodiya, V. Patil, and N. Naik, “AI-integrated autonomous robotics for solar panel cleaning and predictive maintenance using drone and ground-based systems,” Scientific Reports, vol. 15, no. 1, p. 32187, 2025, doi: 10.1038/s41598-025-17313-6

O. Ozturk, B. Hangun, and O. Eyecioglu, “Detecting snow layer on solar panels using deep learning,” in 2021 10th International Conference on Renewable Energy Research and Application (ICRERA), pp. 434–438, 2021, doi: 10.1109/ICRERA52334.2021.9598700

K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556, 2015. Published at ICLR 2015, doi: 10.48550/arXiv.1409.1556

K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016, doi: 10.1109/CVPR.2016.90

B. Shah and H. Bhavsar, “Time complexity in deep learning models,” Procedia Computer Science, vol. 215, pp. 202–210, 2022, doi: 10.1016/j.procs.2022.12.023

M. Taghavi and M. Shoaran, “Hardware complexity analysis of deep neural networks and decision tree ensembles for real-time neural data classification,” in 2019 9th International IEEE/EMBS Conference on Neural Engineering (NER), pp. 407–410, 2019, doi: 10.1109/NER.2019.8716983

S. Kornblith, J. Shlens, and Q. V. Le, “Do better ImageNet models transfer better?,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2656–2666, 2019, doi: 10.1109/CVPR.2019.00277

N. Tajbakhsh, J. Y. Shin, S. R. Gurudu, R. T. Hurst, C. B. Kendall, M. B. Gotway, and J. Liang, “Convolutional neural networks for medical image analysis: Full training or fine tuning?,” IEEE Transactions on Medical Imaging, vol. 35, no. 5, pp. 1299–1312, 2016, doi: 10.1109/TMI.2016.2535302

Intel Corporation, “oneAPI deep neural network library (oneDNN).” https://www.intel.com/content/www/us/en/developer/tools/oneapi/onednn.html, 2024. Last accessed: March 2026.

NVIDIA Corporation, “CUDA toolkit documentation v12.0.” https://docs.nvidia.com/cuda/, 2024. Last accessed: March 2026.

NVIDIA Corporation, “cuDNN developer guide.” https://docs.nvidia.com/deeplearning/cudnn/, 2024. Last accessed: March 2026.

J. Huang, T. Wang, X. Zhu, M. Wei, T. Wu, X. Wu, and M. Huang, “A parallel optimization of the fast algorithm of convolution neural network on CPU,” in 2018 10th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA), pp. 5–9, 2018.

doi: 10.1109/ICMTMA.2018.00009

D. Datta, D. Mittal, N. P. Mathew, and J. Sairabanu, “Comparison of performance of parallel computation of CPU cores on CNN model,” in 2020 International Conference on Emerging Trends in Information Technology and Engineering (ic-ETITE), pp. 1–8, 2020,

doi: 10.1109/ic-ETITE47903.2020.142

B. Hangun and S. Bayar, “An OpenMP-based parallel implementation of image enhancement technique for dark images,” Signal, Image and Video Processing, vol. 18, no. 5, pp. 4115–4130, 2024, doi: 10.1007/s11760-024-03058-8

T. Ben-Nun and T. Hoefler, “Demystifying parallel and distributed deep learning: An in-depth concurrency analysis,” ACM Computing Surveys, vol. 52, no. 4, pp. 1–43, 2019, doi: 10.1145/3320060

A. Saad Althiban, H. M. Alharbi, L. A. Al Khuzayem, and F. Elbouraey Eassa, “Automated models for predicting software defects in hybrid message passing interface (MPI) and open multi-processing (OpenMP) parallel programs using deep learning,” IEEE Access, vol. 13, pp. 65939–65954, 2025, doi: 10.1109/ACCESS.2025.3559807

B. Hangün and Ö. Eyecio?lu, “Performance comparison between OpenCV built-in CPU and GPU functions on image processing operations,” International Journal of Engineering Science and Application, vol. 1, no. 2, pp. 34–41, 2017.

M. Jordà, P. Valero-Lara, and A. J. Peña, “cuConv: CUDA implementation of convolution for CNN inference,” Cluster Computing, vol. 25, no. 2, pp. 1459–1473, 2022, doi: 10.1007/s10586-021-03494-y

H. Dong, T. Li, J. Leng, L. Kong, and G. Bai, “GCN: GPU-based cube CNN framework for hyperspectral image classification,” in 2017 46th International Conference on Parallel Processing (ICPP), pp. 41–49, 2017, doi: 10.1109/ICPP.2017.13

S. Shi, Q. Wang, P. Xu, and X. Chu, “Benchmarking state-of-the-art deep learning software tools,” in Proceedings of the 7th IEEE International Conference on Cloud Computing and Big Data Analytics (ICCCBDA), pp. 99–104, 2018, doi: 10.1109/ICCCBDA.2018.8386507

C. Coleman, D. Narayanan, D. Kang, T. Zhao, J. Zhang, L. Nardi, P. Bailis, K. Olukotun, C. Ré, and M. Zaharia, “DAWNBench: An end-to-end deep learning benchmark and competition,” in NeurIPS ML Systems Workshop, 2017.

S. Markidis, S. W. D. Chien, E. Laure, I. B. Peng, and J. S. Vetter, “NVIDIA tensor core programmability, performance and precision,” in IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), pp. 522–531, 2018, doi: 10.1109/IPDPSW.2018.00091

G. M. Amdahl, “Validity of the single processor approach to achieving large scale computing capabilities,” AFIPS Conference Proceedings, vol. 30, pp. 483–485, 1967, doi: 10.1145/1465482.1465560

M. D. Hill and M. R. Marty, “Amdahl’s law in the multicore era,” IEEE Computer, vol. 41, no. 7, pp. 33–38, 2008, doi: 10.1109/MC.2008.209

TUBITAK ULAKBIM, “TRUBA high performance computing center.” https://www.truba.gov.tr/, 2024. Last accessed: March 2026.

J. Y. Lin, C. Ograda-Bratu, S. Huo, and S. Keshav, “Detecting snow covered solar panels dataset.” Scholars Portal Dataverse.doi: 10.5683/SP2/ITQPHZ

A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “PyTorch: An imperative style, high-performance deep learning library,” in Advances in Neural Information Processing Systems 32 (NeurIPS), pp. 8024–8035, 2019, doi: 10.48550/arXiv.1912.01703

J. L. Gustafson, “Reevaluating Amdahl’s law,” Communications of the ACM, vol. 31, no. 5, pp. 532–533, 1988, doi: 10.1145/42411.42415




DOI (PDF): https://doi.org/10.20508/ijsmartgrid.v10i3.815.g431

Refbacks

  • There are currently no refbacks.


www.ijsmartgrid.com; www.ijsmartgrid.org

ilhcol@gmail.com; icolak@gazi.edu.tr

Online ISSN: 2602-439X

Publisher: ilhami COLAK

https://www.ilhamicolak.org/english.htm

Cited in SCOPUS Q1, SCIMAGO, Google Scholar and CrossRef

 

 

LinkedIn Logo Follow us on LinkedIn

Google Scholar Statistics of ijSmartGrid

https://scholar.google.com/citations?user=qP9ZUKAAAAAJ&hl=tr&authuser=1 

 

SCIMAGO information:

https://www.scimagojr.com/journalsearch.php?q=21101271431&tip=sid&clean=0#google_vignette

Q2, h-index:13