Conference paper · 2022 · Open access

Enabling On-Device Smartphone GPU based Training: Lessons Learned

Anish Kumar Das, Young D. Kwon, Jagmohan Chauhan, Cecilia Mascolo

IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), 533–538 · Published 21 March 2022 · DOI 10.1109/percomworkshops53856.2022.9767442

Open PDFView at the publisherCite this paper

In brief

“Enabling On-Device Smartphone GPU based Training: Lessons Learned” is a 2022 conference paper by Anish Kumar Das and 3 co-authors, published in IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops). OpenAlex counts 19 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
19
Most cited in
2024 (7)
Field-weighted impact
1.81
References
19
Authors
4

Abstract

Deep Learning (DL) has shown impressive performance in many mobile applications. Most existing works have focused on reducing the computational and resource overheads of running Deep Neural Networks (DNN) inference on resource-constrained mobile devices. However, the other aspect of DNN operations, i.e. training (forward and backward passes) on smartphone GPUs, has received little attention thus far. To this end, we conduct an initial analysis to examine the feasibility of on-device training on smartphones using mobile GPUs. We first employ the open-source mobile DL framework (MNN) and its OpenCL backend for running compute kernels on GPUs. Next, we observed that training on CPUs is much faster than on GPUs and identified two possible bottlenecks related to this observation: (i) computation and (ii) memory bottlenecks. To solve the computation bottleneck, we optimize the OpenCL backend’s kernels, showing 2x improvements (40-70 GFLOPs) over CPUs (15-30 GFLOPs) on the Snapdragon 8 series processors. However, we find that the full DNN training is still much slower on GPUs than on CPUs, indicating that memory bottleneck plays a significant role in the lower performance of GPU over CPU. The data movement takes almost 91% of training time due to the low bandwidth. Lastly, based on the findings and failures during our investigation, we present limitations and practical guidelines for future directions.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Das, A. K., Kwon, Y. D., Chauhan, J., & Mascolo, C. (2022). Enabling On-Device Smartphone GPU based Training: Lessons Learned. IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), 533–538. https://doi.org/10.1109/percomworkshops53856.2022.9767442

MLA 9

Das, Anish Kumar, et al. “Enabling On-Device Smartphone GPU based Training: Lessons Learned.” IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), 2022, pp. 533–538. https://doi.org/10.1109/percomworkshops53856.2022.9767442.

Chicago

Das, Anish Kumar, Young D. Kwon, Jagmohan Chauhan, and Cecilia Mascolo. 2022. “Enabling On-Device Smartphone GPU based Training: Lessons Learned.” IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops): 533–538. https://doi.org/10.1109/percomworkshops53856.2022.9767442.

Harvard

Das, A. K. et al. (2022) ‘Enabling On-Device Smartphone GPU based Training: Lessons Learned’, IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), pp. 533–538. doi:10.1109/percomworkshops53856.2022.9767442.

IEEE

A. K. Das, Y. D. Kwon, J. Chauhan, and C. Mascolo, “Enabling On-Device Smartphone GPU based Training: Lessons Learned,” IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), pp. 533–538, 2022, doi: 10.1109/percomworkshops53856.2022.9767442.

BibTeX

@inproceedings{das2022enabling,
  title     = {{Enabling On-Device Smartphone GPU based Training: Lessons Learned}},
  author    = {Das, Anish Kumar and Kwon, Young D. and Chauhan, Jagmohan and Mascolo, Cecilia},
  booktitle = {IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops)},
  year      = {2022},
  pages     = {533--538},
  doi       = {10.1109/percomworkshops53856.2022.9767442},
  url       = {https://doi.org/10.1109/percomworkshops53856.2022.9767442}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Citations by year

YearCitationsShare of the best year
202211
202322
202477
202544
202655

Its field-weighted citation impact is 1.81: it has been cited 1.81 times as often as the average work of the same field, type and year (1.00 is the world average).

Topic and keywords

Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques

  • mobile GPUs
  • OpenCL
  • kernel optimization
  • memory bottleneck

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Das, A. K., Kwon, Y. D., Chauhan, J., & Mascolo, C. (2022). Enabling On-Device Smartphone GPU based Training: Lessons Learned. IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), 533–538. https://doi.org/10.1109/percomworkshops53856.2022.9767442

How many times has this paper been cited?

19 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Anish Kumar Das, Young D. Kwon, Jagmohan Chauhan, Cecilia Mascolo.

What is the DOI of this paper?

10.1109/percomworkshops53856.2022.9767442, permanently at https://doi.org/10.1109/percomworkshops53856.2022.9767442. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.