| Presentation | 2023-03-23 Modeling performance of deep learning for image recognition on a GPU server Tetsuya Matsushita, Shinobu Miwa, Hayato Yamaki, Hiroki Honda, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Deep learning frameworks have an input pipeline that executes data transfer and processes on CPU and GPU in a pipeline manner. When using the input pipeline, the required performance of each hardware in a GPU server differs depending on the combination of hardware. In this research, we model the execution time of an image recognition application based on deep learning by using various architectural parameters on the GPU server, in order to clarify the required performance of each hardware. As a result of training ResNet50 with ImageNet on two types of GPU servers, our model can predict the execution time with a mean absolute percentage error of 10.8%. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | deep learning / GPU server / performance modeling |
| Paper # | CPSY2022-39,DC2022-98 |
| Date of Issue | 2023-03-16 (CPSY, DC) |
| Conference Information | |
| Committee | DC / CPSY / IPSJ-SLDM / IPSJ-EMB / IPSJ-ARC |
|---|---|
| Conference Date | 2023/3/23(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Amagi Town Disaster Prevention Center (Tokunoshima) |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Tatsuhiro Tsuchiya(Osaka Univ.) / Michihiro Koibuchi(NII) / Hiroyuki Ochi(Ritsumeikan Univ.) / / Hiroshi Inoue(Nagoya Institute of Technology) |
| Vice Chair | Toshinori Hosokawa(Nihon Univ.) / Kota Nakajima(Fujitsu Lab.) / Tomoaki Tsumura(Nagoya Inst. of Tech.) |
| Secretary | Toshinori Hosokawa(Nihon Univ.) / Kota Nakajima(Chiba Univ.) / Tomoaki Tsumura(JAIST) / (Hitachi) / (Tokyo Inst. of Tech.) / (Meiji Univ.) |
| Assistant | / Ryohei Kobayashi(Tsukuba Univ.) / Takaaki Miyajima(Meiji Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Dependable Computing / Technical Committee on Computer Systems / Special Interest Group on System and LSI Design Methodology / Special Interest Group on Embedded Systems / Special Interest Group on System Architecture |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Modeling performance of deep learning for image recognition on a GPU server |
| Sub Title (in English) | |
| Keyword(1) | deep learning |
| Keyword(2) | GPU server |
| Keyword(3) | performance modeling |
| 1st Author's Name | Tetsuya Matsushita |
| 1st Author's Affiliation | The University of Electro-Communications(UEC) |
| 2nd Author's Name | Shinobu Miwa |
| 2nd Author's Affiliation | The University of Electro-Communications(UEC) |
| 3rd Author's Name | Hayato Yamaki |
| 3rd Author's Affiliation | The University of Electro-Communications(UEC) |
| 4th Author's Name | Hiroki Honda |
| 4th Author's Affiliation | The University of Electro-Communications(UEC) |
| Date | 2023-03-23 |
| Paper # | CPSY2022-39,DC2022-98 |
| Volume (vol) | vol.122 |
| Number (no) | CPSY-451,DC-452 |
| Page | pp.pp.31-36(CPSY), pp.31-36(DC), |
| #Pages | 6 |
| Date of Issue | 2023-03-16 (CPSY, DC) |