|
High performance 2D convolution utilizing the AVX512on a multi-core architecture |
|---|---|
| รหัสดีโอไอ | |
| Creator | 1. Isamail Masamae 2. Panyayot Chaikan |
| Title | High performance 2D convolution utilizing the AVX512on a multi-core architecture |
| Publisher | Research and Development Office, Prince of Songkla University |
| Publication Year | 2564 |
| Journal Title | Songklanakarin Journal of Science and Technology (SJST) |
| Journal Vol. | 43 |
| Journal No. | 4 |
| Page no. | 1230-1236 |
| Keyword | AVX512, Advanced Vector Extension, 2D convolution, OpenMP, x64 assembly language |
| URL Website | https://rdo.psu.ac.th/sjstweb/index.php |
| ISSN | 0125-3395 |
| Abstract | Convolution is a time consuming operation, especially for signal and image processing, which led us to develop anefficient implementation of 2D convolution for a multi-core architecture utilizing AVX512 intrinsics and OpenMP. For singleprecision convolution, our algorithm is on average 2.30, 3.88, 5.75, and 19.95 times faster than the IPP, OpenCV, Baziotis'salgorithm, and MKL libraries. For double precision convolution, our algorithm is on average 3.12, 5.10, and 16.95 times fasterthan the OpenCV, Baziotis's algorithm, and MKL libraries. We have also developed a hybrid 2D convolution algorithm, writtenin C and assembly, to further augment the processing speeds for small kernel sizes. |
