Parallelization of Sobel edge detection algorithm
- 0 Collaborators
This project mainly uses DPC++ programming to realize the edge detection algorithm based on Sobel operator.Use parallelization to process convolution operations to speed up image processing.Through comparative experiments on different devices, the efficiency of the algorithm is tested. ...learn more
Project status: Published/In Market
Intel Technologies
DevCloud,
DPC++,
Intel FPGA,
Intel® Core™ Processors
Overview / Usage
With the advent of the fourth industrial revolution, industrial intelligence has become an unstoppable trend. The rise of machine learning, artificial intelligence, unmanned driving and other fields has caused the CPU to encounter bottlenecks such as low parallelism and insufficient bandwidth in the data processing process, which has brought about the extension of data processing methods and computing architecture. In order to meet the needs of diversified computing, more and more scenarios have begun to introduce hardware such as GPUs and FPGAs for acceleration. Heterogeneous computing has emerged as the times require, and it continues to play an increasingly important role with the development of the computing industry.
The main work of this experiment is to run the Sobel algorithm on multiple different platforms, use different parallel frameworks, and conduct comparative experiments on different devices, and analyze and compare the final results. Verify the characteristics of different frameworks and devices.
Methodology / Approach
Learn to use DPC++ programming to realize the edge detection algorithm based on Sobel operator. Run the parallel code on the FPGA device and GPU device provided by the DevCloud platform. Use images of different sizes and the results of the final operation of different equipment to carry out comparative experiments.
Technologies Used
-
Devcloud 平台
内存:192 GB RAM
CPU Intel® Xeon® Scalable 6128 processors
Intel® Xeon® Scalable 8256 processors
Intel® Xeon® E-2176 P630 processors (with Intel® Graphics Technology)
GPU Intel® Xeon® E-2176 P630 processors (with Intel® Graphics Technology)
Intel® Iris® Xe MAX
FPGA Intel® Arria® 10 FPGAs
Intel® Stratix® 10 FPGAs
开发语言:DPC++
- TaiShan 200 服务器 2280 均衡型
处理器:鲲鹏 920 5250处理器 48核 2.6GHz;
操作系统 Centos 7.6
配置环境: openMPI
开发语言: c++
- 个人 PC 机
操作系统: Ubuntu18.04
处理器: Intel(R)Core(TM)i7-9750H CPU @2.60GHz 2.59GHz
RAM 24.0GB(23.9 GB可用 )
GPU NVIDIA GEFORCE RTX 2060 内存 6GB
CUDA版本: 11.1
opencv: 3.4.10