Search
[Cover Paper] Cao Zhuang, Zhang Tian, et al. from National University of Defense Technology | Efficient Mapping and Flexible Interconnection: Multi-FPGA-Based Acceleration for 3D Convolutional Neural
1 day ago /
Read about 0 minute
Author：小编
Three-dimensional convolutional neural networks (3D CNNs) hold significant application value and potential in lung nodule detection. However, due to their computational complexity and high memory overhead, accelerating them with a single FPGA hardware presents challenges. To address this, this paper proposes an efficient mapping strategy for multi-FPGA platforms, leveraging system-level large-scale parallelism to enhance computational efficiency.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
