AMD GPU-based Selected Acceleration of PRACH and PUSCH in 5G NR
收藏DataCite Commons2025-03-04 更新2025-04-16 收录
下载链接:
https://dataverse.lib.nycu.edu.tw/citation?persistentId=doi:10.57770/RGHEIH
下载链接
链接失效反馈官方服务:
资源简介:
The software-defined physical layer is a recent trend in the 5G/B5G open radio access network. Most implementations are based on a central processing unit, i.e., Intel XEON and ARM. This work proposes an uplink acceleration on an AMD graphics processor unit and OpenCL to show a new opportunity in 5th Generation New Radio HIGH-PHY. The key accelerators are the physical random access channel and the physical uplink shared channel under a real-time test environment. Referring to 5G NR, we balance and match slot formats and GPU resources to ensure the PHY works properly on the AMD MI.50 with fine-tuned parallelism and memory mapping. This work suggests that the AMD GPU is suitable for 5G ORAN.
提供机构:
NYCU Dataverse
创建时间:
2025-03-04



