Highly parallel GEMV with register blocking method on GPU architecture

Jian Yin, Hui Yu, Weizhi Xu, Yuxuan Wang, Zhu Tian, Yingping Zhang, Bochuan Chen

Abstract: Highlights•We propose a register blocking method for GEMV on GPU.•The proposed method can improve the parallelism and reuse data on chip at the same time.•Different block sizes are tested to found the best block size on a GPU platform.