Vector Lane Threading

Multi-lane vector processors achieve excellent computational throughput for programs with high data-level parallelism (DLP). However, application phases without significant DLP are unable to fully utilize the datapaths in the vector lanes. In this paper, we propose vector lane threading (VLT), an architectural enhancement that allows idle vector lanes to run short-vector or scalar threads. VLTenhanced vector hardware can exploit both data-level and thread-level parallelism to achieve higher performance. We investigate implementation alternatives for VLT, focusing mostly on the instruction issue bandwidth requirements. We demonstrate that VLT’s area overhead is small. For applications with short vectors, VLT leads to additional speedup

• Type: Conference
• Year: 2006
• Where: ICPP
• Authors: Suzanne Rivoire, Rebecca Schultz, Tomofumi Okuda, Christos Kozyrakis
• Keywords: Distributed And Parallel Computing, ICPP 2006, Idle Vector Lanes, Multi-lane Vector Processors, Vector Lanes