过去一年中添加的文章，按日期排序
WASP: Exploiting GPU Pipeline Parallelism with Hardware-Accelerated Automatic Warp Specialization
131 天前 - … Finally, we design and implement a compiler that can … All warps in processing blocks
share a physical register … a program into sub-contexts for execution on parallel processors …
share a physical register … a program into sub-contexts for execution on parallel processors …
Retargeting and Respecializing GPU Workloads for Performance Portability
133 天前 - … thread does, and the amount of memory and register … of the generated code against
the baseline GPU compiler (… regardless of the compiler thanks to the shared frontand back-…
the baseline GPU compiler (… regardless of the compiler thanks to the shared frontand back-…