[摘要]:Memory-intensive operations and their memory access latency are often the performance bottleneck in parallel applications. In this paper, we investigate the concept of active memory operation which is an active data processing operation performed on the m |