Tolerating Cache-Miss Latency with Multipass Pipelines

Multipass pipelining uses compile-time scheduling to exploit parallelism and persistent advance execution to achieve memory-latency tolerance, while maintaining the simplicity of an in-order design.