Design and effectiveness of small-sized decoupled dispatch queues

Continuing demands for high degrees of Instruction Level Parallelism (ILP) require large dispatch queues in modern superscalar microprocessors. However, such large queues are inevitably accompanied by high circuit complexity which correspondingly limits the pipeline clock rates. This is due to the fact that most of today's designs are based upon a centralized dispatch queue which depends on globally broadcasting operations to wake up and select the ready instructions. As an alternative to this conventional design, we propose the design of hierarchically distributed dispatch queues, based on the access/execute decoupled architecture model. Simulation results based on 14 data intensive benchmarks show that our DDQ (Decoupled Dispatch Queues) design achieves performance comparable to a superscalar machine with a large dispatch queue. We also show that our DDQ can be designed with small-sized, distributed dispatch queues which consequently can be implemented with low hardware complexity and high clock rates.