Using age registers for a simple load–store queue filtering

One of the main challenges of modern processor design is the implementation of a scalable and efficient mechanism to detect memory access order violations as a result of out-of-order execution. Traditional age-ordered associative load and store queues are complex, inefficient, and power-hungry. In this paper, we introduce two new LSQ filtering mechanisms with different design tradeoffs, but both explicitly rely on timing information as a primary instrument to rule out dependence violation and enforce memory dependences. Our timing-centric design operates at a fraction of the energy cost of an associative LQ and SQ with no performance degradation.