high performance computing on graphics processing units: hgpu.org

Trace-driven cache simulation is the most widely used method to evaluate different cache structures. Several techniques have been proposed to reduce the simulation time of sequential trace-driven simulation. An obvious way to achieve fast parallel simulation is to simulate the individual independent sets of a cache concurrently on different compute resources. We propose improvements to current trace-driven simulation methods to make them faster and more economical. This paper considers the use of generic GPU to accelerate cache simulation which exploits set-partitioning as the main source of parallelism. We develop more efficient parallel simulation techniques by introducing more knowledge into the compute unified device architecture (CUDA) on the GPU. Our experimental result shows that the new algorithm can produce performance results for a number of set-associative caches in just a single simulation run and still produce the exact results.