Dataset-Simulation-Shared-VS-Privare L2 cache Per Cluster

This dataset contains the simulation results supporting a study on the deep-learning-based optimization of multicore cache architectures. The evaluated architectural parameters include cache capacity, cache associativity, and the number of cores per cluster. The dataset covers eight benchmarks and reports the corresponding Average Memory Access Time (AMAT), cache miss rate, stall cycles, and Instructions Per Cycle (IPC) for the evaluated architectural configurations. The data were used to train and evaluate a deep learning model and to identify the most appropriate configuration for each benchmark and performance objective. The files are provided under restricted access because the dataset also supports ongoing related research and planned publications. Access may be granted by the corresponding author upon reasonable request for legitimate scientific purposes.