Epsim: A Scalable and Parallel Marssx86 Simulator With Exploiting Epoch-Based Execution

Abstract

In general, a detailed modeling and evaluation of computer architectures make a cycleaccurate simulator necessary. As the architectures become increasingly complex for parallel, cloud, and neural computing, nowadays, the complexity of the simulator grows rapidly, and thus its execution is too slow or infeasible for practical use. In order to alleviate the problem, many previous studies have focused on reducing the simulation time in a variety of ways such as using sampling methods, adding hardware accelerators, and so on. In this paper, we propose a new parallel simulation framework, called Epoch-based Parallel SIMulator, to obtain scalable speedup with large number of cores. The framework is based on a well-known cycle-accurate full-system simulator, MARSSx86. From the simulator, we build an epoch, that is an execution interval, where the architectural simulation by PTLSim does not involve any interaction with QEMU. Therefore, we can simulate epochs independently, i.e., execute multiple epochs completely in parallel by PTLSim with their live-in data. Our performance evaluation shows that we achieve 12.8× speed on average with 16-core parallel simulation from the SPEC CPU2006 benchmarks and the PARSEC benchmarks, providing the performance scalability.

EPSim-C: A Parallel Epoch-Based Cycle-Accurate Microarchitecture Simulator Using Cloud Computing

“…Further details can be found in Kim et al [33]. Figure 5 presents the speedup achieved by our existing epoch-based parallel microarchitecture simulator, called EPSim, using a multicore platform on the SPEC CPU2006 benchmarks [33]. EPSim obtained an average speedup of 1.0×, 3.7×, 6.9×, and 13.0× using 1, 4, 8, and 16 cores, respectively, in a multicore machine.…”

“…As mentioned in Section 2.2, the interaction related overhead takes up a considerable portion of the total simulation time. In order to accelerate the MARSSx86 simulator by eliminating the causal interactions, EPSim [33] was developed. EPSim is an epoch-based cycle-accurate parallel MARSSx86 simulator.…”

“…As a result, the epoch-based parallel simulator has large-scale parallelism because epochs can be simulated independently in a completely parallel manner. Further details can be found in Kim et al [33]. Figure 5 presents the speedup achieved by our existing epoch-based parallel microarchitecture simulator, called EPSim, using a multicore platform on the SPEC CPU2006 benchmarks [33].…”

“…Some threads perform functional modeling, and others play the cycle-accurate simulation. EPSim [33], which is our base model, was designed to perform epoch-based parallel microarchitecture simulation that produces deterministic results. Each epoch is identified based on the interactions between cycle-accurate timing simulations and functional emulations such that an epoch can be simulated independently of other epochs in parallel, meaning the epochs can exploit large-scale parallelism.…”

“…In this paper, we propose an EPSim-C (EPSim on Cloud) simulator that extends EPSim [33], an epoch-based parallel microarchitecture simulator, to run in a cloud computing environment. Therefore, we can dramatically improve the performance of the simulation by exploiting the intrinsically massive parallelism of epochs defined in EPSim.…”