讲者简介:Trevor E. Carlson’s research interests include a number of areas of computer architecture including performance modeling, fast and scalable simulation methodologies, highly-efficient microarchitectures, secure processor design and efficient accelerator design. During his PhD, in collaboration with the Intel ExaScience Lab, he co-developed the Sniper Multi-core Simulator which is being used by hundreds of researchers to evaluate the performance and power-efficiency of next generation systems which continues to be used to explore next-generation processor design. His processor and security works have been selected to appear in top computer architecture, security and design automation conferences (such as ASPLOS, DAC, ICCAD, ISCA, HPCA, MICRO, MICRO Top Picks, and USENIX Security). He is currently working to standardize and help to deploy the Capstone (USENIX Security 2023) work to allow for a new class of flexible, high-performance and trustless memory protection mechanisms. He has recently been awarded Amazon, Intel and VMWare Research Awards, and his work has received six Best Paper Awards or Best Paper Nominations in conferences such as the International Symposium on Microarchitecture (MICRO) and the International Symposium on Performance Analysis of Systems and Software (ISPASS).
报告题目:Building Better Hardware with Fast, Heterogenous Simulation Methodologies
报告摘要:The future of processing is diverse. Modern processing systems are complex, heterogeneous systems, that include both CPUs as well as a diverse set of accelerators including GPUs. Researchers would like to build new systems that can continue to achieve higher performance and efficiency, but this can be difficult when simulating these systems is time consuming and complex. New methodologies are needed to address the complexity of today’s systems. In this talk, we will discuss performance modeling, microarchitecture optimization and sampled simulation of modern hardware systems. The highlight of the talk with be the discussion of two of our new simulator methodologies, Photon and Pac-Sim, that target GPUs and CPUs, respectively, and allow for the fast simulation and workload understanding without the need for upfront workload analysis.