Performance Modeling of Computer Systems

This book addresses the gap between performance modeling theory and practice by providing a systematic approach for learning and applying performance modeling relevant principles. It covers essential hardware systems including CPUs, GPUs, accelerators, and memory systems. For each system category, a three-pronged approach is employed: teaching the theoretical background succinctly, presenting modeling methodologies and detailed microarchitectural (cycle-accurate or other) techniques, and providing hands-on implementation guidance using freely available open source simulation frameworks. Furthermore, exercises reinforce learning through combined theoretical analysis and practical implementation.

This comprehensive approach offers multiple advantages. It provides thorough coverage of diverse systems, enabling professionals to build expertise without depending on specific job opportunities for learning. It allows students and new professionals to explore multiple areas before specializing with greater confidence. Finally, it serves as a practical reference for quick refreshers and skill development throughout one's career progression.

What You'll Learn

• Master performance modeling of CPUs, GPUs, accelerators, and memory systems using top open-source tools.
• Refresh core computer architecture concepts essential for accurate modeling.
• Apply best practices to real-world scenarios and start contributing with confidence.
• Build hands-on skills through exercises using widely adopted simulation frameworks.
• Adapt modeling techniques to new hardware and emerging technologies.

Who Is This Book For

Beginner to Intermediate for learning and as a refresher for advanced users. Students and new professionals.

Part I: Foundations. - Chapter 1: Introduction to Performance Modeling. - Chapter 2: Architecture Through the Modeler' s Lens. - Chapter 3: Workload Selection and Experimental Validity


(optional). - Chapter 4: Getting Started with ChampSim. - Part II: The Simulation Pipeline. - Chapter 5: Trace Collection and Management. - Chapter 6: Sampling and Simulation. - Efficiency. - Chapter 7: Understanding Simulation Output. - Part III: Memory System Modeling. - Chapter 8: Memory Hierarchy Architecture for Modelers. - Chapter 9: Cache Modeling in Practice. - Chapter 10: Main Memory and DRAM Modeling. - Chapter 11: Memory Modeling for Multicore Systems. - Part IV: CPU Core Modeling. - Chapter 12: CPU Architecture for Modelers. - Chapter 13: CPU Simulation in Practice. - Chapter 14: Advanced Simulation and Validation. - Part V: GPU Modeling. - Chapter 15: GPU Architecture for Modelers. - Chapter 16: GPU Simulation in Practice. - Part VI: ML Accelerator Modeling. - Chapter 17: The Accelerator Modeling Challenge. - Chapter 18: Accelerator Modeling in Practice. - (optional) Chapter 19: Accelerator Design Space Exploration. - Part VII: Advanced Topics and Practice. - Chapter 20: Analytical and Hybrid Modeling. - Chapter 21: Building a Modeling Practice. -


(optional) Chapter 22: Case Studies.