Exascale Computing: From Shared Memory to GPU Architecture

Explore the cutting-edge world of exascale computing in this comprehensive lecture delivered by Dr. Meifeng Lin from Brookhaven National Laboratory. Delve into the intricacies of US exascale computing initiatives and gain insights into quantum computing advancements. Examine various high-performance computing (HPC) architectures, focusing on shared memory systems and the OpenMP programming model. Compare shared and distributed memory models, understanding their unique characteristics and applications. Discover the power of GPU computing, analyzing its architecture, peak performance, and growing popularity in comparison to traditional CPUs. Learn about GPU programming techniques and explore OpenCL as a framework for heterogeneous computing. This in-depth presentation covers essential topics in advanced computing, providing a solid foundation for understanding the future of computational power and its applications in scientific research and technological innovation.

Introduction
Exascale Computing
US Exascale Computing
Quantum Computing
HPC architectures
Shared memory architectures
OpenMP
OpenMP Parallelization
Shared vs Distributed Models
Shared vs Distributed Memory
Shared Memory Programming Model
GPU Computing
GPU Architecture
Peak Performance
CPU vs GPU
GPU popularity
GPU programming
OpenCL