09. Modern CPU Architecture [HPC in Julia]

In this video we will discuss the aspects of modern CPU architecture that are important to know when optimising your code. We will discuss CPU cache as well as the speed of operations on modern CPUs. At the end of the video, we dive into hardware SIMD instructions that can be used to process data in parallel. This module was designed as an MPAGS (Midlands Physics Alliance Graduate School) module and aimed at postgraduates and early career researchers. Timestamps: 00:00:00 Introduction 00:00:10 Performance Difference Example 00:04:16 von Neumann Architecture revisited 00:05:09 Cache 00:10:21 Cache Architecture 00:14:09 Not all CPU operations are created equal 00:15:15 Branch Prediction 00:17:51 Arithmetic Operations 00:20:16 SIMD 00:24:50 SIMD in LLVM instructions 00:26:55 AVX512 on high end processors 00:29:49 Outro Recommended Videos: https://www.youtube.com/watch?v=SAk-6gVkio0 https://www.youtube.com/watch?v=WwkuAqObplU Infographic Blog Post: http://ithare.com/infographics-operation-costs-in-cpu-clock-cycles/ Useful links: Course Information Website: https://jamiemair.github.io/mpags-high-performance-computing/overview/ Julia Documentation: https://docs.julialang.org/en/v1/ MPAGS Registration: https://warwick.ac.uk/fac/sci/physics/mpags/registration/ Course Notes: https://jamiemair.co.uk/courses/hpc