Goals

• Describe steps a CPU performs surrounding actual arithmetic

• Anticipate when performance will be affected by cache size

• Exercises

  • Fit a model to data

  • Read & write models using a probabilistic programming language

  • Estimate uncertainties in model parameters using Markov chain Monte Carlo

• Project

  • Plan data structures to optimize for memory access

  • Consider whether autodifferentiation and/or probablistic program model is appropriate for your project

Readings

• Prior to Monday's class

  • Submit Project Proposal

  • Writing Scientific Software Ch 6: Computer Organization (12pg)

  • Either

    • Introduction to High Performance Computing for Scientists and Engineers Ch 1: Modern Processors (32pg) -or-

    • Introduction to High-Performance Computing Ch 1-1.3.4.1: Single-processor Computing (13pg)

• Prior to Wednesday's class

  • Writing Scientific Software Ch 17.3: Profiling (2pg)

  • Why does Julia work so well?

Project Proposal

Lab

Lab 4: Higher-level Languages & Probabilistic Programming (due Sept 20)

• Exercise 1: Auto-differentiation & Optimization (as in minimization/maximization of a function)

• Exercise 2: Probabilistic Programming Language

Additional Resources

• Monday Q&A: Compilation, Code inspection, Vectorization, Memory Hierarchy