15-418/15-618: Parallel Computer Architecture and Programming, Fall 2021: Schedule

Notes

  • The exact topics of the lectures are subject to change.
  • We do not anticipate changing any of the other dates (exams, assignments, etc.)
  • To watch the lecture videos, sign in to YouTube using your andrew.cmu.edu account.
<
Date Topic Assignment  

Aug 30 Why parallelism? (slides; Lecture 1 videos: Part1 (ignore slides 2 and 3), Part2)
Sep 1 Modern multicore processors (slides; lecture 2 video)
Sep 3 Graphic processing units and CUDA (slides , video (part 1/2), video (part 3)) Assignment 1 out

Sep 8 Parallel programming models (slides; lecture 4 video)
Sep 10 Parallel programming basics (slides , lecture 5 video)

Sep 13 Work distribution and scheduling (slides , lecture 6 video) Assignment 1 due, assignment 2 out
Sep 15 Locality, communication, and contention (slides , lecture 7 video)
Sep 17 Application case studies (slides , lecture 8 video)

Sep 20 Workload-driven performance evaluation (pdf ,video (part 1), video (part 2), video (part 3))
Sep 22 Snooping-based cache coherence (pdf , video (part 1), video (part 2), video (part 3))
Sep 25 Directory-based cache coherence (pdf , video (part 1), video (part 2))

Sep 27 Snooping implementation (pdf , video (part 1), video (part 2)) Assignment 3 out
Sep 29 Performance Measurement and Tuning (pdf, video (part 1), video (part 2)) Assignment 2 due
Oct 1 Memory consistency (slides , lecture 14 video)

Oct 4 Scaling a Website (slides , lecture 15 video (part 1), exam review video (part 2) (NOTE: ignore the discussion of exam logistics - it is outdated)
Oct 6 Exam 1
Oct 8 Interconnection networks (pdf, pptx , lecture 16 video (part 1), lecture 16 video (part 2))

Oct 11 Implementing synchronization (pdf, video (part 1), video (part 2), video (part 3))
Oct 13 Fine-grained synchronization, lock-free programming (pdf, video (part 1), video (part 2), video (part 3)) Assignment 3 due, Assignment 4 out
Oct 15 Transactional memory (pdf, video (part 1), video (part 2), video (part 3))

Oct 18 Heterogenous parallelism (pdf, video (part 1), video (part 2), video (part 3))
Oct 20 Tolerating latency through prefetching (slides , lecture 21 video) Project assignment out
Oct 22 Domain-specific programming languages (slides , lecture 22 video)

Oct 25 Domain-specific programming on graphs (slides , lecture 23 video)
Oct 27 Deep neural networks (slides , lecture 24 video) Assignment 4 due
Oct 29 ML accelerators at Amazon (guest lecture by Randy Huang and Ron Diamant)

Nov 1 Project Idea Meetings
Nov 3 MPI, OpenMP, Cilk implementation (Part A slides, Part B slides, Part A video (MPI), Part B video (OMP, Cilk)) Project proposal due

Nov 17 Exam 2