Information about the final project, followed by a guest lecture covering fractal trees, which combine the strengths of B-trees and append-to-file.  Analysis of search and insert in a simplified fractal tree.

Lecture 19: How TokuDB Fractal Tree Indexes Work

Students vote on which ray-tracer images are accurate enough to be included in the competition, and these entries are tested against each other for speed.  Students also provide feedback on the course.

Lecture 23: Final Competition and Student Feedback

Meeting for 6.172 industry mentors.  Description of mentorship role, expectations, overview of course and how it fits into EECS curriculum.

Industry mentor (MITPOSSE) overview

Primer on ray tracing techniques, given to prepare students for the final project.  Includes some ray tracing background and a code overview covering classes and high-level execution.

Lecture 18: Primer on Ray Tracing Techniques

Lecture covering distributed systems at the cluster, data center, and planet scales.  Topics include message passing, examples of the need to scale, Google's programming model, and cloud computing.

Lecture 20: Distributed Systems

Lecture covering nondeterministic programming, including mutual exclusion, implementation of mutexes, and locking anomalies.

Lecture 15: Nondeterministic Programming

Lecture covering analysis of multithreaded algorithms, including divide-and-conquer recurrences, loop parallelism in Cilk++, and matrix multiplication and merge sort examples.

Lecture 14: Analysis of Multithreaded Algorithms

The first part of the lecture covers parallelism analysis, caches, and synchronization correctness.  The second focuses on compiler optimization questions: is the optimization legal, faster, and automatic?

Lecture 21: Quiz 2 Review

Guest lecture by Ravi Soundararajan of VMware, covering ten case studies in performance engineering.

Lecture 22: A Tale of 10 Bugs: Performance Engineering at Vmware

Discussion of project 3 beta.  Lecture covering multicore programming, including shared-memory hardware, concurrency platforms, and race conditions.

Lecture 12: Multicore Programming

Performance Engineering of Software Systems

Lecture covering the impact of synchronization and memory on parallel performance, using OpenMP instead of Cilk.  Topics include granularity of parallelism, true and false sharing, and load balancing.

Lecture 17: Performance Issues in Parallelization

Lecture covering dynamic storage allocation, including reference counting, a graph abstraction, and updating pointers.

Lecture 10: Dynamic Storage Allocation

Lecture covering synchronizing without locks, including memory consistency, lock-free protocols, the ABA problem, and reducer hyperobjects.

Lecture 16: Synchronizing without Locks

Lecture covering cache-efficient algorithms, with tiled and recursive matrix multiplication examples.

Lecture 8: Cache-Efficient Algorithms

Lecture covering parallelism, scheduling theory, the Cilk++ runtime system, and Cilk chess programs.

Lecture 13: Parallelism and Performance

Basic performance engineering.  Bentley's rules (modifying data, modifying code) and the traveling salesman problem.

Lecture 3: Basic Performance Engineering

Lecture covering memory systems, including cache concepts, access pattern concepts, data structure transformations, and computation transformations.

Lecture 7: Memory Systems and Performance Engineering

Lecture covering compiler hacks, when to optimize, data-flow analysis and optimizations, and instruction scheduling.  Discussion of quiz, including common mistakes.

Lecture 11: What Compilers Can and Cannot Do

Lecture covering bit hacks.  Examples include swap, population count, and the n-queens problem, used to highlight backtracking search and bitvector representation.

Lecture 2: Bit Hacks

Discussion of project 2.2 beta, how pair programming can lead to less time debugging, and importance of working well in groups.  Lecture covering more cache-efficient algorithms, including a heat diffusion simulation.

Lecture 9: Cache-Efficient Algorithms II

Really good course. Thanks for sharing the same for the benefits of all. It would be really great if slides and class room exercises can be shared.

Good course

Good to know that there exist courses that focuses on software performance. The videos are sometime blurry, but mostly good. I hope we can get a copy of the slides used in this class.

Nice course that focuses on software performance.

Overview of computer architecture, profiling a program, and a set of example programs.  Student performance on project 1 is also discussed.

Lecture 4: Computer Architecture and Performance Engineering

The best podcast episodes on parenting. Interviewing experts and talking through common issues surrounding parenting. This topic is filled with important lessons and insights for anyone listening. Add episodes you're interested into your queue below!

OwlTail

Best Parenting Podcasts

The Best Multi-Part Documentaries, with documentaries on Obama, Federer vs Nadal, the making of Star Wars, Netflix vs Blockbuster and more. Add episodes you're interested into your queue below!