Concurrency and Parallelism From The Ground Up

on YouTube ↗29m

" summary

Covers how CPUs, interrupts and OS schedulers create 15–60 ms time slice quanta and why ticks (Windows 15.6 ms, Linux ~4 ms) matter for context switches; shows cooperative vs preemptive multitasking, virtual memory and cache-coherency effects on races; demonstrates how atomic instructions and locks (and their fairness limits) try to fix races; states global interpreter lock prevents concurrent Python bytecode execution and argues IOCP and epoll async I/O scales for thousands of connections while multiprocessing suits CPU-bound work.

AI-written summary · about

" chapters (22)

1. 0:00Amber Brown Concurrency and Parallelism From The Ground Up
2. 0:37Implementing Concurrency & Parallelism From The Ground Up
3. 2:17'concurrency' is not recognized as an internal or external command, operable program or batch file.
4. 2:59How do programs run?
5. 4:531985 (Windows 1.9)
6. 5:27Cooperative Multitasking
7. 8:501994 (Windows 95)
8. 9:26Pre-emptive Multitasking Concurrency through Distrust
9. 9:58Timeslices & Ticks Distributing time
10. 11:21Timeslice Tradeoff Expensive & Responsive VS Performant & Laggy
11. 12:01CPU registers Virtual Memory table Program Counter
12. 12:45Threads & Processes More than just a choice in gunicorn
13. 14:422001 (Windows XP)
14. 14:53Simultaneous Multithreading 1 core, 2 threads
15. 15:29Symmetric Multiprocessing oh h*ck
16. 16:11Synchronisation Locks, Semaphores, and Mutexes, oh my!
17. 18:47Thread Safety "Safe" as in space travel
18. 20:11Death of the Tick "Tickless" kernels
19. 20:52Coprocessing GPGPU & Friends
20. 22:29Cooperative Python Single-threaded Fun
21. 22:57Kernel-based Networking I/O Completion Ports epoll kqueue
22. 25:21Decreasing Overhead No inter-thread communication

" related talks

• 48m“Is Parallel Programming still hard? PART 2 of 2”P. McKenney, M. Michael, M. Wong · CppCon 2017
• 39mLet's talk concurrencyJosé Valim · Rocky Mountain Ruby 2012
• 59m“Is Parallel Programming still hard? PART 1 of 2”P. McKenney, M. Michael, M. Wong · CppCon 2017
• 59mAn Introduction to Multithreading in C++20Anthony Williams · C++ on Sea 2022
• 37mConcurrency vs ParallelismNarendran R · PyCon APAC 2018
• 51mFrom Linux Threads To ForkJoinPoolAndrey Ershov · JavaDay Ukraine 2017
• 36mConcurrency and Parallelism with Python for HumansSantiago Basulto · PyCon Latam 2019
• 25mLet me sleep on it improving concurrency in unexpected waysNir Soffer · PyCon Israel 2017