Summary of Concurrency is not Parallelism by Rob Pike

00:00:00 - 00:30:00

Rob Pike discusses concurrency and how it is not parallelism in this YouTube video. He explains that concurrent execution can be achieved by adding a concurrent procedure to an existing design, and introduces the idea of a gopher. He also discusses how goroutines are threads that are dynamically created and multiplexed onto operating system threads.

• 00:00:00 Rob Pike discusses the importance of concurrency and parallelism in programming, and how these are two different concepts that are related. He then introduces the idea of a gopher, a program that helps to eliminate unnecessary paperwork. However, because go is a concurrent language, gophers are not very effective at moving items between piles.
• 00:05:00 The video discusses how concurrent execution can improve the performance of a program. It discusses how concurrent execution can be achieved by adding a concurrent procedure to an existing design.
• 00:10:00 Go routines are like threads, but they run in the same address space and are executed concurrently. This design allows a function to run while another function is executing, similar to the ampersand operator in the shell.
• 00:15:00 The video discusses how concurrent programming is not parallelism, but they are much cheaper and easier to create. Goroutines are threads that are dynamically created and multiplexed onto operating system threads. Select lets you control the behavior of a program by listening to multiple channels at once.
• 00:20:00 Rob Pike discusses concurrency in a language-independent way, explaining how the channel idea in languages like Erlang makes it easy to compose concurrent tasks. He shows an illustrative example of how this works in practice, explaining how a balancer sends work to workers based on their load and how the workers communicate with one another using channels.
• 00:25:00 This video discusses the concept of concurrency, and how it is intrinsic to the design of a pool of workers and a load balancer. The author demonstrates how a simple, correct implementation is possible, by building a heap and dispatching tasks using channels and go routines.
• 00:30:00 This talk by Rob Pike provides an overview of the concept of concurrency, explaining that it is not parallelism. Concurrency is powerful, but it can be easy to parallelize code using it. Pike also provides a history of the development of concurrent programming languages, and a blog post discussing the same topic.