Goroutines, User Threads, and Kernel Threads: What's the Difference?

Differentiating Goroutines, User Threads, and Kernel Threads

Introduction

The concepts of goroutines, user threads, and kernel threads can be confusing when exploring the intricacies of Go's concurrency model. This article aims to clarify these concepts and address related questions.

Question 1: Defining OS Threads and Goroutines

Question:

• The "Effective Go" introduction mentions goroutines. What does "os threads" refer to? Is it user or kernel threads?

Answer:

• "OS threads" refers to kernel threads.

Question 2: Understanding P and G in the Go Scheduler

Question:

• Why is the number of P (processor/scheduling contexts) equal to the number of CPU cores?
• If all CPUs are servicing Go programs, what happens to other programs in the OS that require CPU allocation?

Answer:

• P serves as a mapping layer between goroutines and kernel threads. There is one P per kernel thread.
• Go programs can allocate all available CPU cores by default (GOMAXPROCS), but this doesn't prevent other OS processes from executing. Even with concurrent operations, processes spend significant time waiting for I/O operations, allowing the kernel scheduler to allocate CPU time to other programs.

Question 3: Number of Kernel Threads Generated

Question:

• How many kernel threads are generated by the OS system?

Answer:

• The number of kernel threads created by the OS is dynamic and depends on the system load.

The above is the detailed content of Goroutines, User Threads, and Kernel Threads: What's the Difference?. For more information, please follow other related articles on the PHP Chinese website!