When and Why Does the Go Scheduler Allocate New M and P Processors?

When and Why the Go Scheduler Allocates New M and P Processors

Understanding the Go runtime's scheduling model is crucial for optimizing application performance. One key aspect of this model is the allocation of M (machines) and P (processors).

When Are M and P Created?

M processors are created when:

• A new goroutine is started ( unless there is a spare idle M)
• An existing M blocks on an A system call (creates a new operating system thread known as an M)

P processors are created when all the local runqueues are full one goroutine is running its neighbor goroutines ready to run are put in a global queue and it contains a single p.

Blocking Tasks and P Reuse

In the provided test code, the goroutines perform blocking database operations. Blocking tasks remove M processors from P processors.

In this case, creating new M processors is necessary to handle the blocked goroutines. The initial 8 (number of virtual cores) M processors will not be sufficient for the second batch. New M processors will be allocated as needed.

Additional Resources

• https://www.programmersought.com/article/79557885527/
• https://blog.golang.org/go-goroutine-os-thread-and-cpu-management

The above is the detailed content of When and Why Does the Go Scheduler Allocate New M and P Processors?. For more information, please follow other related articles on the PHP Chinese website!