When does the Go scheduler create new M and P?

Just learned the golang GMP model, now I understand how goroutines, operating system threads and golang context/processors cooperate with each other. But I still don't understand when M and P will be generated?

For example, I have a test code to run some operations on the database, and there are two test cases (two batches of goroutines):

func Test_GMP(t *testing.T) {
    for _ = range []struct {
        name string
    }{
        {"first batch"},
        {"second batch"},
    } {
        goroutineSize := 50
        done := make(chan error, goroutineSize)
        for i := 0; i < goroutineSize; i++ {
            go func() {
                // do some databases operations...
                // each goroutine should be blocked here for some time...

                // propogate the result
                done <- nil
            }()
        }

        for i := 0; i < goroutineSize; i++ {
            select {
            case err := <-done:
                assert.NoError(t, err)
            case <-time.After(10 * time.Second):
                t.Fatal("timeout waiting for txFunc goroutine")
            }
        }
        close(done)
    }
}

From my understanding, if M is created when needed. In the first batch of goroutines, 8 (number of virtual cores on my computer) operating system threads will be created, the second batch will only reuse these 8 operating system threads without creating new threads. Is this correct?

We would be grateful if you could provide more material or blogs on this topic.

Correct answer

M can be reused only if your process is not blocked or does not make any system calls. In your case, you have blocking tasks in your go func(). Therefore, the number of M is not limited to 8 (the number of virtual cores on my machine). The first batch will block and remove from P and wait for the blocking process to complete while a new M is created associated with P.

See below for more details,

• //m.sbmmt.com/link/a4a1108bbcc329a70efa93d7bf060914
• go-goroutine-OS thread and CPU management

The above is the detailed content of When does the Go scheduler create new M and P?. For more information, please follow other related articles on the PHP Chinese website!