Fade Observation

First question:
I noticed the fade program doesn't use Serial.begin(9600). Why? I'm new and thought that affects how fast the Arduino operates?

Second question:
I noticed an interesting observation with the fade program. I removed the delay() function at the bottom of the original program and adjusted the fadeAmount.

When fadeAmount = 5, there is a most noticeable and slowest blink compared to other fadeAmounts (I tried 1,3, 15, etc.) for which the blink is fast and looks nearly identical with each setting other than 5.

The blink observation gives weird results without the delay() function, but agrees with the code when the delay() function is used.

My question is why it gives weird behavior without the delay() function.
Can you explain this behavior to me? Because I thought with fadeAmount = 1 the blink would be slowest and thus most noticeable compared to fadeAmount = 5. But that is not the case. I tried pins 11, 10, 9, 6, 5, and 3 and they all give similar results.

int led = 9;
int brightness = 0;
int fadeAmount = 5;


void setup() {
  pinMode(led, OUTPUT);
}


void loop() {

  analogWrite(led, brightness);

  brightness = brightness + fadeAmount;

  if (brightness == 0 || brightness == 255) {
    fadeAmount = -fadeAmount ;
  }

  // Removed delay(30);
}

Serial is used for communication. It doesn't change the speed of the micro controller. The number 9600 is the baud rate, or the speed of communication. Not all programs need to send or receive data from the computer while running.

PWM used in analogWrite depends on microcontroller's internal timekeeping. When you remove the delay you very quickly change the duty cycle of the PWM. However, the instruction to change this duty cycle takes some time. If during that time comes the interrupt to turn on or off the pin, the microcontroller does that and goes back to the task of changing the PWM. Without the delay you don't give it enough time to settle on one value before asking it do do something else.
Why is fadeAmount 5 so special? It appears it just happens to be inside some "resonant frequency" (I use that term loosely here) range where you perceive the effect the strongest.

Thank you for responding.

I think that answers the questions.

However, the instruction to change this duty cycle takes some time.

True

If during that time comes the interrupt to turn on or off the pin,

There is no interrupt to turn on or off the pin; PWM pins are driven by timer hardware.

Fair enough.
What about the rest? Would you say that my explanation to the observed behavior is close enough?