1 const int buttonPin = 2;
 2 const int ledPin = 13;
 3 
 4 int ledState = HIGH;
 5 int buttonState;
 6 int lastButtonState = LOW;
 7 
 8 unsigned long lastDebounceTime = 0;
 9 unsigned long debounceDelay = 50;
10 
11 void setup() {
12   pinMode(buttonPin, INPUT);
13   pinMode(ledPin, OUTPUT);
14   digitalWrite(ledPin, ledState);
15 }
16 
17 void loop() {
18   int reading = digitalRead(buttonPin);
19   if (reading != lastButtonState) {
20     lastDebounceTime = millis();
21   }
22   if ((millis() - lastDebounceTime) > debounceDelay) {
23     if (reading != buttonState) {
24       buttonState = reading;
25       if (buttonState == HIGH) {
26         ledState = !ledState;
27       }
28     }
29   }
30   digitalWrite(ledPin, ledState);
31   lastButtonState = reading;
32 }
 

原文的程式碼 暫時去掉註釋，便於顯示重點。

編譯結果：

Sketch uses 1116 bytes (3%) of program storage space. Maximum is 30720 bytes.
Global variables use 19 bytes (0%) of dynamic memory, leaving 2029 bytes for local variables. Maximum is 2048 bytes.

優化手段1：選用夠用的資料型別，這裡面的int物件，都是小於255的，可以換成byte：

const byte buttonPin = 2;
const byte ledPin    = 13;
byte ledState = HIGH;
byte buttonState;
byte lastButtonState = LOW;

unsigned long lastDebounceTime = 0;
unsigned long debounceDelay    = 50;

void setup() {
  pinMode(buttonPin, INPUT);
  pinMode(ledPin,    OUTPUT);
  digitalWrite(ledPin, ledState);
}

void loop() {
  byte reading = digitalRead(buttonPin);
  if (reading != lastButtonState) {
    lastDebounceTime = millis();
  }
  if ((millis() - lastDebounceTime) > debounceDelay) {
    if (reading != buttonState) {
      buttonState = reading;
      if (buttonState == HIGH) {
        ledState = !ledState;
      }
    }
  }
  digitalWrite(ledPin, ledState);
  lastButtonState = reading;
}

編譯結果：

Sketch uses 1074 bytes (3%) of program storage space. Maximum is 30720 bytes.
Global variables use 17 bytes (0%) of dynamic memory, leaving 2031 bytes for local variables. Maximum is 2048 bytes.

效果：

程式空間減少42位元組；SRAM減少2位元組。

繼續淦：

裡面有一個函式在相近的位置，用了兩遍，且貌似挺複雜的，就是：millis();

將其合二為一，但是又要增加一個全域性變數，試試效果：

const byte buttonPin = 2;
const byte ledPin    = 13;
byte ledState = HIGH;
byte buttonState;
byte lastButtonState = LOW;
unsigned long lastDebounceTime = 0;
unsigned long debounceDelay    = 50;
unsigned long ml = 0;

void setup() {
  pinMode(buttonPin, INPUT);
  pinMode(ledPin,    OUTPUT);
  digitalWrite(ledPin, ledState);
}
void loop() {
  byte reading = digitalRead(buttonPin);
  ml = millis();
  if (reading != lastButtonState) {
    lastDebounceTime = ml;
  }
  if ((ml - lastDebounceTime) > debounceDelay) {
    if (reading != buttonState) {
      buttonState = reading;
      if (buttonState == HIGH) {
        ledState = !ledState;
      }
    }
  }
  digitalWrite(ledPin, ledState);
  lastButtonState = reading;
}

　　編譯結果：

Sketch uses 1052 bytes (3%) of program storage space. Maximum is 30720 bytes.
Global variables use 17 bytes (0%) of dynamic memory, leaving 2031 bytes for local variables. Maximum is 2048 bytes.

又幹掉22位元組的程式空間，SRAM竟然沒變？

(1166-1052)/1166 = 104/1166 = 9.8%

將近10%的優化成果，這還沒使用骨灰級的暫存器大法來優化pinMode和digitalRead/Write。

後續再試：constbytedebounceDelay=50; 結果還是1052、17，看來是編譯器自動將其優化了。