I am working on an embedded system using STM32 microcontrollers, and I want to optimize GPIO writes for performance, power efficiency, and reliability. Specifically, I want to avoid unnecessary GPIO writes when the pin state is already as desired.
Here is my current approach:
if (temperature > 15) {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_SET);
} else {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_RESET);
}
This code will write to the pin even if its state hasn't changed, which could waste resources and power. To address this, I check the current pin state before writing:
if (temperature > 15 && HAL_GPIO_ReadPin(GPIOB, GPIO_PIN_0) == GPIO_PIN_RESET) {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_SET);
} else if (temperature <= 15 && HAL_GPIO_ReadPin(GPIOB, GPIO_PIN_0) == GPIO_PIN_SET) {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_RESET);
}
In this second method, I check the current pin state before writing, ensuring that the write operation only occurs if the state needs to be changed.
However, I considered that reading the GPIO pin might introduce some delay, so I defined a variable to store the pin state. Now, instead of reading the pin state, I read the variable:
bool pinState = false; // false: LOW, true: HIGH
if (temperature > 15 && pinState == false) {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_SET);
pinState = true;
} else if (temperature <= 15 && pinState == true) {
HAL_GPIO_WritePin(GPIOB, GPIO_PIN_0, GPIO_PIN_RESET);
pinState = false;
}
In this third method, I store the pin state in a variable, avoiding reading the hardware state every time.
Now I have some doubts about which approach is best:
In the first method, I write to the pin every time (i.e., constant writes).
In the second method, I check the hardware state before writing.
In the third method, I store the pin state in a variable and avoid reading the hardware state.
Which of these three methods would be the most efficient in terms of performance, power efficiency, and reliability? What is the best approach?
and I want to optimize GPIO writes for performance, power efficiency, and reliability.
and HAL_GPIO_...
are contraversions.
Which of these three methods would be the most efficient in terms of performance, power efficiency, and reliability? What is the best approach?
There isn’t a single "best approach" — it depends entirely on the specific context and requirements.
If you use direct register access like GPIO->BSRR = PIN_MASK << (reset * 16) instead of calling the HAL_ function, you can save dozens of clock cycles (or more). Additionally, the overhead of your if
statement will likely be higher than a single write to the peripheral. The STM32 GPIO hardware is quite efficient and won't make unnecessary changes if the requested state is already set.
Know you hardware. If you want to go low on power you need to know your hardware very well. No magic tricks. It takes time and effort.
In my opinion, rather than focusing on micro-optimizations, it's more important to prioritize key aspects like entering sleep mode quickly and staying in low-power mode as long as possible. Also, consider going bare metal. While the HAL functions simplify development, they come with performance and energy costs, so be mindful of when to use them.