ESP8266 Community Forum

- Sat Oct 04, 2014 6:38 pm #1191 OK! I finally did a test with this.

I wrote the following function. The comment is my tested performance:

Code: Select all//Through testing, I've found that these cause the pin to be at that state for time = 1/10th of a us. (plus 1/15?th of a us for the call)
static void PTON( uint8_t time )
{
   while(time--)
   {
      WRITE_PERI_REG( PERIPHS_GPIO_BASEADDR + GPIO_ID_PIN(PTGPIO), 1 );
   }
}

So, looks like you can get up to around .1us maximum performance.

If you use the GPIO API, the performance is TERRIBLE. Do not use it for anything that needs speed.

- Sat Oct 04, 2014 6:53 pm #1192 What variations do you see? Any larger gaps when wifi data is processed?

- Sun Oct 05, 2014 1:46 am #1209 Yes, it was all over the place unless I disabled interrupts.

- Sun Oct 05, 2014 2:13 am #1211 Does disabling interrupts kill active connections?
For what time do you think it's save to disable interrupts?