Original function is here

#define STRESS_ARRAY_SIZE (128*1024)
static void* _stress_function(__attribute__((unused)) void* arg)
{
    uint8_t * _stress_array  = malloc(sizeof(uint8_t)*STRESS_ARRAY_SIZE);
    
    uint32_t cntr = 0;
	while(stress_running){
	    cntr++;
		volatile double s = sqrt(rand()); //this helps slow down the while loop
		
		const int do_memcpy_every_nth = 50; //do not perform memcpy every loop because cpu can get hot very quickly
		if ((cntr % do_memcpy_every_nth) == 0){            
		    memcpy(_stress_array,_stress_array+STRESS_ARRAY_SIZE/2,sizeof(uint8_t)*STRESS_ARRAY_SIZE/2);
		}
	}
	//printf("stress thread exiting\n");
	
	free(_stress_array);
	return NULL;
}

It tested this myself and have a small tip for you: setting cpu mode to performance helps to burn more power in the cpu (run this right before starting the temp calibration):

voxl-set-cpu-mode perf

Also you should start the calibration when the board is a bit colder, maybe room temp 20-25C

Try it!

the source code is here

You can run in debug mode to get more info, but it seems the temperature of the IMU has not really changed from the start (34C vs 31C). The cpu stressing may not be enough. You should check output of top while the stressing of CPU is running (the application just burns cpu cycles in order to heat up the cpu, which will heat up the IMU to help with temperature calibration). How long did the procedure run before it stopped?