TDA4AL-Q1: A72 Frequency Scaling

 Use k3conf to adjust the frequencies of clock 202 and clock 203 and conduct the following tests:

1. Utilize a while loop to reduce the idle time of A72 to zero, and record the stable temperature after a period of time.
2. In an idle state, run the following program and measure the time in seconds.
   

   #!/ti_fs/bin/ksh
   
   i=0
   while [ $i -ne 1000000 ]
   do
       i=$(($i+1))
   done
       echo "End"

We got the following result:

• Settings of **500M, 1G, 1.5G, 1.6G, 2.9G, 3.0G, and 3.2G** will cause the system to crash.
• 1.9G, 2.2G, 2.5G, 2.8G, and 3.1G are invalid configuration values.
• From 200M to 900M, there is a state of high computational power and high temperature.
• From 1.4G to 2.7G, computational power and frequency show a positive correlation trend.
• 1.2G and 1.3G are in a state of high computational power and high temperature.
• At 1.1G, the computational power is the lowest and the temperature is the lowest.
• At 2.7G, the computational power is the highest and the temperature is the highest.

The experimental results do not match the conventional understanding of the correlation between CPU frequency and computing power. What could be the reason?