Cortex-A8 Forced Memory Caching

You need a process of a large array of numbers (~ 1 Mb) in real time using a function, for example. void processData(char* data).

The following test was launched on the target platform:

int j = 10;
while(j--)
    processData(dataPtr);

with the same data every time. He showed the following result:

• 1st launch takes ~ 22.5 m.
• 2nd run, and others ~ 12.5 ms

In my opinion, this may be due to the fact that on the second start, the data is already in the processor’s cache, so it works much faster.

Problem: in the real case, the data will be different each time.

Is there a way to do some sort of “preloading” the data into the cache?