Hi,
I am working with the AudioSample project on the OMAP-L137/C6747 EVM and I am having trouble understanding some observations I have made using the clock profiler on this program. I am using the profiler by placing a breakpoint in the infinite loop within this code in order to judge the number of cycles taken by one execution of this loop.
The first thing I noticed is that using the default number of buffers (four), the number of cycles varies greatly between each execution. While most executions will take maybe 200,000 clock cycles, some will take several million. This is fixed by reducing the number of buffers to 2, but I still do not understand what is causing this change.
The second thing I noticed is that the number of clock cycles this loop takes is unaffected by the sampling frequency I choose. My current understanding is that the clock profiler will tell me the number of clock cycles between two breakpoints, however if this is the case, I would expect each execution of this loop to require the number of clock cycles it takes to sample the data which by my calculations would be ~108.8k clock cycles =(16 * 3*10^8)/44,100. Instead each execution takes ~40k clock cycles, and this value is unchanged when I reduce the sampling frequency to 8k.
I am new to CCS and believe both problem could stem from a simple misunderstanding I have regarding the operation of the clock profiler, but I have been able to find answers to these questions in my research so far. Any information that could improve my understanding would be greatly appreciated.
Thanks,
Alan
