Hi RandyP:
In the thread ----
http://e2e.ti.com/support/dsp/tms320c6000_high_performance_dsps/f/112/p/265550/929354.aspx#929354
I followed your advice and I searched "C6000 Training" keyword on
http://processors.wiki.ti.com.
So I found the useful workshop -
TMS320C6000 DSP Optimization Workshop.In the workshop There was
one pdf file --- op6000_student_guide_v1.51.pdf.In addition I saw
the page like below and I couldn't agree more!
Question1: why spend time optimizing if real-time performance has
alreadybeen met with plain C code.
In my opinion,Of course you couldn't spend time optimizing if
performance has beenmet.And its very easy to test whether the
performance of your code met your need using profiler or
CLK_getltime or clock.Before you start to optimize your C code you
must to prove that .its very important.
But as to Question2 -Why spend time optimizing if the algorithm
cannot be improved.I didn't know how to test that.In another word I
didn't know when to stop optimizing.I didn't know what was the optimum!
I didn't know what was the best performance of the algorithm.
I just know I had to start to optimize my poor performance algorithm
because the performance couldn't meet my need.But I didn't know when
to stop optimizing.
That was a rather terrible thing.Only knowing when to start But not
knowing when to stop.It was unbelievable.What can I do? Help me please!