We use MATLAB as the implementation language for functional prototyping of both hardware and software. In the Profiler data from execution of a scientific application, there are entries for:
- num2str
- strcat
- int2str
- blanks
The above routines are called for each reference to other routines we wrote, in For loops. The problem: these routines consume a significant amount of Total time. Also, Self time for the individual routines is significant.
What is the function of these routines? Can you recommend options for reducing their impact on application performance? Thanks for your assistance
Best Answer
num2str() and int2str() should be replaced with sprintf() unless you need the behavior of automatically formatting arrays.
I am not sure there is anything faster than blanks. Possibly the obscure char.empty()