Optimize total programm run time

[freemin7]

Assuming i have a particular program with particular inputs with a total run-time want to optimize

./executeable work_file

Do i really need to recompile the executable and add a progress point before the successful exit and loop the executetable using something like bash

coz run --- bash -c "for i in {1..10000}; do ./executeable work_file; done;" 

Or is there a more elegant way?

[emeryberger]

You could of course also have the code execute a loop comprising the whole program, but if you don't want to do that, what you propose would work. Ideally you'd have some individual thing that you really want to have happen more frequently (so you'd place a throughput progress point) or a begin-end pair which you'd like to take less time (which would call for latency progress points).

[freemin7]

If i want to optimize total program run-time, would i need evenly distributed progress points throughout the program to ensure drawing the correct conclusions?

[emeryberger]

Let's move this to discussions.

[freemin7]

Where do these happen? Nevermind. Found them.