Why is 2014b faster?

I am running simulink via matlab scripts on data files. I am finding that 2017b runs them much slower than that of 2014b. 2014b takes 3-4min, while 2017b takes 25min for one file. I am not sure why this is or what can be done to fix it.

Any thoughts on what to look at or be concerned with would be greatly appreciated.

Answer 1

[Rik]

This question pops up from time to time.

Every now and then, some functions are redesigned. Of course not all functions would be re-designed every release, but I would imagine that some functions get an overhaul more often than others. This is done to improve the speed, and/or precision, and/or flexibility of the function. For some specific use cases this will result in much slower code, even if on average everybody's code performance might improve.

Sometimes a lot of things get a major overhaul. The last time this happened very prominently was with R2014b, where the internal design of graphics elements was changed. This resulted in a wealth of new possibilities, and a lot of old code breaking (although that would in general only happen for undocumented uses).

The only real answer is: enjoy the release that works fastest, and use the debugging tool and the profiler to get a better look at the bottlenecks.

If you want help on a specific situation, you should post more details.

[Chase Ward]

Ok thanks for your advice. I have called into support and they basically explained the same thing to me. Go in and use the upgrade advisor tool. Hopefully after doing a run through with that I will see an improvment. Looks like have a bit of learning to do!