CUDA behaviour differences between GTX 1070 laptop and desktop

I need to make optimisations for several computational algorithms for desktop GPUs GTX 1070. I have GTX 1070 installed in my laptop ( ) .
The main question is: if I do optimisation for GTX 1070 laptop version would GTX 1070 desktop behaviour be different or not? Say, desktop edition will work slower after optimisation and laptop faster or otherwise.
As I got the only difference between these GPUs is clock freq. Memory onboard and CUDA cores are the same. So, is it correct that all optimisations made for laptop edition of 1070 take same effect for desktop edition of 1070?
I don’t want to buy desktop pc only for this task.

Depending on how much the clocks differ between the two GTX 1070s, it is at least theoretically possible that a kernel would be compute bound on of the GPUs but memory bound on the other.

I would consider the chances of that happening reasonably slim, so to first order you should be fine with developing on your laptop while deploying on your desktop. What are the clocks (core and memory) on the desktop GTX 1070 vs the laptop GTX 1070?

Looks like I was wrong. Number of cuda cores differs.
Here are the specs for both from Nvidia:



I guess more cores is not a problem cause I can use the same number of cores as for desktop edition. But clocks…