LabVIEW Idea Exchange

cancel
Showing results for 
Search instead for 
Did you mean: 
altenbach

Installer Builder option to require SSE2 support

Status: New

In newer LabVIEW versions, we can build executables that require SSE2 support an this option is even the default.

 

If we built an installer, we can select many other system requirements, such as the minimal OS version or the presence of the LabVIEW development system.  However, there is no option to prevent installation if the CPU does not support SSE2.

 

This means that an application that requires SSE2 support will install just fine on a non-SSE2 system and will fail only once we try to actually run.

 

I suggest to add a system requirement option to the installer builder so the application will not install if the hardware is not suitable. It should maybe even be enabled by default.

 

Here's how it could look like.

 

 

3 Comments
Mads
Active Participant

Or - make it possible to build and ship two versions automatically, with and without the SSE2 requirement, and make the installer figure out which one to install...That way we can almost forget the whole thing, and just remove the non-SSE2-version if we know that the application will in fact need it to perform adequately.

Yamaeda
Proven Zealot

I agree in principle. However i dont like having an extra option which'd allow a non SSE program to require it from the CPU or the opposite. If SSE is enabled in the compile it should be required on the target.

 

/Y

G# - Award winning reference based OOP for LV, for free! - Qestit VIPM GitHub

Qestit Systems
Certified-LabVIEW-Developer
Mads
Active Participant

If the application is set to be optimized with SSE2, then the installer should refuse to install on CPUs without such support - period. No need for any new input in the app builder.

 If a feature like the one I suggested above was implemented (to allow us to compile with AND without SSE2, and let the installer decide which one to install), then a choice to require it would not make sense either.

 

So in a way I agree with you Yamaeda, but the core of the idea is still good.