[Razor12911]

I made it to work like that for a reason. Remember you said your intel you was not detected. So after dxgi is done detecting gpus using its methods, if any gpus were missed then registry code adds the remaining gpus. This is why I didn’t use “else”.
But then you might say why not just get rid of dxgi if doesn’t detect all gpus.

Here is why, registry skips gpus with names that have already been detected which means if SLi/Crossfire was used, registry method will not be able to detect this but dxgi will be able to.
While dxgi has issues of detecting intel gpu, registry helps dxgi
Plus we have no idea if the same registry idea works on older operating systems so it’s best to have both methods working together

So you can see why both code have to exist because they nullify the disadvantages they both have, furthermore. Dxgi has issues detecting gpu size in x86 because the value type used in x86 has the same issue as with wmi adapterram where gpu over 4gb is misread.

Tl;dr both code help mitigate their own disadvantages so the code works flawlessly