Small Compatibility Fix for CUDA 12.8.

[Schmiedium]

The latest version of CUDA ships libnvvm with version 2.0, which fails the check for an appropriate version number. This commit fixes this so that 2.0 correctly passes the minimum version number check.

#163 Referencing this issue, this PR should provide a fix for that when using CUDA 12.8

[LegNeato]

I don't think this is right, 1.5 will incorrectly continue / not take the error path. (5 < 6) && (1 < 1) -> (true && false) -> false

[LegNeato]

You probably want something like:

    if major <= 1 && minor < 6

[LegNeato]

As mentioned in #100, we are unsure of what the implications here are. I'd be inclined to enable and deal with the fallout though.

[Schmiedium]

You're right, I messed up the logic there. Thank you for catching that!

I did some digging and found here nvvm version 2.0 is supposed to be a breaking change. I have CUDA 12.8 installed though, and don't see any issues with the examples we have built.

Everything seems to compile just the same and run fine. If you're okay with the fallout from trying to support newer versions, then I think it would be a good thing to do. I think people trying to use the project in the future will install the newest version of CUDA available, and if this prevents that it could be a headache.

[LegNeato]

Right, if you look at #100 I pasted that link in 👍 . It doesn't look the libnvvm API was breaking, and I don't think there is any case where we generate multiple versions concurrently so it is probably fine.

[LegNeato]

Thanks!!! 🚀