Add lane masks bit-shift example, update to CUDA 12.9.1 and HIP SDK 7.1

[j-stephan]

Motivation

Adds more meaningful examples to illustrate the HIP documentation's lane-masks bit-shift section. This PR requires an update to CUDA 12.9.1.

Technical Details

Adds example code, integration into CMake and Makefile build system, .gitignore, README.md and Visual Studio files. Updates the top-level files where necessary.

Test Plan

Code tested locally on a RDNA GPU.

Test Result

Works as expected.

Added/Updated documentation?

• Yes
• No, does not apply to this PR.

Included Visual Studio files?

• Yes
• No, does not apply to this PR.

Submission Checklist

• New examples contain proper CMake and Make files.
• I have updated relevant CI workflows and the new examples are being built and tested in CI.
• Check and guard against unsupported ASICs if relevant dependent libraries have limited support.
• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

[j-stephan]

@randyh62 @neon60 Apart from CUDA CI stuff, this is complete and ready for review.

[j-stephan]

The CUDA issue observed here is fixed by #394 and its upgrade to CUDA 12.8.

[zichguan-amd]

HIP SDK 7.1.1 is out, I wonder if it's a compatibility issue with HIP SDK 6.4 causing the CI to fail

[j-stephan]

Yes, looks like CUB removed that header file with CUDA 12.8. Time to migrate to a newer HIP SDK :-)