Add Ascend NPU device support.

[Ginray]

Summary

This PR is the first step in the adaptation of Ascend NPU to Liger Kernel: adding NPU device support. For details, refer to [RFC] Native Ascend NPU Support for Liger Kernel, Section 2.1: Device Support Integration.

Details

Key Modifications:

1. Add the installation method and basic function adaptation for NPU.
2. Directly import via triton.language.math on NPU to avoid errors caused by non-existent interfaces.

Testing Done

Verification Status:
We have conducted verification on Atlas 800T A3, and basic test cases such as test_softmax and test_swiglu have passed. We will continue to improve it in the future.

• Hardware Type:
• run make test to ensure correctness
• run make checkstyle to ensure code style
• run make test-convergence to ensure convergence

[Tcc0403]

LGTM

[ji-huazhong]

LGTM