feat: INT8 scaled matmul Triton kernel

[namgyu-youn]

Summary:
Introduce a Triton kernel for INT8 scaled matrix multiplication.

Motivation: The existing kernel (torch._int_mm) crashes with small dimension inputs in vLLM (failure log: https://gist.github.com/vkuzo/5bf389079442bb9851ef315cdcb797b4).

Micro benchmark results show the Triton kernel is ~2.6x faster than int_scaled_matmul (full log: https://gist.github.com/namgyu-youn/aa2fc5d444fdc4b52b35db555087e2ce).

Test plan:

pytest -sv test/kernel/test_int8mm_triton.py

Future work:

• Add autotuning for dynamic block size selection
• Add Swizzling (GROUP_M) for L2 cache optimization
• Add SplitK for K-dimension parallelization (improves decode with small M)
• Replace int_scaled_matmul in Int8Tensor after e2e validation

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/3558

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• B200 runners are down due to network issues

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[namgyu-youn]

@pytorchbot label "topic: new feature"