[Kernel][Quantization] feat: add Gluon kernels for AWQ quantization

[AlpinDale]

Still a WIP. Need to build triton from source.

$ apt install zlib1g-dev
$ git clone https://github.com/triton-lang/triton.git && cd triton
$ uv pip install -r python/requirements.txt
$ uv pip install -ve .  # may take a while, needs to download a 1.2 GiB llvm archive

$ APHRODITE_USE_GLUON_AWQ=1 aphrodite run Orion-zhen/Qwen3-0.6B-AWQ -q awq --dtype float16 --enforce-eager

Does not work yet, needs some quirks ironed out.