proposal: add MetaX accelerator support (#2308)

* modelserver: add MetaX vLLM baseline profile

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* modelserver: add MetaX eight-GPU validation profile

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* docs: add MetaX prefill calibration results

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* docs: add MetaX maintainer placeholders

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* docs: update Metax mainter

Signed-off-by: Lianjie Zhang (ljzhang) <ljzhang@metax-tech.com>

* update maintainers

Signed-off-by: archlitchi <mengxuan.li@dynamia.ai>

* update Metax vLLM image

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* feat(guides): add MetaX C500X vLLM P/D disaggregation overlay

Ship a reduced 1P+1D / Qwen3-14B overlay so MetaX C500X can follow the
community well-lit P/D path over NixlConnector, using the public vLLM image.

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* update

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>

* my suggested changes to metax guide integration v1

Signed-off-by: greg pereira <grpereir@redhat.com>

* update_parameters

Signed-off-by: archlitchi <mengxuan.li@dynamia.ai>

---------

Signed-off-by: Jifei Wang <jifei.wang@dynamia.ai>
Signed-off-by: Lianjie Zhang (ljzhang) <ljzhang@metax-tech.com>
Signed-off-by: archlitchi <mengxuan.li@dynamia.ai>
Signed-off-by: greg pereira <grpereir@redhat.com>
Co-authored-by: Lianjie Zhang (ljzhang) <ljzhang@metax-tech.com>
Co-authored-by: archlitchi <mengxuan.li@dynamia.ai>
Co-authored-by: greg pereira <grpereir@redhat.com>