feat: support Global Performance Tuning

[zewenli98]

Description

TensorRT's Global Performance Tuner searches internal builder knobs for faster engines. Torch-TensorRT exposes the same capability in-process for Dynamo TRT subgraphs. Please see the description for details.

Fixes # (issue)

Type of change

• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

Checklist:

• My code follows the style guidelines of this project (You can use the linters)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas and hacks
• I have made corresponding changes to the documentation
• I have added tests to verify my fix or my feature
• New and existing unit tests pass locally with my changes
• I have added the relevant labels to my PR in so that relevant reviewers are notified

[jianhu-nv]

Actually we cannot guarantee the TensorRT will not crash when setting a build route, for example, some driver level exception might be raised and it will not be caught by the try-catch block, and probably will crash the entire Python program. So I suggest we spawn a new process like the implementation in trtexec, so that even if the new child process crashes, the parent process remains unaffected and can continue to the next iteration.