[IR][Script] Support indexing/slicing of both shared/global tensors

[yaoyaoding]

This PR adds the support of indexing and slicing of global and shared tensors.

New instructions added:

• GlobalSlice
• GlobalIndex
• SharedSlice
• SharedIndex

In Tilus Script, we can use the following syntax to indexing/slicing a shared/global tensor:

a: fp16[32, 64]
a[1, 2] # indexing, got fp16
a[1, :] # slice, got fp16[64]
a[:, 3] # slice, got fp16[32]

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.