Graphbolt: Enable separate stream for CUDA memory copy and computation

[peizhou001]

🔨Work Item

IMPORTANT:

• This template is only for dev team to track project progress. For feature request or bug report, please use the corresponding issue templates.
• DO NOT create a new work item if the purpose is to fix an existing issue or feature request. We will directly use the issue in the project tracker.

Project tracker: https://github.com/orgs/dmlc/projects/2

Description

Depending work items or issues

[mfbalin]

We will probably handle this automatically when we finalize the design of the pipelining and executor logic for the sampling stage.

[mfbalin]

This is already implemented in the dataloader with the overlap_feature_fetch switch.

[mfbalin]

@frozenbugs is there more to be done for this issue? We already support feature copy overlap when features are pinned.