Make P2P in warmup/cooldown stage to sync comm.

Add microbatch_group_size_per_vp_stage as configurable.

Add DualPipe-V support.

Add support for DualPipe.

Update overlapped time.

Add support for 1F1B-interleave-overlap.

Add visualization for 1F1B overlap.

Update implementation for 1F1B overlapping.

Add support for zero-bubble-1P.

Add bubble rate formula.

Refactor schedule execution model and simplify execution flow

Fix bug.

Add VPP support and refactor project.