Tracing Support in XDNA2

[Victor-Jung]

This PR adds trace generation via control packets in the Chrome Trace Event Format for visualization with Perfetto. Useful information such as kernel runtime, vector unit utilization, and DMA stalls for each AIE core can be visualized this way.

Below is an example of the generated trace for a double-buffered Add node:

Note: This PR builds on top of PR #179 and should be reviewed after #179 is merged.

Added

• Two new code transformations to configure the AIE cores and MemTile tracing packets.
• A new code transformation to configure the host-side buffer to store the trace.

Changed

• Use --dump-intermediate to dump the npu.write32 instructions in main_physical_with_elfs.mlir (used to parse the trace)
• Update the XDNA2 test harness to declare a buffer for the trace. The buffer size can be changed with the --trace-buffer-size <size> flag.

Fixed

PR Merge Checklist

1. The PR is rebased on the latest devel commit and pointing to devel.
2. Your PR reviewed and approved.
3. All checks are passing.
4. The CHANGELOG.md file has been updated.
5. If the docker was modified, change back its link after review.