mpl: add option to store clustering data as dbGroups

[AcKoucher]

For #7959.

Next step is to allow browsing groups using the Hierarchy Browse widget in the GUI.

[maliberty]

@eder-matheus please review as I'm getting backed up

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[eder-matheus]

The code looks good, but I would like to see how creating the groups affects the flow after MPL. It's possible that we need to change some other tools and prevent them of looking at this groups created by MPL, since they are just for visualization and debugging.

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[maliberty]

If necessary we could a distinguishing value to dbGroupType

[AcKoucher]

I added a new value to dbGroupType and I'm working to make the placer understand that it should ignore it.

[AcKoucher]

I updated this branch locally with master now that both #7992 and #7986 we're merged and I still get different placement results when running the flow. It looks like DPL is also affected by the new group, so apparently there's more work needed to allow the changes here.

[maliberty]

Yes we support groups in dpl/dpo too

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[AcKoucher]

@maliberty @eder-matheus I ran a secure-CI enforcing the new flag and I got the same results as running without it, so this should be good to merge.

[eder-matheus]

Changes look good, but you have a failure on this unit test:
mpl.keep_clustering_data.tcl (Failed)

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[eder-matheus]

@AcKoucher I believe your unit test is failing in the bazel build due to a missing dependency on the input files. I had something similar in a recent PR, see how I fixed it here: https://github.com/The-OpenROAD-Project/OpenROAD/pull/8554/files#diff-85df7b235e804a7f2ad2d172187227f957651f66e912f40a0f41fc521c86ba53R332-R337

[AcKoucher]

Thanks @eder-matheus.

There were two issues. The first one was the actual test result, the other was a missing bazel resource as you pointed out. I missed that there are multiple tests using the same resource and that's being addressed later in the file:

OpenROAD/src/mpl/test/BUILD

Lines 112 to 126 in 9a0e3df

	[
	filegroup(
	name = test_name + "_resources",
	srcs = [":regression_resources"] + glob(
	[
	test_name + ".*",
	] + {
	"io_constraints6": ["testcases/io_constraints6.def"],
	"io_constraints7": ["testcases/io_constraints6.def"],
	"io_constraints8": ["testcases/io_constraints6.def"],
	}.get(test_name, []),
	),
	)
	for test_name in ALL_TESTS
	]

The last commit should handle it.

[github-actions]

clang-tidy review says "All clean, LGTM! 👍"

[AcKoucher]

@maliberty The check failure looks unrelated to the changes here:

docker: Error response from daemon: failed to create task for container:
Unavailable: error reading from server: EOF: unavailable.
[...]

[maliberty]

@mguthaus FYI