[Lane 2 · kotlin-core] StableHLO coverage — lower conv3d to StableHLO #1398

Sub-issue of #1389 ([Feature]: StableHLO converter coverage — lower every TensorOps op).

Lane: 2 · Kotlin core
Skill needed: Kotlin; reading a little MLIR text. No SKaiNET runtime or kernel knowledge needed — converters only emit text.
Size: m (1–2 days)
Blocked by: nothing

What to do

1. Open skainet-compile/skainet-compile-hlo/src/commonMain/kotlin/sk/ainet/compile/hlo/converters/NeuralNetOperationsConverter.kt and find an op it already lowers that is similar.
2. Add conv3d to its supportedOperations and emit: stablehlo.convolution with three spatial dims — generalise the existing conv2d emission (dimension numbers [b, f, 0, 1, 2]).
3. Add a one-node export test next to the converter's existing tests in skainet-compile/skainet-compile-hlo/src/commonTest (copy one of its tests).
4. Run ./gradlew :skainet-compile:skainet-compile-hlo:jvmTest, then regenerate the matrix: ./gradlew :skainet-compile:skainet-compile-hlo:jvmTest --tests '*HloCoverageDumpTest' :skainet-compile:skainet-compile-dag:jvmTest --tests '*BackwardCoverageDumpTest' generateDocs and commit the changed docs/modules/ROOT/pages/reference/ files.

Acceptance

• StableHLO column shows ✅ for conv3d
• Export test passes
• Result reported back on the parent issue

Notes

If something here turns out to be wrong or much bigger than the size label, comment on this issue or open a specific bug instead of blocking the parent.

Coverage matrix: docs/modules/ROOT/pages/reference/ops-status-matrix.adoc · How lanes work: Issue taxonomy