Docs: drop PR-split L1a/L1b/L2/L5/L6 tags from distributed bring-up code#638
Merged
ChaoWao merged 1 commit intohw-native-sys:mainfrom Apr 22, 2026
Merged
Conversation
The distributed bring-up series (hw-native-sys#608 / hw-native-sys#610 / hw-native-sys#613) used `(L2)` / `(L5)` / `(L6)` as split-step labels in PR titles and commit messages. These tags leaked into docstrings, comments, and the allreduce_distributed example where they collide with the framework hierarchy in `docs/hierarchical_level_runtime.md` — which also calls its levels L0–L6, but means CORE / DIE / CHIP / HOST / POD / SuperNode / Cluster. Readers who encounter "L6 teardown order" or "the L1a root-info handshake" naturally reach for the topology definition and find the two numberings disagree (framework L6 = cluster, split-step L6 = `Worker` bootstrap loop). PR hw-native-sys#613's body already announced the cleanup for new and touched code; this commit finishes it across the files that were left behind: - `examples/workers/l3/allreduce_distributed/main.py`: replace the fake "L1a..L6 stack" table with an unlabeled component list and a pointer to the real topology doc. - `tests/ut/py/test_worker/test_bootstrap_context_{sim,hw}.py`: drop `(L5)` from the module titles and rewrite `L6 teardown order` / `L6's ChipContext` / `L5 one-shot bring-up` / `L1a root-info handshake` / `paired L1b UT` to name the thing (`Worker bootstrap loop`, `ChipContext`, `bring-up`, etc.). - `tests/ut/py/test_worker/test_bootstrap_channel.py`: `(L2 bootstrap mailbox)` → `(per-chip bootstrap mailbox)` — the old tag was ambiguous with framework L2 (CHIP). - `tests/ut/py/test_worker/test_platform_comm.py`: `L1a HCCL backend` → `HCCL backend`; `Known issue inherited from L1a` → `... from the HCCL backend`; `L1a observed CANN error 507018` → `The C++ HCCL UT observed ...`. The two `L2-boundary contract` references are left alone — those correctly refer to the framework L2 boundary documented in `hierarchical_level_runtime.md`. - `src/common/platform_comm/comm_sim.cpp`: `L1a contract alignment notes` → `HCCL backend contract alignment notes`. - `src/a2a3/platform/onboard/host/device_runner.cpp`: `L1a 507018` → `HCCL 507018`; `L1a C++ hardware UT` → `HCCL C++ hardware UT`. No runtime behavior change — comments and docstrings only. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
PRs #608 / #610 / #613 used
(L2)/(L5)/(L6)as split-step labels in their titles and commit messages. Those tags leaked into docstrings, comments, and theallreduce_distributedexample where they collide with the framework hierarchy indocs/hierarchical_level_runtime.md— which also calls its levels L0–L6, but means CORE / DIE / CHIP / HOST / POD / SuperNode / Cluster.A reader seeing
L6 teardown orderorL1a root-info handshakereaches for the topology definition and finds the two numberings disagree (framework L6 = cluster, split-step L6 =Workerbootstrap loop). PR #613's body already announced the cleanup for new and touched code; this PR finishes it on the files that were left behind.examples/workers/l3/allreduce_distributed/main.py— replace the fake "L1a..L6 stack" table with an unlabeled component list and a pointer to the real topology doc.tests/ut/py/test_worker/test_bootstrap_context_{sim,hw}.py— drop(L5)from module titles; rewriteL6 teardown order/L6's ChipContext/L5 one-shot bring-up/L1a root-info handshake/paired L1b UTto name the thing (Worker bootstrap loop,ChipContext,bring-up, etc.).tests/ut/py/test_worker/test_bootstrap_channel.py—(L2 bootstrap mailbox)→(per-chip bootstrap mailbox); the old tag was ambiguous with framework L2 (CHIP).tests/ut/py/test_worker/test_platform_comm.py— dropL1aprefixes in the HCCL backend references. The twoL2-boundary contractsentences are left alone — those correctly refer to the framework L2 boundary documented inhierarchical_level_runtime.md.src/common/platform_comm/comm_sim.cpp—L1a contract alignment notes→HCCL backend contract alignment notes.src/a2a3/platform/onboard/host/device_runner.cpp—L1a 507018→HCCL 507018;L1a C++ hardware UT→HCCL C++ hardware UT.No runtime behavior change — comments and docstrings only.
Test plan
grep -rnE '\bL1[ab]\b|\bL5\b|\bL6\b'acrosssrc/,python/,examples/,tests/reports only the legitimate framework-hierarchy uses indocs/hierarchical_level_runtime.mdandsrc/common/hierarchical/types.h.clang-format,clang-tidy,cpplint,ruff,pyright) all pass.🤖 Generated with Claude Code