feat: Implement TensorCast broadcast tree materialization#172
Open
FernanDAlumin wants to merge 49 commits into
Open
feat: Implement TensorCast broadcast tree materialization#172FernanDAlumin wants to merge 49 commits into
FernanDAlumin wants to merge 49 commits into
Conversation
… write RDMA serving for HomeBatchPut
- Thread activated HOST_SHARED stable-backing refs and keepalives through source spans and segmented put-side communicator exports. - Add stable-backed lightweight source-view keys that preserve the existing BatchPayloadCommunicatorSource schema while splitting view-key cleanup from stable chunk MR ownership. - Teach READ_PLAN_REQUEST admission to resolve stable-backed source views on the requested rail through StableLocalBackingState::ensure_chunk(), bypassing broker-owned raw tensor registration. - Validate with targeted daemon/communicator tests, tensorcast daemon build, and the share-remote SGLang replay showing stable_backing_view exports with cached-token hits intact
- Add missing TYPE_CHECKING imports in tensorcast/__init__.py for names
exposed via _LAZY_ATTRS so they satisfy __all__
- Restructure broadcast transport claim with try/except/else so pyright
narrows replica/edge as non-None on the success path
- Pass transport_request_id and transport_scheduling_group as typed
named arguments instead of via dict[str, object] **kwargs to match
materialize_by_artifact_id_v2 overloads
Add after the guard so strict type checkers can narrow to non-optional inside the try/finally bodies. Resolves 22 union-attr errors in broadcast_repository.py.
FernanDAlumin
changed the title
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
2 participants
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
Test Plan
PYTHONPATH=/data/tot/tensorcast:/data/tot/tensorcast/proto/gen/python /data/tot/.venv/bin/python -m pytest tests/python/global_store/test_broadcast_service.py12 passedgit diff --checkbazel test //daemon:grpc_service_impl_no_lease_materialize_testcudaErrorInsufficientDriver: CUDA driver version is insufficient for CUDA runtime version.