Skip to content

feat(embeddings): add parallel processing and thread count control#169

Merged
donhardman merged 2 commits into
masterfrom
feature/embeddings-threads
Jun 19, 2026
Merged

feat(embeddings): add parallel processing and thread count control#169
donhardman merged 2 commits into
masterfrom
feature/embeddings-threads

Conversation

@donhardman

Copy link
Copy Markdown
Member
  • Add rayon dependency for parallel execution support
  • Update TextModel and model traits to include threads parameter
  • Implement scoped rayon thread pools for candle-based models
  • Update ONNX pipelined inference to respect worker limits
  • Modify FFI interface and C++ wrapper to accept thread count
  • Increment library version to 4 for breaking FFI changes
  • Propagate thread settings through KNN and text-to-vector pipelines
  • Update Jina, local models, and test suites for new signatures
  • Allow zero threads to signify use of all available CPUs

BREAKING CHANGE: TextModel::predict and MakeVectEmbeddingsFn signatures changed to include thread count parameter.

@github-actions

github-actions Bot commented May 13, 2026

Copy link
Copy Markdown

Windows test results

  5 files    5 suites   25m 41s ⏱️
517 tests 503 ✅ 14 💤 0 ❌
525 runs  511 ✅ 14 💤 0 ❌

Results for commit ff8354a.

♻️ This comment has been updated with latest results.

@github-actions

github-actions Bot commented May 13, 2026

Copy link
Copy Markdown

Linux debug test results

  8 files    8 suites   23m 31s ⏱️
540 tests 520 ✅ 14 💤 6 ❌
554 runs  534 ✅ 14 💤 6 ❌

For more details on these failures, see this check.

Results for commit ff8354a.

♻️ This comment has been updated with latest results.

@github-actions

github-actions Bot commented May 13, 2026

Copy link
Copy Markdown

Linux release test results

  8 files    8 suites   9m 2s ⏱️
540 tests 526 ✅ 14 💤 0 ❌
554 runs  540 ✅ 14 💤 0 ❌

Results for commit ff8354a.

♻️ This comment has been updated with latest results.

- Add rayon dependency for parallel execution support
- Update TextModel and model traits to include threads parameter
- Implement scoped rayon thread pools for candle-based models
- Update ONNX pipelined inference to respect worker limits
- Modify FFI interface and C++ wrapper to accept thread count
- Increment library version to 4 for breaking FFI changes
- Propagate thread settings through KNN and text-to-vector pipelines
- Update Jina, local models, and test suites for new signatures
- Allow zero threads to signify use of all available CPUs

BREAKING CHANGE: TextModel::predict and MakeVectEmbeddingsFn signatures
changed to include thread count parameter.
@donhardman
donhardman force-pushed the feature/embeddings-threads branch from c99e78b to ff8354a Compare June 15, 2026 07:32
@donhardman
donhardman merged commit 057d176 into master Jun 19, 2026
421 of 452 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant