ENH: Review connected-component buffer size guarantees

Đang mở
#1,928 0 bình luận 0 reaction 0 người được giao Xem trên GitHub

Maintainer thường phản hồi trong vòng 6 ngày

Chưa có ai nhận issue này.

Đánh giá

Độ khó
5/5
Thời gian dự kiến
Hơn một tuần
Mức phù hợp với người mới
32/100
Loại issue
Tính năng
Độ rõ ràng
Khá rõ ràng
Mức độ hoạt động
Sôi nổi
Công nghệ
cpp
Lĩnh vực
backend, performance

Hướng nghiên cứu

Start by reading ConnectedComponentEngine and the linked simplnx #1764 change, then trace its 2D row, tile, replay and finalization buffers, 3D plane and grouped-transfer buffers, and ScratchBuffer<T> callers. The issue asks for agreement on exact allocation guarantees, growing run and label data, fallback behavior, and test ownership before choosing an implementation. Done means a reviewed plan covering representative 2D/3D cases, both storage modes, actual disk-backed storage, runtime and memory, and preserved output, error, and cancellation behavior.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.

Mô tả

enhancement Out-of-Core

Summary

Review the data-sized working buffers in ConnectedComponentEngine and establish which buffers need exact element-count allocation under the agreed data-payload accounting rule. This work is deferred from simplnx #1764 to a separate branch. The intended scope, including dynamically growing run and label state, needs agreement before implementation.

Mike's review note includes algorithm working arrays in data payload and names the connected-component engine. The engine uses existing std::vector buffers in its 2D and 3D paths. In the 2D path, the planner counts a requested number of run elements, while std::vector::reserve only promises a capacity of at least that number. The 2D value is a planning target; it is not itself a working-memory reservation. The 3D path also uses vectors for some buffers associated with working reservations.

These vectors existed before the accounting change in #1764. This observation does not establish that a supported standard library has allocated extra capacity or that a user has encountered a memory overrun. Normal allocator overhead is outside the payload accounting rule, and the rule does not cap the whole process.

Inspect the 2D row, tile, replay and finalization buffers and the 3D plane and grouped-transfer buffers. Decide separately how the rule applies to growing run, union-find and label data and to the existing low-memory fallback. The generic ScratchBuffer<T> currently uses std::vector for non-bool types, so its name alone does not establish an exact-capacity guarantee. Avoid changing it globally without reviewing its other callers.

The engine serves the Connected Component Image Filter, Threshold Maximum Connected Components Image Filter, and Morphological Watershed Image Filter. It runs with in-memory as well as out-of-core stores. Preserve output, error and cancellation behavior; check runtime and memory in both storage modes. Use independent small label/count expectations, representative 2D and 3D cases, tight and sufficient allowances, and tests that witness actual disk-backed storage. Verify which engine path each test reaches. Review proposed buffer ownership and test strategy before selecting an implementation.

Ngôn ngữ chính
C++
Star
18
Fork
15
Merge trung bình
4 ngày 23 giờ
Pull request đã merge (30 ngày)
42

Chuẩn bị môi trường

Bắt đầu từ đâu

  1. Đọc hết issue, rồi đọc hướng dẫn đóng góp của dự án.
  2. Bình luận trên issue rằng bạn sẽ nhận — tránh hai người làm cùng một việc.
  3. Fork repository và làm thay đổi trên một nhánh.
  4. Mở pull request có tham chiếu số hiệu của issue.

Issue khác của BlueQuartzSoftware/simplnx

Tất cả issue của BlueQuartzSoftware/simplnx

Issue tương tự

Thêm issue về C++

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.