guan404ming opened a new issue, #799: URL: https://github.com/apache/mahout/issues/799
# QDP POC Roadmap > Quantum Data Plane development progress tracking ## Phase 1: Foundation & Infrastructure | Status | Task | |--------|------| | ✅ | Rust workspace & config | | ✅ | Implement robust `build.rs` for Kernels | --- ## Phase 2: The Core | Status | Task | |--------|------| | ✅ | CUDA Kernel Implementation | | ✅ | Rust CPU Pre-processing | | ✅ | GPU Memory Allocation and Load | --- ## Phase 3: Zero-copy and Safety | Status | Task | |--------|------| | ✅ | Define `DLManagedTensor` and DLPack structure | | ✅ | Implement resource `tensor_holder` and deleter | | ✅ | Core engine API refinement | --- ## Phase 4: Python Binding | Status | Task | |--------|------| | ✅ | PyO3 class wrapping | | ✅ | Implement DLPack protocol | --- ## Phase 5: Input Format Support | Status | Task | |--------|------| | ⬜ | PyTorch tensor input ([#717](https://github.com/apache/mahout/issues/717)) | | ⬜ | NumPy array input ([#718](https://github.com/apache/mahout/issues/718)) | | ⬜ | TensorFlow tensor input ([#719](https://github.com/apache/mahout/issues/719)) | | ⬜ | Pandas DataFrame input | --- ## Phase 6: Verification and Test | Status | Task | |--------|------| | ✅ | Fidelity test | | ✅ | Benchmark | | ⬜ | Different devices testing | | ⬜ | Benchmark with real dataset | --- ## Phase 7: More Encoders & Architecture (optional) | Status | Task | |--------|------| | ✅ | Amplitude encoding | | ⬜ | Angle encoding | | ⬜ | Basis encoding | | ⬜ | Multi-architecture GPU optimization | --- ## Phase 8: Integration & Release | Status | Task | |--------|------| | ⬜ | Documentation | | ⬜ | Integrate with QuMat | | ⬜ | Example notebooks | | ⬜ | Performance benchmarks | | ⬜ | PyPI package publishing | --- ## Open Issues by Category ### Meta/Tracking Issues | Issue | Title | Description | |-------|-------|-------------| | [#752](https://github.com/apache/mahout/issues/752) | ROADMAP for new input format support | Master tracking issue | | [#716](https://github.com/apache/mahout/issues/716) | Supply more input formats | Parent for torch/numpy/tf | | [#772](https://github.com/apache/mahout/issues/772) | [RFC] Benchmark roadmap | Benchmark planning | --- ### Input Format Support (tracked by #752) #### PyTorch Support > **Parent:** [#717](https://github.com/apache/mahout/issues/717) - Supply torch.Tensor in input format | Issue | Task | |-------|------| | [#726](https://github.com/apache/mahout/issues/726) | GPU tensor zero-copy via dlpack protocol | | [#727](https://github.com/apache/mahout/issues/727) | Add PyTorch tensor optimizations, tests, docs | #### NumPy Support > **Parent:** [#718](https://github.com/apache/mahout/issues/718) - Supply np.ndarray in input format | Issue | Task | |-------|------| | [#721](https://github.com/apache/mahout/issues/721) | Add NumPy 1D array input for single sample | | [#722](https://github.com/apache/mahout/issues/722) | Add NumPy batch input + memory optimizations | | [#723](https://github.com/apache/mahout/issues/723) | Add NumPy input documentation and more tests | | [#787](https://github.com/apache/mahout/issues/787) | NumPy file size check to prevent OOM (followup) | | [#789](https://github.com/apache/mahout/issues/789) | NumPy input speed & memory improvement (followup) | #### TensorFlow Support > **Parent:** [#719](https://github.com/apache/mahout/issues/719) - Supply tf.Tensor in input format | Issue | Task | |-------|------| | [#728](https://github.com/apache/mahout/issues/728) | Add TensorFlow eager tensor + CPU/GPU detection | | [#729](https://github.com/apache/mahout/issues/729) | GPU TF tensor support with dlpack | | [#730](https://github.com/apache/mahout/issues/730) | TF-Specific Optimizations and Documentation | --- ### Bugs | Issue | Title | |-------|-------| | [#696](https://github.com/apache/mahout/issues/696) | [Bug] CUDA memory illegal address error | | [#743](https://github.com/apache/mahout/issues/743) | Fix async amplitude pipeline out-of-bounds write | | [#765](https://github.com/apache/mahout/issues/765) | Unify Null Value Handling (Batch vs Streaming) | --- ### Performance & Optimization | Issue | Title | |-------|-------| | [#714](https://github.com/apache/mahout/issues/714) | Merge norm and encode into a kernel method | | [#746](https://github.com/apache/mahout/issues/746) | DLPack shape/strides: Support batch 2D tensor | --- ### Infrastructure & Tooling | Issue | Title | |-------|-------| | [#678](https://github.com/apache/mahout/issues/678) | Make CI runs when push on local branch | | [#736](https://github.com/apache/mahout/issues/736) | Add clippy & rustfmt as linter | | [#750](https://github.com/apache/mahout/issues/750) | Reconstruct project files into multiple platforms | -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
