pov-shuffle 0.1.2__tar.gz

pov_shuffle-0.1.2/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

pov_shuffle-0.1.2/MANIFEST.in

@@ -0,0 +1,2 @@
+include setup.py
+recursive-include src *.h *.cpp *.cu

pov_shuffle-0.1.2/PKG-INFO

@@ -0,0 +1,9 @@
+Metadata-Version: 2.4
+Name: pov-shuffle
+Version: 0.1.2
+Requires-Python: <4,>=3.12
+License-File: LICENSE
+Requires-Dist: numpy~=2.4
+Requires-Dist: nvidia-cutlass~=4.2
+Requires-Dist: torch~=2.12
+Dynamic: license-file

pov_shuffle-0.1.2/README.md

@@ -0,0 +1,147 @@
+# POV-Shuffle
+_**P**arallel **O**ffset **V**irtual-block Shuffle_
+
+A fully parallel algorithm for efficiently shuffling large datasets without copy,
+while sufficiently approximating a uniform shuffle within few iterations.
+
+## Usage
+
+```python
+import povs
+
+# A numpy array or PyTorch tensor on CPU or CUDA device
+dataset = ...
+
+# Shuffle the dataset in place
+povs.shuffle(
+    dataset,  # A numpy array or PyTorch tensor on CPU or CUDA device
+    iterations=3,
+    options=povs.Options(virtual_block_size=2),  # Optional: specify full or partial algorithm options.
+                                                 # Unspecified parameters are chosen automatically.
+)
+```
+- For more details see `help(povs.shuffle)` and `help(povs.optim_options_for_dataset)`.
+
+## Installation
+```bash
+uv add pov-shuffle
+```
+
+### Build Customization
+Because of optimizations within the CUDA extension, the possible combinations of certain parameters need to be known at build-time.
+
+See [setup.py](setup.py) for the list of environment variables that can be set at build-time in order to add support for different parameter sets.
+
+In order to set environment variables consistently across builds we recommend using UV's [`extra-build-variables`](https://docs.astral.sh/uv/reference/settings/#extra-build-variables).
+Example:
+```toml
+# pyproject.toml
+[tool.uv.extra-build-variables.pov-shuffle]
+POVS_CUDA_INSTANTIATIONS = "4,32,64,float;2,64,16,int"  # Add support for 4x32x64xFloat and 2x64x16xInt (VBlockSize x PBlockSize x InstanceSize x DType)
+POVS_CUDA_INSTANCE_SIZES = "32,96,128"  # Add support for these instance sizes, in cartesian combination with supported dtypes and supported values of other parameters.
+```
+- Build parameters can be queried at runtime via `povs.get_build_params()`
+
+## Performance
+
+### Shuffle Time
+
+Shuffle time per deck size with 4-iterations POV-shuffle on instances of shape `(128 x float16)`, using the NVIDIA Ada Lovelace architecture.
+
+For a fair comparison with the algorithm, which offers close-to-uniform, zero-copy, in-place shuffling,
+the baseline used here is `numpy.shuffle`, which is also a uniform shuffle (Fisher-Yates), performed in-place and without copy.
+
+[![alt text](./data/time_per_deck_size/2026-06-13T20:54:09/plot.png)](./data/time_per_deck_size/2026-06-13T20:54:09)
+
+### Bias Convergence
+
+[Positional TVD bias](./docs/GLOSSARY.md#positional-tvd), [N-gram TVD biases](./docs/GLOSSARY.md#n-gram-tvd)
+and [LSTM predictability bias](./docs/GLOSSARY.md#lstm-predictability)
+measured for the baseline `numpy.shuffle` and for increasing iterations of the POV-shuffle,
+on a dataset of 1k distinct instances and estimating the event distributions from the observation
+of 3k independent shuffling episodes.
+
+[![alt text](./data/bias_per_iter/2026-06-22T19:18:55/plot.png)](./data/bias_per_iter/2026-06-22T19:18:55)
+
+### Breaking Point
+
+Minimum number of POV-shuffle iterations required for each bias metric to converge to the observed value
+of the respective metric for the baseline `numpy.shuffle`, as a function of the deck size.
+
+[![alt text](./data/breaking_point_per_deck_size/2026-06-22T14:05:38/report/plot.png)](./data/breaking_point_per_deck_size/2026-06-22T14:05:38/report)
+
+## Algorithm
+
+### How it works
+1. Partition the array into physical blocks of a specified size.
+2. For each iteration:
+   - Pick a random offset, so every block start is shifted from its original position, with the rightmost blocks wrapping around the array.
+   - Randomly assign each few physical blocks to a virtual block, so every virtual block is contiguous by parts.
+   - Each worker thread shuffles its assigned virtual block in place, using a standard shuffle algorithm (e.g. Fisher-Yates).
+
+Because there is no overlap between virtual blocks, the algorithm can be fully parallelized without facing race conditions,
+and doesn't require a temporary copy of the whole dataset to perform the shuffle in place.
+
+Compared to a traditional local-block shuffle, the virtual block assignment significantly reduces positional bias,
+while the random offset prevents the occurrence of shuffle artifacts from the physical block boundaries.
+
+When applied to higher rank tensors, the shuffle happens along the axis 0,
+with each indexable multidimensional object along that axis being treated as a flat 1D instance
+(e.g. for a tensor with shape `(I, M, N)`, we shuffle the `I` instances,
+each instance being a `(M, N)` matrix treated as an array of length `M*N`).
+
+### Trade-offs
+
+- **Block Size**:
+  - Larger blocks (both physical and virtual) increase the portion of the data that needs to be loaded into each worker at each iteration.
+  - On the other hand, smaller physical blocks increase the total number of physical blocks, so the host program has to do more non-parallel shuffles when randomly assigning them to virtual blocks.
+  - Therefore, as a rule of thumb one should use larger blocks to shorten the time per iteration, or smaller blocks if the priority is reducing the data transfer to workers.
+  - Remarkably, so far we have observed little impact of this parameter on the amount of iterations needed for shuffle convergence.
+
+### Diagrams
+Algorithm flowchart:
+```mermaid
+flowchart TD
+  A([Array]) --> B["Partition into physical blocks\n(size = physical_block_size)"]
+  B --> iter
+
+  subgraph iter ["For each iteration"]
+      direction TB                                                                                                                                                                                                         
+      C[/"Pick a random offset\n(not a multiple of block size)"/]                                                                                          
+      --> D["Randomly assign physical blocks\nto virtual blocks\n(virtual_block_size blocks each)"]
+      --> par
+
+      subgraph par ["In parallel — one worker per virtual block"]
+          direction LR
+          E["Read assigned physical blocks\nat +offset  (wrapping around array)"]
+          --> F["Shuffle in local memory"]
+          --> G["Write back to\noriginal positions"]
+      end
+  end
+                                                                                                                                                           
+  iter --> H([Done])
+```
+
+Sequence diagram with 2 workers shuffling 4 physical blocks (2 virtual blocks) in parallel:
+```mermaid
+sequenceDiagram
+  participant A as Array
+  participant W0 as Worker 0                                                                                                                                
+  participant W1 as Worker 1
+
+  Note over A,W1: Pick offset · Randomly assign virtual blocks
+
+  par VB0
+      W0->>A: Read blocks B1, B3 at +offset
+      A-->>W0: local copy
+      W0->>W0: shuffle
+      W0->>A: Write back to B1, B3
+  and VB1
+      W1->>A: Read blocks B0, B2 at +offset
+      A-->>W1: local copy
+      W1->>W1: shuffle
+      W1->>A: Write back to B0, B2
+  end
+
+  Note over A,W1: Non-overlapping virtual blocks — no race conditions
+```

pov_shuffle-0.1.2/pyproject.toml

@@ -0,0 +1,91 @@
+
+[build-system]
+build-backend = "setuptools.build_meta"
+requires = [
+    "wheel",
+    "setuptools",
+    "nvidia-cutlass   ~=4.2",
+    "torch            ~=2.12",
+]
+
+[project]
+name = "pov-shuffle"
+version = "0.1.2"
+requires-python = ">=3.12,<4"
+
+dependencies = [
+    "numpy            ~=2.4",
+    "nvidia-cutlass   ~=4.2",
+    "torch            ~=2.12",
+]
+
+[dependency-groups]
+dev = [
+    "matplotlib       >=3.10.9",
+    "pytest           >=8.0",
+    "setuptools       >=80.9",
+    "ruff             ~=0.14",
+    "jupyter          ~=1.1",
+    "tqdm             >=4.67.3",
+    "pandas           >=3.0.3",
+    "pyyaml           >=6.0.3",
+    "pybind11-stubgen >=2.5.5",
+]
+
+[tool.pytest.ini_options]
+filterwarnings = [
+    "ignore::UserWarning:povs.torch:",
+]
+
+[tool.setuptools.package-data]
+povs = ["build_params.json"]
+
+[tool.uv]
+reinstall-package = ["pov-shuffle"]
+python-preference = "only-managed"
+
+[tool.uv.extra-build-variables.pov-shuffle]
+POVS_CUDA_CARTESIAN_INSTANCING = "false"
+POVS_CUDA_INSTANTIATIONS = "2,16,8,float;2,32,8,float;2,16,128,c10::Half;2,32,1,int"
+
+# === Ruff ========================================================================================
+
+[tool.ruff]
+line-length = 120
+indent-width = 4
+extend-include = []
+exclude = [
+    ".eggs",
+    ".git",
+    ".ruff_cache",
+    ".venv",
+    "__pypackages__",
+    "build",
+    "dist",
+    "setup.py",
+]
+
+[tool.ruff.format]
+preview = true
+quote-style = "double"
+indent-style = "space"
+skip-magic-trailing-comma = false
+line-ending = "auto"
+
+[tool.ruff.lint]
+exclude = [""]
+fixable = ["ALL"]
+unfixable = []
+select = [
+    "F",   # Pyflakes
+    "E",   # Errors
+    "W",   # Warning
+    "I",   # isort
+    "TCH", # Typecheck
+    "SIM", # flake8-simplify
+#    "TID", # flake8-tidy-imports
+]
+ignore = [
+    "E501",  # Line too long
+    "E731",  # Do not assign a lambda expression, use a def
+]

pov_shuffle-0.1.2/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+