Skip to content

Commit 13f7fd1

Browse files
matthewfeickertmatthewfeickert
committed
feat: Add 'Reproducible ML for Scientists with Pixi' draft
1 parent 5d067d1 commit 13f7fd1

File tree

9 files changed

+110
-0
lines changed

9 files changed

+110
-0
lines changed

‎papers/matthew_feickert/acknowledgements.md‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
## Acknowledgements
2+
3+
Matthew Feickert is supported by the U.S. National Science Foundation (NSF) under Cooperative Agreement PHY-2323298 (IRIS-HEP) and by the US Research Software Sustainability Institute (URSSI) via grant G-2022-19347 from the Sloan Foundation.
4+
Ruben Arts is supported by Prefix.dev.
5+
John Kirkham is supported by NVIDIA.

‎papers/matthew_feickert/banner.png‎

495 KB
Loading

‎papers/matthew_feickert/conda-forge.md‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,3 @@
1+
## conda-forge
2+
3+
Technical overview of conda-forge and conda packages.

‎papers/matthew_feickert/cuda.md‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
## CUDA
2+
3+
* Overview of CUDA 12 packages
4+
* Importance of the work
5+
* What it enables

‎papers/matthew_feickert/introduction.md‎

Lines changed: 22 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,22 @@
1+
## Introduction
2+
3+
"Reproducible" research can mean many things and is a multipronged problem.
4+
Focus primarily on computational reproducibility.
5+
Like all forms of reproducibility, there are multiple "levels" of reproducibility.
6+
7+
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum sapien
8+
tortor, bibendum et pretium molestie, dapibus ac ante. Nam odio orci, interdum
9+
sit amet placerat non, molestie sed dui. Pellentesque eu quam ac mauris
10+
tristique sodales. Fusce sodales laoreet nulla, id pellentesque risus convallis
11+
eget. Nam id ante gravida justo eleifend semper vel ut nisi. Phasellus
12+
adipiscing risus quis dui facilisis fermentum. Duis quis sodales neque. Aliquam
13+
ut tellus dolor. Etiam ac elit nec risus lobortis tempus id nec erat. Morbi eu
14+
purus enim. Integer et velit vitae arcu interdum aliquet at eget purus. Integer
15+
quis nisi neque. Morbi ac odio et leo dignissim sodales. Pellentesque nec nibh
16+
nulla. Donec faucibus purus leo. Nullam vel lorem eget enim blandit ultrices.
17+
Ut urna lacus, scelerisque nec pellentesque quis, laoreet eu magna. Quisque ac
18+
justo vitae odio tincidunt tempus at vitae tortor.
19+
20+
* Introduction
21+
* Computational reproducibility
22+
* Computational reproducibility vs. scientific reuse

‎papers/matthew_feickert/main.md‎

Lines changed: 23 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,23 @@
1+
---
2+
# Ensure that this title is the same as the one in `myst.yml`
3+
title: Reproducible Machine Learning Workflows for Scientists with Pixi
4+
abstract: |
5+
Scientific researchers need reproducible software environments for complex applications that can run across heterogeneous computing platforms.
6+
Modern open source tools, like Pixi, provide automatic reproducibility solutions for all dependencies while providing a high level interface well suited for researchers.
7+
Combined with the recent emergence of the entire CUDA software stack — from compilers to development libraries — being supported on conda-forge, researchers are now able to easily specify their exact hardware acceleration requirements and software dependencies and get portable computational environments locked down to the hash level.
8+
---
9+
10+
:::{include} introduction.md
11+
:::
12+
13+
:::{include} conda-forge.md
14+
:::
15+
16+
:::{include} cuda.md
17+
:::
18+
19+
:::{include} pixi.md
20+
:::
21+
22+
:::{include} acknowledgements.md
23+
:::

‎papers/matthew_feickert/mybib.bib‎

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,6 @@
1+
@software{pixi,
2+
author = {Arts, Ruben and Zalmstra, Bas and Vollprecht, Wolf and de Jager, Tim and Morcotilo, Nichita and Hofer, Julian},
3+
license = {BSD-3-Clause},
4+
title = {{pixi}},
5+
url = {https://github.com/prefix-dev/pixi/releases/tag/v0.48.0}
6+
}

‎papers/matthew_feickert/myst.yml‎

Lines changed: 41 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,41 @@
1+
version: 1
2+
extends: ../papers.yml
3+
project:
4+
# Update this to match `scipy-2025-<folder>` the folder should be `<firstname_surname>`
5+
id: scipy-2025-matthew_feickert
6+
# Ensure your title is the same as in your `main.md`
7+
title: Reproducible Machine Learning Workflows for Scientists with Pixi
8+
# subtitle:
9+
description: |
10+
Scientific researchers need reproducible software environments for complex applications that can run across heterogeneous computing platforms.
11+
Modern open source tools, like Pixi and the CUDA conda-forge packages, provide reproducibility solutions for while providing high level semantics well suited for researchers.
12+
# Authors should have affiliations, emails and ORCIDs if available
13+
authors:
14+
- name: Matthew Feickert
15+
email: [email protected]
16+
orcid: 0000-0003-4124-7862
17+
affiliations:
18+
- University of Wisconsin&ndash;Madison
19+
corresponding: true
20+
- name: Ruben Arts
21+
affiliations:
22+
- Prefix.dev
23+
- name: John Kirkham
24+
affiliations:
25+
- NVIDIA
26+
keywords:
27+
- reproducible
28+
- machine learning
29+
- hardware acceleration
30+
- CUDA
31+
- conda-forge
32+
- Pixi
33+
# Add the abbreviations that you use in your paper here
34+
abbreviations:
35+
CUDA: Compute Unified Device Architecture
36+
exports:
37+
- id: pdf
38+
format: typst
39+
template: https://github.com/curvenote-templates/scipy.git
40+
article: main.md
41+
output: full_text.pdf

‎papers/matthew_feickert/pixi.md‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,5 @@
1+
## Pixi
2+
3+
* Modern manager [@pixi]
4+
* Automatic and non-optional hash level lock files upon workspace mutation
5+
* Multi-platform

0 commit comments

Comments
 (0)