Numba: a LLVM-based Python JIT compiler

By Siu Kwan Lam, Antoine Pitrou, Stanley Seibert. 2015

1,806 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Hardware and Architecture, Information Systems and Management, ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Haiyang Huang
2. A. Lewis
3. Jesús Torrado
4. Daniel F. Schmidt
5. Tom Brown
6. K. Jarrod Millman
7. Geoffrey I. Webb
8. David P. Schlachtberger

In Hardware and Architecture

1. Jonas Horsch
2. Angus Dempster
3. Tom Brown
4. A. Lewis
5. Thibault Formal
6. Benjamin Piwowarski
7. Geoffrey I. Webb
8. John Healy

In Information Systems and Management

1. Matthew Brett
2. Jonas Horsch
3. Angus Dempster
4. Tom Brown
5. A. Lewis
6. Thibault Formal
7. Benjamin Piwowarski
8. Geoffrey I. Webb

In ML Systems & Hardware

1. Jonas Horsch
2. Tom Brown
3. David P. Schlachtberger
4. Angus Dempster
5. A. Lewis
6. Thibault Formal
7. Benjamin Piwowarski
8. Geoffrey I. Webb

Citations in the corpus

• Array programming with NumPy (2020, background): “[44] S.”
• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction (2018, methodology): “ed that MulticoreTSNE is a heavily optimized implementation wri−en in C++ based on Van der Maaten’s bhtsne [26] code. In contrast our UMAP implementation was wri−en in Python (making use of the numba [10] library for performance). MulticoreTSNE was run in single threaded mode to make fair comparison”
• Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualization (2020, methodology): “We implemented PaCMAP in Python by adapting code from TriMAP (Amid and Warmuth, 2019) and using the packages ANNOY (Bernhardsson, 2019) and Numba (Lam et al., 2015).”
• openTSNE: a modular Python library for t-SNE dimensionality reduction and embedding (2019, neutral): “This result likely stems from the fact that UMAP is written in pure Python and uses numba (Lam, Pitrou, and Seibert 2015) to accelerate computation.”
• SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking (2021, methodology): “For storing the index, we use a custom implementation based on Python arrays, and we rely on Numba [14] to parallelize retrieval.”
• TensorFlow Eager: A Multi-Stage, Python-Embedded DSL for Machine Learning (2019, background): “Outside of DSLs, there are several projects that provide just-in-time (JIT) compilation for Python, of which Numba (Lam et al., 2015) and PyPy (Bolz et al., 2009) are two examples.”
• Cobaya: code for Bayesian analysis of hierarchical physical models (2020, methodology): “those containing long explicit loops) can also be implemented using popular just-in-time compilers or language extensions such as NUMBA [16] or Cython [17], or coded in C/Fortran and interfaced to Python.”
• SPLADE v2: Sparse Lexical and Expansion Model for Information Retrieval (2021, methodology): “For storing the index, we use a custom implementation based on Python arrays, and we rely on Numba [15] to parallelize retrieval.”
• Fast and flexible bacterial genomic epidemiology with PopPUNK (2018, methodology): “36.2 to compile functions (Lam et al. 2015) and exploited multithreading of Mash sketching and distance calculations.”