Metraipolis › archive › dc_17e956cd0247

GPU Acceleration of Awkward Arrays: Using Python cuda.compute — arXiv:2610.08238

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.08238
License (as archived)
CC0-1.0
Archived
2026-10-07T11:24:42.041Z
SHA-256 of the archived text
5d9cded14fc82fca837cc348099e55dc16bd4e7bbbbabc000f3f7c6c91f990a5
Stable URL of this page
https://metraipolis.com/documento/dc_17e956cd0247

How to cite

GPU Acceleration of Awkward Arrays: Using Python cuda.compute — arXiv:2610.08238 — arXiv (en), https://arxiv.org/abs/2610.08238 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_17e956cd0247 (Metraipolis archive dc_17e956cd0247, sha256 5d9cded14fc82fca837cc348099e55dc16bd4e7bbbbabc000f3f7c6c91f990a5)
BibTeX
@misc{dc17e956cd0247,
  title = {GPU Acceleration of Awkward Arrays: Using Python cuda.compute — arXiv:2610.08238},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.08238},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_17e956cd0247; sha256 5d9cded14fc82fca837cc348099e55dc16bd4e7bbbbabc000f3f7c6c91f990a5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.08238. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Awkward Array is a widely used library in high-energy physics (HEP) for representing and manipulating nested, variable-length data in Python. Previous CHEP contributions have explored GPU acceleration for Awkward Array, demonstrating the feasibility and performance benefits of CUDA-based backend while also identifying limitations related to irregular data access, fine-grained kernel launches, and composability of operations. In this contribution, we present recent developments that build directly on these earlier efforts by introducing a CUDA execution model for Awkward Array based on the Python CUDA Core Compute Libraries (CCCL). Using CCCL, we eliminate the need for custom CUDA kernels and can instead use a high-level Python interface. The CCCL-based approach also enables fusion of multiple Awkward operations into a reduced number of CUDA kernels, addressing kernel launch overhead observed in earlier GPU implementations. Lazy execution allows expression graphs to be constructed and optimized prior to kernel generation, improving performance for analysis workflows involving jagged arrays, combinatorial operations, and reductions. In contrast to earlier approaches, this design also emphasizes extensibility, allowing user-defined Python code to be incorporated into GPU execution paths with minimal boilerplate and without breaking existing analysis semantics. We present performance studies that demonstrate improvements over previously reported eager GPU execution strategies for representative HEP analysis patterns. These developments extend the GPU capabilities of Awkward Array toward a more composable and sustainable backend, aligned with the needs of Python-based analysis at the HL-LHC and beyond.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC) ; Performance (cs.PF); Software Engineering (cs.SE).
Pagina del lavoro: https://arxiv.org/abs/2610.08238 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T11:15:54.478Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).