Metraipolis › archive › dc_0a9d8b60cd82

Ascend-SpMM: Accelerating Sparse Matrix-Dense Matrix Multiplication on Ascend NPUs — arXiv:2611.01936

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.01936
License (as archived)
CC0-1.0
Archived
2026-11-03T18:26:40.734Z
SHA-256 of the archived text
b6115d6c5399e12daba33ab3c8fb93918faad35abff36d92daa1084b02b40f46
Stable URL of this page
https://metraipolis.com/documento/dc_0a9d8b60cd82

How to cite

Ascend-SpMM: Accelerating Sparse Matrix-Dense Matrix Multiplication on Ascend NPUs — arXiv:2611.01936 — arXiv (en), https://arxiv.org/abs/2611.01936 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0a9d8b60cd82 (Metraipolis archive dc_0a9d8b60cd82, sha256 b6115d6c5399e12daba33ab3c8fb93918faad35abff36d92daa1084b02b40f46)
BibTeX
@misc{dc0a9d8b60cd82,
  title = {Ascend-SpMM: Accelerating Sparse Matrix-Dense Matrix Multiplication on Ascend NPUs — arXiv:2611.01936},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01936},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0a9d8b60cd82; sha256 b6115d6c5399e12daba33ab3c8fb93918faad35abff36d92daa1084b02b40f46},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01936. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Sparse Matrix-Dense Matrix Multiplication (SpMM) is a fundamental kernel in graph neural networks and scientific computing, yet its optimization on Ascend NPUs remains largely unexplored. The physically decoupled Cube/Vector architecture and single-threaded scalar instruction dispatch of modern NPUs introduce unique challenges. We present Ascend-SpMM, a novel and systematic framework tailored to the physically decoupled architecture of Ascend this http URL introduce a memory-centric cost model that quantifies hardware affinity and guides a conflict-bounded, densityaware task-mapping strategy. The strategy maps dense regions to the high-throughput Cube path and sparse residuals to the flexible Vector path, with a theoretical bound on cross-processor conflicts. To execute these workloads with high efficiency, we further develop highly customized sub-kernels. These kernels leverage hierarchy-aware pipelining and round-robin ring buffer-based implicit synchronization to effectively mask memory latency and reduce instruction pipeline stalls. Extensive experiments demonstrate that Ascend-SpMM achieves a geometric-mean speedup of 12.70-20.55X over the strongest evaluated NPU baseline for each matrix, accelerates end-to-end GNN training by up to 6.8X, and delivers performance competitive with state-of-the-art GPU implementations. The source code of Ascend-SpMM is publicly available at this https URL .

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC) ; Databases (cs.DB).
Pagina del lavoro: https://arxiv.org/abs/2611.01936 — 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 2026-11-03T18:10:22.824Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).