Metraipolis › archive › dc_1acaacda61da

Dissecting How Die Scaling Breaks GPU Fine-grained Scheduling — arXiv:2609.24270

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.24270
License (as archived)
CC0-1.0
Archived
2026-10-09T06:53:12.597Z
SHA-256 of the archived text
ab97610936ab9ee202958ad663f3c18c5e0cd8d7d2d717b416c75f0954b49e16
Stable URL of this page
https://metraipolis.com/documento/dc_1acaacda61da

How to cite

Dissecting How Die Scaling Breaks GPU Fine-grained Scheduling — arXiv:2609.24270 — arXiv (en), https://arxiv.org/abs/2609.24270 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1acaacda61da (Metraipolis archive dc_1acaacda61da, sha256 ab97610936ab9ee202958ad663f3c18c5e0cd8d7d2d717b416c75f0954b49e16)
BibTeX
@misc{dc1acaacda61da,
  title = {Dissecting How Die Scaling Breaks GPU Fine-grained Scheduling — arXiv:2609.24270},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.24270},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1acaacda61da; sha256 ab97610936ab9ee202958ad663f3c18c5e0cd8d7d2d717b416c75f0954b49e16},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24270. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Modern GPUs are no longer physically symmetric. Die scaling leads to both manufacturing-driven floorsweeping and cache and memory partitioning. The former creates chip-specific compute topologies, while the latter causes non-uniform memory access. These asymmetries are substantial. Topology-oblivious compute unit allocation can lead to up to 1.33x performance variation, while remote accesses increase HBM latency by up to 67% and nearly double L2 latency. However, these asymmetries are hidden behind the GPU's logical resource abstractions and can vary across chips. We develop lightweight characterization methods to uncover per-chip compute topology and memory affinity. We then use the discovered information to make existing fine-grained scheduling asymmetry-aware, considering not only how many resources are allocated but also which physical resources are assigned. Across full-GPU kernel execution, intra-application multiplexing, and inter-application co-location, asymmetry-aware scheduling improves mainstream kernels by up to 1.22x, multiplexed LLM inference by up to 14.3%, and avoids up to 1.33x performance variation.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-22 in cs.DC).
Categorie arXiv: Hardware Architecture (cs.AR) ; Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2609.24270 — 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 2026-10-09T06:50:13.566Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).