Metraipolis › archive › dc_8917d0105c45
Measuring and Reducing WebGPU Dispatch Overhead for LLM Inference — arXiv:2608.08730
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.08730
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:52:48.810Z
- SHA-256 of the archived text
0d5c12e971a1e54e3237e0affcef12dee2025448fc8d7a0df493c3bebae998a8- Stable URL of this page
https://metraipolis.com/documento/dc_8917d0105c45
How to cite
Measuring and Reducing WebGPU Dispatch Overhead for LLM Inference — arXiv:2608.08730 — arXiv (en), https://arxiv.org/abs/2608.08730 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8917d0105c45 (Metraipolis archive dc_8917d0105c45, sha256 0d5c12e971a1e54e3237e0affcef12dee2025448fc8d7a0df493c3bebae998a8)
BibTeX
@misc{dc8917d0105c45,
title = {Measuring and Reducing WebGPU Dispatch Overhead for LLM Inference — arXiv:2608.08730},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.08730},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8917d0105c45; sha256 0d5c12e971a1e54e3237e0affcef12dee2025448fc8d7a0df493c3bebae998a8},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.08730. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large Language Models are deployed to multiple types of environments, from internet browsers to edge devices, and WebGPU serves as a modern cross-platform standard. The engines for browser-based LLM inference have proliferated, yet the overhead of WebGPU per-operation dispatch remains poorly characterized. In this work, we introduce a sequential-dispatch measurement method and show that naive single-operation measurements overestimate per-dispatch cost by conflating dispatch with synchronization. Using our method, we measure the per-dispatch cost and show that it is independent of data type used. We show that the dispatch overhead, not kernel quality, is the bottleneck at batch size 1, and isolate the dispatch count as the cause. Therefore, we conclude that at batch size 1, the effective approach to LLM inference optimization in WebGPU is reducing dispatch count. Our findings point to dispatch amortization, in the inference engines and in the WebGPU specification, as a path to practical browser-based inference. Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-08-11 in cs.DC). Categorie arXiv: Machine Learning (cs.LG) ; Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF). Pagina del lavoro: https://arxiv.org/abs/2608.08730 — 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 2026-10-19T08:50:23.781Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-08-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).