Metraipolis › archive › dc_3e2962b005cc

DWISE: Denoising Workload-Informed Scheduling Engine — arXiv:2610.39852

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.39852
License (as archived)
CC0-1.0
Archived
2026-11-01T02:25:15.522Z
SHA-256 of the archived text
20aa690321c851cabeb45757159a27688ae8023c98984c014797e02177b9c97a
Stable URL of this page
https://metraipolis.com/documento/dc_3e2962b005cc

How to cite

DWISE: Denoising Workload-Informed Scheduling Engine — arXiv:2610.39852 — arXiv (en), https://arxiv.org/abs/2610.39852 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3e2962b005cc (Metraipolis archive dc_3e2962b005cc, sha256 20aa690321c851cabeb45757159a27688ae8023c98984c014797e02177b9c97a)
BibTeX
@misc{dc3e2962b005cc,
  title = {DWISE: Denoising Workload-Informed Scheduling Engine — arXiv:2610.39852},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39852},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3e2962b005cc; sha256 20aa690321c851cabeb45757159a27688ae8023c98984c014797e02177b9c97a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39852. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Unlike autoregressive large language models (AR-LLMs), diffusion LLMs (dLLMs) generate text by iteratively denoising many tokens in parallel, offering a promising path to efficient inference. Existing serving frameworks, however, schedule dLLM requests with policies designed for AR models, namely first-come-first-served (FCFS) admission and prefill priority. We find that both policies are ill-suited to dLLMs, wasting much of their speed. FCFS ignores that a dLLM request's decoding cost is highly heterogeneous and unknown at arrival, causing severe head-of-line blocking. Prefill priority, widely used in AR-LLM serving to keep time-to-first-token low, sharply cuts dLLM efficiency, as rounds that run only prompts leave most of the batch idle. We present DWISE, a Denoising Workload-Informed Scheduling Engine for efficient dLLM serving. DWISE schedules requests by their remaining work, measured in a block-level unit shared by prefill and denoising, so both phases share a round. To predict the unknown denoising workload, DWISE combines a lightweight prompt-based predictor at admission with a single refinement from the first denoising round's hidden states, which capture model-specific information absent from the prompt and require no additional model execution. Implemented in SGLang and evaluated on five models from three dLLM families under bursty workloads, DWISE reduces mean end-to-end latency over SGLang by 5.4$\times$ on average and up to 14.9$\times$ while preserving task accuracy. The code will be released publicly.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.39852 — 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 2026-11-01T02:12:44.190Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).