Metraipolis › archive › dc_bdc479529ccd

HeteroReason: Heterogeneous FPGA-GPU Acceleration for Disaggregated Speculative Reasoning — arXiv:2609.28717

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.28717
License (as archived)
CC0-1.0
Archived
2026-10-08T04:51:02.574Z
SHA-256 of the archived text
9938c03988ba1057dcb7a60cad0912c4bb95e236be94d57694dac2cb9176d4d1
Stable URL of this page
https://metraipolis.com/documento/dc_bdc479529ccd

How to cite

HeteroReason: Heterogeneous FPGA-GPU Acceleration for Disaggregated Speculative Reasoning — arXiv:2609.28717 — arXiv (en), https://arxiv.org/abs/2609.28717 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bdc479529ccd (Metraipolis archive dc_bdc479529ccd, sha256 9938c03988ba1057dcb7a60cad0912c4bb95e236be94d57694dac2cb9176d4d1)
BibTeX
@misc{dcbdc479529ccd,
  title = {HeteroReason: Heterogeneous FPGA-GPU Acceleration for Disaggregated Speculative Reasoning — arXiv:2609.28717},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.28717},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bdc479529ccd; sha256 9938c03988ba1057dcb7a60cad0912c4bb95e236be94d57694dac2cb9176d4d1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.28717. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Reasoning Models (LRMs) have achieved state-of-the-art performance in reasoning tasks by utilizing Chain-of-Thought (CoT) reasoning. To achieve fast execution speed, speculative reasoning techniques adopt a lightweight draft model for candidate token generation followed by process reward models (PRMs) for verification and a strong target model for refinements. This paper identifies that the existing speculative reasoning paradigm follows a strictly forward-only reasoning trajectory, which lacks robustness and can lead to severe error propagation if early reasoning steps are suboptimal. Furthermore, executing these disparate inference schemes, including sequential drafting and parallel verification on homogeneous GPU platforms, can lead to severe resource underutilization. To address this, we propose HeteroReason, an algorithm-hardware co-designed heterogeneous FPGA-GPU inference paradigm specifically tailored for LRM speculative reasoning. At the algorithmic level, we introduce a backtracking-enhanced workflow that enables the system to recover from low-quality states and explore alternative reasoning trajectories, significantly improving reasoning robustness. At the system level, the draft model is offloaded to the FPGA while deploying the PRM and target models on GPUs. A specialized workflow is optimized to achieve prefill-decode disaggregation, which exploits shadow synchronization to overlap GPU-side refinements with FPGA-side token updates to effectively hide synchronization latency. To mitigate inherent sequential constraints, we propose a step-ahead speculation and refinement scheduling scheme, transitioning the system from a sequential execution scheme to a parallel pipeline. Experimental evaluations show an average 4.2% accuracy improvement, with 1.01x-1.42x latency speedups and 1.25x-1.57x improvements in energy efficiency compared to homogeneous GPU baselines.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AR).
Categorie arXiv: Hardware Architecture (cs.AR).
Pagina del lavoro: https://arxiv.org/abs/2609.28717 — 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 2026-10-08T04:28:37.286Z dalla pagina https://arxiv.org/catchup/cs.AR/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).