CORTEXA
← Browse
arxivcs.ARcs.AI2026-07-23

Hardware-Software Co-Design for Float16 On-Device Training on RISC-V Single-Core

Benjamin Hubinet, Pierre-Alain Moellic, Olivier Savry, Olivier Potin, Jean-Baptiste Rigaud

By leveraging standard RISC-V extensions, namely Zfh (scalar float16) and Zvfh (vector float16), this work proposes an open-source framework to enable complete on-device training on resource-constrained RISC-V single-core. Our approach allows memory footprint reduction by about 50% as compared to using float32 and with minimal model performance degradation. We also facilitate transfer learning and fine-tuning scenarios by incorporating layer-freezing capabilities. Our work builds onto AIfES, an open-source, modular and generic DNN training and inference framework for embedded systems that can be extended with custom hardware-specific functions. The benefits of float16 is further emphasized by outlining the low area overhead of Zfh on a RV64GC super-scalar out-of-order FPGA softcore (+1.15% LUT6 and +0.05% FF at 175MHz). Finally, we discuss the architecture of a Zvfh implementation within the same RISC-V core.

View free PDFSource page

Related papers

arxivcs.ARcs.AI2026-07-15

Kaleido: Algorithm-Hardware Co-Design for Video Diffusion Transformers by Exploiting Latent Space Correlations

Wenxuan Miao, Haosong Liu, Weiming Hu, Zihan Liu, Aiyue Chen, Jianlin Yu, et al.

Video diffusion transformers (vDiTs) generate high quality video but introduce extremely high compute cost due to the long diffusion timesteps and self attention computation. As diffusion timesteps are reduced, the computation cost of self attention becomes the dominant bottlenec…

View free PDFSource page
arxivcs.ARcs.AIcs.DC2026-07-28

MDTransformer: A Hardware-Software Co-Design of Mode-Division Photonic Transformer Accelerator with Inverse-Designed Coherent Crossbar

Solomon Micheal Serunjogi, Rachmad Vidya Wicaksana Putra, Ayat Taha, Muhammad Shafique, Mahmoud Rasras

Recently, photonic transformer accelerators (PTAs) have successfully achieved significant speedup and energy efficiency improvements over electronic accelerators for expediting Transformer inference. However, state-of-the-art rely on expensive multi-wavelength light generation an…

View free PDFSource page
arxivcs.ARcs.AI2026-06-26

Agentic Hardware Design as Repository-Level Code Evolution

Cunxi Yu, Chenhui Deng, Nathaniel Pinckney, Brucek Khailany

We present HORIZON, a self-evolving agent framework that treats hardware design as repository-level code evolution. A Markdown harness is compiled into a project pack containing domain knowledge, an executable evaluator, an acceptance predicate, and a git/runtime policy; a hands-…

View free PDFSource page
arxivcs.ARcs.AI2026-07-24

Sparse by Command: Task-Conditional Compute Skipping for Multi-Task Inference Accelerators

Afzal Ahmad, Gaoyu Mao, Shoubo Hu, Hui-Ling Zhen, Mingxuan Yuan, Xinyu Chen, et al.

Multi-task inference models share a single backbone across diverse tasks, yet execute identical computation regardless of which task is active - wasting energy and cycles on task-irrelevant operations. We observe that the task command, typically available before inference begins,…

View free PDFSource page
arxivcs.ARcs.AI2026-07-20

BRIM: Workload-Balanced Dual-Sided Bit-Serial Sparse Inference Accelerator

Varun Manjunath, Ruokai Yin, Donghyun Lee, Arkapravo Ghosh, Priyadarshini Panda

Bit-serial accelerators exploit bit-level sparsity to reduce DNN inference cost, but existing designs exploit sparsity on only one operand, bounding the speedup. Extending sparsity exploitation to both operands simultaneously yields compounding reductions in partial products but…

View free PDFSource page