Modern deep learning workloads are increasingly co-located on a single GPU to improve hardware utilization, yet existing MPS-based sharing mechanisms expose concurrency without providing QoS-aware control. Deadline-constrained training jobs can suffer severe slowdown when co-executed with best-effort workloads, because...
De-Sheng Wang, Fu-Zhen Yong, Si-Chao Chen et al.· ACM Transactions on Architec...· 0 citations
Cross-chain asset transfer protocols are vital for blockchain interoperability. While many focus on functionality and scalability, privacy is often overlooked. The Payment Channel Hub (PCH) addresses this by offering a centralized service that enables fast, private off-chain cross-chain transactions by obscuring user t...
Yuanming Shao, Yu-Ming Feng, Weizhe Zhang et al.· IEEE Transactions on Informa...· 0 citations
Parallax is presented, a fine-grained and accurate performance prediction framework tailored for training–inference co-execution and introduces interpretable modeling strategies for the two primary GPU sharing paradigms.
Zesen Hu, Desheng Wang, Sichao Chen et al.· ACM Transactions on Architec...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.