Local deployment of MoE-based LLMs on CPU-GPU nodes is attractive for cost and privacy reasons, yet limited GPU memory forces most experts into CPU memory. Prefill is then compute-bound and benefits from aggressive GPU loading, while decode is I/O-bound and requires conservative PCIe transfers. These demands cannot be...
En-Da Yu, De-Zun Dong, Zhao-Ning Zhang et al.· 0 citations
This work proposes to decouple low latency from fault tolerance, and design a hybrid solution that seamlessly integrates client-coordinated two-phase commit (2PC) with consensus, respectively to reduce latency in normal situations and tolerate faults in faulty situations, and designs xRaft, which can adaptively switch...
Yan Chen, Xinyi Yu, Shengyun Liu et al.· Conference on Applications,...· 0 citations
This work develops a non-clairvoyant scheduling algorithm that does not know task work in advance, yet loses only a small additive term relative to the clairvoyant lower bound, and proves a performance upper bound under uniform random data placement.
H. Kang, Yiwei Zhao, Kunal Agrawal et al.· ACM Symposium on Parallelism...· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
