# H100 cluster Published articles for H100 cluster. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## vExpert: Virtualizing Expert Storage for Adaptive Load Balancing in Distributed MoE Inference DevFeed: [vExpert: Virtualizing Expert Storage for Adaptive Load Balancing in Distributed MoE Inference]() Original publisher: [Read original article]() Author: Wenxun Wang, Xiuhong Li, Yida Wang, Chen Tang, Zongle Huang, Ke Hong, Yu Wang, Yongpan Liu Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Distributed Computing]() Topics: [Mixture of Experts (MoE)](), [Multi-GPU](), [Load Balancing]() Tags: [adaptive](), [balancing](), [cs-dc](), [distributed](), [distributed-inference](), [dram](), [h100](), [h100-cluster](), [inference](), [latency](), [load-balancing](), [memory](), [moe](), [offloading](), [overhead](), [pcie](), [performance](), [ratio](), [routing](), [runtime](), [storage]() ### AI overview vExpert addresses dynamic load imbalance in distributed MoE inference by using host DRAM for full redundancy of virtualized expert storage. Its performance model jointly accounts for host-to-device transfer latency, shared PCIe contention, and load imbalance; a lightweight allocator maps physical experts to virtual slots at runtime, while a disaggregated manager reduces memory overhead across processes. In an evaluation on DeepSeek-V3.1 using an 8-node H100 cluster, the authors report up to 55% improvement in load balance ratio and 1.2x end-to-end speedup. The supplied text is an abstract, so it does not provide further evaluation details. ### Source excerpt arXiv:2611.06057v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models have become prevailing, yet their distributed deployment suffers from severe dynamic load imbalance due to the conflict between static expert placement and inherent routing dynamism. While host DRAM offers capacity to avoid static device storage for load balancing, we identify existing offloading works fundamentally fail in distributed inference due to two overlooked challenges: (1) coupled performance impacts of runtime Host-to-Device transfers due to load dynamism and shared PCIe topology; (2) system-level memory overhead incurred by DRAM usage at scale. In this paper, we propose vExpert that shifts DRAM offloading paradigm from partial residency to full redundancy through virtualized expert storage. We pioneer the first systematic performance model that jointly captures H2D latency, PCIe contention, and load imbalance, building a lightweight allocator to dynamically map physical experts to virtual slots runtime. A novel disaggregated manager is designed to reduce memory overhead in multi-process environments. Evaluated on DeepSeek-V3.1 within an 8-node H100 cluster, vExpert improves load balance ratio by up to 55% and achieves 1.2$\times$ end-to-end speedup. ## IBM and CoreWeave co-design controls for agent workloads DevFeed: [IBM and CoreWeave co-design controls for agent workloads]() Original publisher: [Read original article]() Author: Victoria Gayton Published: 2026-10-02T20:12:25Z Content type: news Language: en Sources: [SiliconANGLE]() Topics: [AI Infrastructure](), [DevOps](), [AI software supply chain security](), [AI Platform](), [Software Engineering]() Tags: [agent-workload-isolation](), [agentic-ai](), [ai](), [ai-training-and-inference](), [architecture](), [aws](), [benchmark](), [brian-belgodere](), [computing](), [coreweave](), [coreweave-sandboxes](), [cube-event-coverage](), [cybersecurity](), [data-provenance](), [dave-vellante](), [fully-connected-2026](), [fullyconnected](), [granite](), [h100-cluster](), [ibm](), [identity-management](), [infrastructure](), [isolation](), [john-furrier](), [model](), [model-development](), [models](), [news](), [nyse-ibm](), [research](), [thecube]() ### AI overview IBM and CoreWeave are working together on identity management and controls for agent workloads. The article describes IBM Research's shift from model training to executing and testing agent code, including isolated execution through CoreWeave Sandboxes and a managed serverless runtime. It also covers infrastructure planning and measuring the performance impact of security controls. ### Source excerpt Agent workload isolation is becoming a practical infrastructure challenge as research teams move beyond training models to running code and testing agents. Systems built for intensive computation must now accommodate workloads that interact with tools, storage and other services. IBM Research's infrastructure must support increasingly varied workloads as its model development practices evolve. Reinforcement learning [...] The post IBM and CoreWeave co-design controls for agent workloads appeared first on SiliconANGLE. ## EU's AI factory plan requires proposals while Lambda gives you 8x H100s in 5 minutes DevFeed: [EU's AI factory plan requires proposals while Lambda gives you 8x H100s in 5 minutes]() Original publisher: [Read original article]() Author: Pieter Levels Published: 2025-10-24T00:00:00Z Content type: opinion Language: en Sources: [@levelsio (Pieter Levels)]() Topics: [AI Factory](), [DGX Cloud](), [Finance](), [X (Twitter)]() Tags: [ai](), [ai-factory](), [ai-startups](), [business](), [china](), [cluster](), [entrepreneurship](), [eu](), [eu-acc](), [europe](), [filesystem](), [founders](), [funding](), [gpu](), [gpus](), [h100-cluster](), [hetzner](), [hosting](), [inference](), [infrastructure](), [jupyter](), [jupyter-notebook](), [lambda](), [tech]() ### AI overview The author says Lambda let them select a location and filesystem, launch an 8x H100 server for $24 per hour, and access a running cluster with Jupyter and terminal access in about five minutes, without a proposal application. They contrast this with the EU AI Factory plan, which they say requires a proposal and may involve months of waiting. As an alternative, they suggest subsidizing GPU rental through existing European hosting companies, estimating that funding could cut the cited hourly price to $12. They acknowledge they personally oppose market intervention, and argue that EU technology policy should listen more to founders and technical experts. These are the author's claims and proposal; the excerpt does not establish the EU plan's full requirements or provider availability. ### Source excerpt 💻 I just wanted to show how easy it is getting a GPU in the regular way compared to the 🇪🇺 EU's "AI Factory" plan where you have to apply for a proposal Funnil...