Tag
IT vendor Core 4 Solutions listed a decommissioned 300-rack AI supercomputer on eBay for $59.7 million, featuring 9,600 Nvidia A100 GPUs and 4,200 AMD EPYC Rome CPUs, with free shipping available via FedEx Ground.
Ahmad 分享了如何通过 Mikrotik CRS804-4DDQ 1.6Tbps 交换机，将 4 条 400G QSFP-DD 拆分为 2x200G QSFP56，连接 8 台配备 ConnectX-7 的 DGX Spark，实现 1.6Tbps 全速互联的集群方案。
DGX Monarch是個人開發的開源專案，讓使用者能在雙DGX Spark叢集上透過ComfyUI進行圖像/影片生成，達到近2倍的渲染速度提升，支援BF16/FP8/NVFP4/INT8等多種精度。作者同時分享了使用AI輔助開發此專案所耗費的各平台訂閱成本。
NVIDIA 发布 AICR v1.0，提供开源、稳定且可验证的 GPU 加速 Kubernetes 集群配置方案，通过版本锁定的配方和签名验证证据简化跨组件兼容性管理。
A DIY inference machine built with dual RTX 3090 GPUs, a ThinkPad motherboard, and a VW Golf radiator, capable of running AI models like Qwen3.8-27B via vLLM.
The author shares their experience building a high-VRAM AI PC with multiple GPUs to run local models for software engineering, detailing solutions to hardware challenges like PSU limitations and GPU mounting.
A technical explanation of how LLM inference requests are routed when they hit a GPU cluster in production environments.
Explored NVIDIA Dynamo, a tool for deploying LLMs across multiple GPU cluster nodes with features like model caching, autoscaling, multinode deployments, and Kubernetes integration.
Advice on building a high-end AI workstation with 8x RTX PRO 6000 GPUs, emphasizing proper infrastructure, cooling, and avoiding reuse of DDR4.
ZCube is a new network architecture that flattens the topology and mixes single/multi-rail access to optimize KV Cache transmission in long-context and PD separation scenarios. In the GLM-5.1 production cluster, it achieved a 33% reduction in switch/optical module costs, a 15% increase in GPU inference throughput, and a 40.6% decrease in TTFT P99.
This paper presents ZCube, a novel network architecture developed by Z.ai, Harnets.AI, and Tsinghua University to address topology-induced congestion in Prefill-Decode disaggregated LLM inference clusters. Production deployments on GLM-5.1 coding workloads achieved a 33% reduction in network CapEx, 15% throughput improvement, and 40.6% reduction in TTFT P99 latency.
Jane Street allowed Dwarkesh Patel to tour their new Texas data center with 4,032 GPUs, each rack pulling 140 kilowatts, highlighting the massive scale and unique networking choices.
Jane Street revealed inside views of its AI training center in Texas, housing 4,032 GPUs, 8,000 kilometers of fiber optics, and a full liquid cooling system, while recounting the 20-year evolution from a humble start with six Dell hosts to today's extreme trading system.
OpenAI is launching Stargate UK, an AI infrastructure partnership with NVIDIA and Nscale to build sovereign compute capabilities in the United Kingdom, with plans for up to 31,000 GPUs over time. The initiative supports the UK's national AI strategy and will enable OpenAI models to run on local UK compute for critical public services, regulated industries, and national security use cases.