We share the cloud tech topics you care about. Feel free to leave a message and tell us what you'd like to learn.
A joint effort by Baidu Baige and Fudan University introduces the LU‑KV framework, which treats KV‑cache budget allocation as a global combinatorial optimization problem, achieving only 0.52% relative performance loss at 80% compression and establishing a new efficiency‑accuracy SOTA on LongBench.
LoongForge applies system‑level optimizations—async data prefetch, fine‑grained communication‑compute overlap via a Megatron distributed optimizer, and per‑microbatch CUDA Graph scheduling—to the GR00T N1.6 Vision‑Language‑Action model, delivering up to 2.3× higher training throughput and a 56.6% reduction in overall training time on an 8×A800 cluster.
The article analyzes the challenges of large‑scale live VM migration, introduces Intel Xeon 6 CPU‑integrated QAT hardware acceleration, compares pre‑ and post‑QAT workflows, and reports a 60% reduction in migration time, 20% CPU savings, and sub‑10 ms downtime in Baidu Smart Cloud production.
The article analyzes the power‑supply challenges of high‑density AI data centers, compares traditional UPS and 800 V DC architectures, and shows how Baidu’s Hanhai U series redesign delivers precise capacity matching, up to 2.5× higher power density, 55% space reduction and up to 15% cost savings.
The article analyzes the architectural challenges of large‑model inference in the Agent era—such as memory‑intensive MLA structures, MoE communication overhead, exploding KV‑Cache size, and tool‑call accuracy—and presents a series of engineering solutions including hierarchical KV‑Cache pooling, sequence parallelism, offloading strategies, and chip‑level adaptations to achieve higher throughput and lower token costs.
The article presents a real‑world migration case where a CPU‑XPU bottleneck limited inference QPS, explains how Btune 2.0’s new latency‑focused diagnostics pinpointed a kernel lock contention in the halolet component, and shows the AI Agent’s automated, cross‑process analysis that restored performance and reduced cost.
The article details Baidu Baige’s end‑to‑end AI infrastructure for embodied intelligence, covering VLA and world‑model architectures, scaling challenges for medium‑sized models, cloud‑based motion‑control pipelines, open‑source integration, hardware‑aware training optimizations, and simulation‑engine improvements that together speed up model development and deployment.
LoongForge is an open‑source, Megatron‑based multimodal training framework that unifies LLM, VLM, VLA and diffusion models, runs seamlessly on NVIDIA GPUs and Baidu Kunlun XPU, and delivers 15%‑45% end‑to‑end training acceleration while scaling linearly on thousands of cards.
The open‑source GLM‑5.1 model, adapted to Baidu Baige's Kunlun XPU via the vLLM‑Kunlun Plugin, delivers record‑breaking SWE‑bench scores, eight‑hour autonomous coding, long‑context handling up to 64K tokens, and scalable deployment across tens of thousands of chips, showcasing end‑to‑end AI acceleration.
Baidu Intelligent Cloud’s seventh‑generation AI confidential virtual machine combines Intel TDX, NVIDIA GPUs, and BlueField DPUs to deliver end‑to‑end encrypted data paths, elastic multi‑GPU scaling, and near‑native performance, proving that high‑sensitivity AI workloads can run securely in the cloud without sacrificing speed.
