Tag: AI Hypercomputer

AI Hypercomputer Generative AI Official Blog June 9, 2025

Accelerate your gen AI: Deploy Llama4 & DeepSeek on AI Hypercomputer with new recipes - Learn about new recipes on GitHub for deploying the latest Llama4 and DeepSeek models on the AI Hypercomputer platform.

AI Hypercomputer LLM Official Blog May 26, 2025

Introducing the next generation of AI inference, powered by llm-d - We’re making inference easier and more cost-effective with llm-d, an open-source, Kubernetes-native, distributed and disaggregated inference platform.

AI Hypercomputer Official Blog May 19, 2025

AI Hypercomputer developer experience enhancements from Q1 25: build faster, scale bigger - The article discusses recent enhancements to Google Cloud's AI Hypercomputer, designed to improve the AI developer experience. These enhancements include Pathways on Cloud for interactive scaling, Xprofiler for performance analysis, optimized container images for popular AI frameworks, and recipes for boosting GPU training efficiency.

AI Hypercomputer LLM Official Blog May 12, 2025

From LLMs to image generation: Accelerate inference workloads with AI Hypercomputer - Google Cloud is enhancing its AI Hypercomputer with new inference capabilities, including the Ironwood TPU, vLLM support for TPUs, and GKE Inference Gateway and Quickstart. JetStream, Google's inference engine, now integrates Pathways for lower latency and supports multi-host inference, while MaxDiffusion delivers improved image generation performance on TPUs. MLPerf™ Inference v5.0 results highlight the powerful inference performance of A3 Ultra (NVIDIA H200) and A4 (NVIDIA HGX B200) VMs.

AI AI Hypercomputer Official Blog April 14, 2025

High performance storage innovations for your AI workloads - Google Cloud introduces high-performance storage innovations to optimize AI workloads. Rapid Storage offers sub-millisecond latency and high throughput, while Anywhere Cache improves read-storage latency by up to 70%. Google Cloud Managed Lustre provides a fully managed parallel file system with sub-millisecond latency and high throughput. Storage Intelligence analyzes object metadata to generate storage insights and optimize costs.

AI Hypercomputer Official Blog TPU April 14, 2025

Introducing Ironwood TPUs and new innovations in AI Hypercomputer - Google Cloud introduces new innovations in AI Hypercomputer, including Ironwood TPUs, enhanced networking, and software capabilities for training and inference. Ironwood TPUs offer 5x more peak compute capacity and 6x the high-bandwidth memory capacity compared to the previous generation.

AI Hypercomputer Official Blog March 10, 2025

Guide: Our top four AI Hypercomputer use cases, reference architectures and tutorials - AI Hypercomputer, a fully integrated supercomputing architecture for AI workloads, offers various use cases with reference architectures and tutorials. It enables affordable inference with JAX, GKE, and NVIDIA Triton Inference Server, especially when paired with Spot VMs for significant cost savings.