Setup and manage your in-house servers and computing infrastructure
Gaudi brings to data centers a new level of AI compute performance and efficiency with massive scale-out capabilities. Optimizing your on-premise AI environment is easier than ever with tools and information on hardware identification and working condition of your hardware. To get started with system setup, review the documents, guides, and leverage the tools for system setup and management. Learn how to measure PCI bandwidth, maintain constant and equal power load and run functional tests on hardware components.
Learn how to download and run Habana specific docker images, create custom docker images, and install bare metal drivers.
Qualify the usage and integration of Gaudi hardware platforms in your systemdesign with qualification tool (hl_qual) for Gaudi.
See environmental variables, common error codes and how to report an issue when training your model.