Test and performance measurement vendor Keysight Technologies has developed Keysight Artificial Intelligence (KAI) to identify performance inhibitors affecting large GPU deployments. It emulates workload profiles, rather than using actual resources, to pinpoint performance bottlenecks.
Scaling AI data centers requires testing throughout the design and build process – every chip, cable, interconnect, switch, server, and GPU needs to be validated, Keysight says. From the physical layer through the application layer, KAI is designed to identify weak links that degrade the performance of AI data centers, and it validates and optimizes system-level performance for optimal scaling and throughput.
AI providers, semiconductor fabricators, and network equipment manufacturers can use KAI to accelerate design, development, deployment, and operations by pinpointing performance issues before deploying in production.
The Keysight AI architecture includes the newly announced KAI Data Center Builder for emulating large-scale AI workloads, and it features four product suites that address AI data center design from pre-silicon simulation through post-deployment system testing and troubleshooting. The suites are: KAI Compute for optimizing digital designs next-generation AI chip development; KAI Interconnect to validate optical and electrical data paths; KAI Network to benchmark AI network performance; and KAI Power to optimize power efficiency and energy management across data center components.
“Scaling AI data centers requires more than component-level validation. Interoperability, performance, and efficiency are system-wide metrics that can only be measured under real-world network conditions,” said Ram Periakaruppan, vice president and general manager of network test and security solutions at Keysight, in a statement. “Keysight’s AI solutions integrate our deep experience in traffic emulation, component, and network compliance validation, and the latest industry standards to emulate every aspect of data center performance: compute, network, interconnect, and power to ensure AI infrastructure meets evolving demands.”
Source:: Network World