AI computing brings workstation AI to the edge
Bringing workstation-grade artificial intelligence (AI) computing to edge applications, Vecow’s ECX-2400 AI computing system uses the Blaize Graph Streaming Processor (GSP) -based Xplorer AI accelerator.
Vecow ECX-2400 AI is designed for smart automated optical inspection (AOI), public security, robotic control, traffic vision, and any industry 4.0 / artificial intelligence of things (AIoT) applications at the edge. In addition to low power and high reliability, it also results in less total cost of ownership (TCO), says Vecow.
The ECX-2400 AI is powered by workstation-grade 10-core 10th Gen Intel Xeon / Core i9 / i7 / i5 / i3 processor, and incorporates the advanced Blaize Xplorer X1600P PCIe series AI accelerator card supporting up to 80 T operations per second of AI inference. Despite its compact size, it is rugged and reliable for mission-critical AI accelerating tasks at the edge. It features smart system protection functions including 12 to 50V DC power input, 80V surge protection, software ignition power control and a 250W system power budget.
The Blaize software programmable GSP architecture provides a low latency and high- performance inference processing solution requiring less memory with fewer access, yielding a low latency AI system that consumes less power than other accelerators, says Vecow. “The Blaize software programmable GSP architecture excels in AI inference for edge-based products and appliances,” said Kev Wang, applied software manager at Vecow.
The Blaize Picasso software development kit (SDK) has been integrated into VHub AI developer software, enabling the users to easily develop, deploy and manage edge AI inference systems with single-core or quad-core Blaize Xplorer AI inference accelerators.
Barrie Mullins, senior director, product marketing at Blaize, commented: “Industrial customers using vision analytics in . . . industrial inspection, manufacturing production line anomaly, human robot interaction, employee safety and smart camera systems can utilise the ECX-2400 AI to advance operational efficiencies and processes.”
There are two versions available. The ECX-2400 AI-P is based on a 10th Gen Intel Xeon / Core i9 / i7 / i5 / i3 (Comet Lake) AI computing system with single core Blaize Xplorer AI accelerator and the ECX-2400 AI-PQ uses a quad-core Blaize Xplorer AI accelerator.
