Xplorer X1600P PCIe Accelerator Platform

 Accelerate low-latency edge AI applications with 30x higher system-level efficiency vs. CPU/GPU/VPU solutions

The Blaize Xplorer X1600P PCIe accelerator platform provides a 30x leap in system-level efficiency to accelerate complete AI applications. X1600P is based on El Cano, the second generation of the revolutionary Blaize Graph Streaming Processor (GSP™) architecture that enables new levels of processing power with energy efficiency ideal for AI inferencing workloads at the edge. The half-height, half-width X1600P accelerator conveniently plugs into any standard PCIe slot. With low power, 100x lower memory bandwidth, low latency, and smaller footprint, the X1600P opens doors to previously unfeasible AI inferencing solutions across a range of edge smart vision use cases.

Programmability to Build Complete AI Apps, Keep Pace with Rapid Evolution of AI Models

The X1600P, like all Blaize GSP-based systems, is fully programmable via the Blaize Picasso Software Development Kit. The hardware and

software are purpose-built to enable developers to build entire AI applications optimized for edge deployment constraints that can run

efficiently in a complete streaming fashion, as well as continuously iterate to keep up with rapid evolutions in neural networks.

Applications

• Smart Retail
• Smart City
• Smart Manufacturing
• Automotive (Industrial)
• Robotics
• Security
• Smart Vision
• Edge Inference server
• Internet of Things

 Features

• 16-core El Cano GSP: 16 TOPS
• Power: 7W Typical use-case
• GSP accelerated programmable ISP
• 2 GB, 4 GB or 8 GB, dual-channel, 32-bit wide LPDDR4
• PCIe Gen 3.0, 4 lanes