Gryf

AI Processing on the Move

In the ever-changing landscape of missions that demand real-time AI data analysis in the field, there arises a need for a novel supercomputer. This supercomputer must be compact enough to be transported on commercial aircraft, capable of handling extensive AI data processing, and secure for use in any environment.

Introducing Gryf, the Source Code at Your Service

The Gryf Edge AI system represents a robust, portable data center encased within a suitcase. Engineered with cutting-edge technology, it boasts effortless deployment and management. Ideal for military, government, and public safety operations, Gryf serves as the ultimate solution.

Durable and Compact

Gryf embodies a portable TSA-standard fly-away kit that’s been engineered for mobility. Encased in a lightweight and resilient carbon fiber housing, it includes removable end caps equipped with a telescoping handle and wheels designed to reduce vibrations, ensuring swift and easy deployment.

Adaptable, Customizable, and Versatile

Tailor-Made for Composable Disaggregated Infrastructure (CDI), Gryf boasts an innovative design that allows each sled or server to independently access the resources of any other sled, offering an impressive 1.5PB of storage and GPU capacity per unit, ideal for managing extensive air-gapped data sets. With the potential to interconnect up to 5 Gryf units, it offers nearly limitless configurability, featuring 30 resource slots.

Gryf brings the capabilities of a supercomputer into a compact, lightweight, and energy-efficient form factor, optimized for rapid deployment and adaptable configurations in challenging operational scenarios. Furthermore, it can be dynamically reconfigured to align with evolving mission demands, granting operators the immediate capacity to construct a highly modular computing platform through its high-density servers, flash-based storage, and accelerator options.

A Grab-and-Go AI Platform
  • Modular AI-Ready Edge Computing System
  • Composable compute, GPUs, storage and network
  • Low SWaP – TSA carry-on sized
  • All NVMe storage sleds are removable for portability
  • 6x Sleds per Chassis
    • Up to 64-core CPU compute server sled
    • Up to 8x NVMe SSDs per drive sled – up to 245TB per sled
    • GPU/FPGA sled – up to 350w
    • Integrated chassis ethernet networking sled
  • Up to 5x Interconnected Chassis
    • Full composability across all 5x Chassis (30 sleds)
    • Up to 8PB NVMe storage to a single compute sled
Made in USA with Standard 1-Year Warranty

Source Code builds all systems in our US-based manufacturing facilities. We inspect and test each system to ensure it is 100% operational and optimized to support rapid deployment.

Source Code offers a comprehensive 1-year warranty with each system. Optional two-year or three-year warranties are also available.

Learn More

© 2024 Aurified Consulting. All Rights Reserved.