TECHNOLOGY OVERVIEW
AI Architecture that Accelerates the Future
GTI’s architecture offers a unique way to design and develop AI hardware. Our suite of accelerator chips and boards enable companies to develop advanced AI capabilities, from Edge to Cloud. Our solutions bring the power of cloud Artificial Intelligence to local devices and improves Cloud AI performance with greater performance and efficiency, providing the utmost in AI customization for new solutions.
Domain-Specific Architecture
The Matrix Processing Engine (MPE™) is a two-dimensional array of digital multiply-add (MAC) units that computes the series of layer operations of a convolutional neural network. The scalable matrix design allows each engine to directly communicate and interact with adjacent engines, optimizing and accelerating data flow.
Clustered computing elements perform inference without having to go off-chip to DRAM.
APiM increases computing efficiency in a revolutionary way
In Memory computing
AI Processing in Memory (APiM) handles AI computing in a revolutionary way by including localized memory that holds the weights and activations for the neural network. APiM eliminates the bottleneck of massive data movement, which increases performance, decreases power consumption and lowers cost.
patent portfolio
CNN Architecture
CNN Data Structure
Memory Architecture
Motion Recognition
Natural Language Processing
Hierarchical Classification
AI Inference Computing Device
Natural Language Translation
publications
White Paper
Intelligent Transportation
Gain insights into the Intelligent Transportation landscape, industry issues, and the direction of the autonomous vehicle market.
White Paper
Accelerating AI-Defined Cars
Gain insights into the automotive landscape, industry issues, and the direction of the autonomous vehicle market.
Frost & Sullivan Report
New Product Innovation Award
Gyrfalcon was recognized in 2019 by leading global research organization Frost & Sullivan for Innovative AI Accelerator Chips.
Research Papers
10+ Paper Submissions Accepted
Gyrfalcon’s research papers have been presented at prestigious technical conferences around the world.
video coding for machines (VCM)
GTI has been active in the MPEG standards and has been recognized by industry analysts for technology innovations leveraging AI and deep learning.
Optimizing video data for machine use will be a key driver for commercializing new solutions and enabling mass adoption and innovation.
development & training tools
Gyrfalcon offers silicon-proven solutions. Register for access to our Developer Portal to gain free access to software, sample code, etc. Purchase development hardware and begin evaluating and prototyping.
DEV PORTAL
Access software & resources to design, test and deploy high-performance AI models.
DEV HARDWARE
Connect our USB dongles and PCI boards to easily create AI enabled apps and solutions.