As policies like "AI+" Action Plan take effect, the adoption of generative AI models across industries is accelerating, driving exponential demand. Developers are seeking efficient solutions to design, train, and deploy large models that integrate AI into products and services, advancing the commercialization and accessibility of generative AI.
JWIPC E608: Generative AI Super Edge Computing
The JWIPC E608 leverages the NVIDIA Jetson Orin™ Nano platform with industrial-grade design, providing cutting-edge inference capabilities for robotics, multimodal intelligent products, and large-scale industry model deployment. Starting January 2025, JWIPC introduces a “Super” upgrade with JetPack 6.1, delivering enhanced performance and cost-efficiency for generative AI applications.
Performance Breakthroughs
· AI Performance Boost: AI capability increases from 40 TOPS to 67 Sparse TOPS, a 1.7x improvement, enabling the latest generative AI inference tasks.
· Enhanced Memory Bandwidth: Upgraded from 68GB/s to 102GB/s, a 1.5x boost, optimizing high-performance generative AI computations.
· Higher CPU Frequencies: Frequency rises from 1.5GHz to 1.7GHz, a 14% increase, ensuring efficient operation of diverse generative AI models.
Seamless Upgrade for Existing Users
Current JWIPC E608 users can enjoy the 1.7x performance increase through a simple software upgrade without the need for hardware replacement.
Versatile Applications
· Robotics & Logistics: Supports emerging applications of generative AI workloads in robotics and multimodal intelligent agents.
· Computer Vision: Effortlessly handles complex vision tasks, from object detection to image segmentation.
· Embodied Intelligence & Generative AI Applications: Seamlessly runs various Transformer models, from LLMs to ViTs, empowering edge devices with advanced AI capabilities.
With the E608 “Super” upgrade, JWIPC continues to lead the evolution of generative AI at the edge, driving the transformation of industry AI models from concepts to real-world deployment.