Computer vision systems are all around us today and play an essential role in keeping us safe at the edge in remote and mobile environments. Computer vision systems are found in a wide range of uses across all industry sectors, for example:
- Public transportation: from improved passenger experiences to predictive maintenance
- Automotive: enhanced driving capabilities from self-parking to autonomous vehicles
- Government and public safety: faster decision making and real-time data e.g. video data from the variety of sensors (UAV/UAS, fixed assets, satellite, etc)¹
- Heavy Industry: from autonomous operations to material handling
Building trust in AI/ML in harsh conditions
In all the uses listed above, the GPU is in an environment outside the Goldilocks environment of a server room, in which the climate is stable and controlled. To truly trust the accuracy of the data model and AI inference outcomes, the GPU must continue to operate even in the harshest conditions under extreme temperatures.
Life on the road requires rugged GPU devices explicitly designed to withstand the variable power, shock, vibrations, and temperatures of being in a vehicle.
Designed rugged to TrueTactical™ standards², the VoyagerVM 4.0 with the 10th generation of Intel Xeon D compute (10 cores) and snap-on expansion GPU module (KLAS-EXP-GPU-A) with an Nvidia Ampere A4500 (5888 CUDA® cores, 46 RT Cores, and 184 Tensor Cores) delivers the optimal compute performance per watt in a small, rugged form factor, making it ideal for life on the road.
Future-proofed data acquisition and inference modeling
The combined VoyagerVM 4.0 and Nvidia GPU expansion module delivers the highest networking capacity (4 x 25GbE and 2 x 25GbE) in its size, making it easier to connect to a more significant number of high-bandwidth sensors e.g. 10GbE cameras as part of a vehicle compute vision system.
The versatile Voyager Nvidia GPU expansion module permits users in all industry sectors to innovate on multiple levels—from image processing and analysis to accelerated processing of complex sonar/radar systems to system training and inferencing of AI engines, all while on the road in the real world.
Maximizing GPU utilization
GPUs are expensive, and in an ideal world, they need to be 100% utilized. Unlike traditional OEM offerings, the simple and straightforward snap-on capability of the VoyagerVM 4.0 Nvidia GPU expansion module means the GPU is a transferable resource that can be used as and when required in the field.
The ability to share and move GPU resources maximizes GPU utilization, ensures organizations do not overspend unnecessarily, and thus gets a greater return on their investment at the edge.
Learn more
To learn more about VoyagerVM 4.0 and expansion modules check out our blogs, datasheets and video overviews:
Blogs:
- Maximize what the edge can do for you! – https://www.klasgroup.com/maximize-what-the-edge-can-do-for-you/
Datasheets and Videos:
- VoyagerGPU 4.0 – https://www.klasgroup.com/products/voyagergpu-4-0/
References:
- Insight.Tech “Mobile Edge Servers Empower First Responders” – https://www.insight.tech/industry/mobile-edge-servers-empower-first-responders
