The AIoM products were created with NVIDIA and Intel. The series offers a hardware optimization strategy to help developers address performance, and SWaP requirements.
Heterogeneous computing is supported by the integration of one or more types of processing cores.
ADLINK’s AIoM offering includes:
Mobile PCI Express Module (MXM) GPU modules: The MXM GPU modules feature NVIDIA Quadro Embedded GPUs based on the Turing and Pascal architectures.
VPU-accelerated SMARC modules: Vizi-AI and Neuron Pi equipped with Intel Movidius Myriad X VPU. These enable developers to quicken the prototyping process. Commercially available options include tight version control and longevity support.
VPU-accelerated COM Express modules: High-performance modules that quickly integrate AI.
Additional form factors include PC/104, VPX, CompactPCI and XMC. Standards like USB3 Vision and GigE Vision are also supported.
The company will show the range at embedded world in a series of demonstrations that include an access control powered by ADLINK’s MXM GPU module and DLAP-3000-CFL platform, an inspection robot based on NVIDIA Jetson TX2, and a Vizi-AI development kit that simplifies scaling to other ADLINK AI products.
The integrated hardware/software approach provides flexibility. Developers can start on the low-cost Vizi-AI and choose a processor (e.g. CPU, GPU, VPU, TPU, NPU) at the time of deployment.