GPU VPU Accelerated Computing
Certain imaging or machine vision applications require industrial PCs with high levels of processing and graphics power. Application examples include artificial intelligence (AI), video analytics, video processing, and autonomous vehicles. Industrial computers with graphics processing units (GPUs) have become a popular tool in both consumer and industrial computing and offer an efficient means for accelerating compute-intensive applications.
GPUs: Powerful Processing for Demanding Applications
Where central processing units (CPUs) feature 4, 8, or 16 cores, GPUs can have hundreds or thousands of small cores, allowing the devices to spread the workload over multiple cores, which enables parallel processing capabilities. GPUs have traditionally been used to render pixels (graphics) in video games. The better the GPU, the better the graphics quality and higher the frame rates. In image processing, a GPU performs the same function, but in reverse. In machine vision, for example, instead of rendering video games onto a screen with millions of pixels, the GPU processes millions of pixels down to help the imaging software interpret and understand images.
NVIDIA GPUs are popular today in both industrial and consumer devices. At the 2021 NVIDIA GTC conference for AI innovators, the company announced the launch of its RTX A5000 and RTX A4000 GPUs, which offer upgrades over previous models. For example, the Quadro RTX 4000, now called simply the RTX A4000, offers 16 GB GDDR6 memory, while the RTX A5000 has 24 GB GDDR6. Both GPUs offer more CUDA cores than prior Quadro and GeForce models, with the RTX A4000 featuring 6144 and the RTX A5000 featuring 8192. Another major upgrade for the two GPUs is memory, as both models improve upon the 264 GB/s of the RTX A3000 — RTX A4000 with 384 GB/s and RTX A5000 with 448 GB/s.
VPUs: Deploy Low-Cost, Low-Power Edge AI Applications
Vision processing units (VPUs), meanwhile, are microprocessors specifically geared toward image acquisition and interpretation. While VPUs are less powerful than GPUs, the low-cost, low-power devices deliver computational power for vision and AI at the edge. One of the first VPUs widely available on the market was released by Movidius, which has since been acquired by Intel. The Intel Movidius Myriad X VPU available today can interface with a CMOS image sensor, preprocess captured images, move the images through a pretrained neural network, and output results while consuming less than 1W of power.
Intel VPUs combine traditional CPU cores and 16 programmable, 128-bit VLIW vector processors that enable customers to run multiple concurrent imaging and vision application pipelines. This VPU also offers a neural compute engine — a dedicated on-chip accelerator designed to run neural networks at high speed and low power. With this engine, the Myriad X VPU delivers more than 1 trillion operations per second of DNN inference performance, meaning that customers can run convolutional neural networks for deep learning in real-time at the edge without compromising on power consumption or accuracy.
Additionally, the VPUs offer 16 configurable MIPI lanes, meaning that users can connect up to eight HD RGB cameras directly to the VPU with support for up to 7 MPixel per second of image signal processing throughput.
CoastIPC offers several different types of industrial PCs that either feature GPUs or offer expandability for GPUs or VPUs. This includes fanless, rugged, and compact models from leading companies such as Advantech and Neousys Technology. Products include processors from Intel, AMD, and NXP and feature options such as IEEE 1588 compatibility, multiple I/O options for cameras, IP67 design, and more. We can customize these industrial PCs with third-party cards or software so they’re ready for deployment out of the box.
CoastIPC carries several industrial PCs that feature NVIDIA’s Jetson edge AI platform. Models include NVIDIA Jetson AGX edge AI computers for intelligent video analytics, NVIDA Jetson Tegra X2 computers with 256 CUDA cores, NVIDIA Jetson Xavier NX computers, and NVIDIA Jetson Nano computers. Contact us today and learn how we can help solve your ITS computing challenges with NVIDIA Jetson technology.
GPU Computers and VPU Computers
CoastIPC offers several industrial PCs that can integrate GPUs or VPUs for compute-intensive applications. Whether you need a PC already equipped with a GPU, a PC with dual GPU slots, or a PC with expansion capabilities for VPUs, CoastIPC can help you build the right platform for your application. Products include fanless industrial PCs, NVIDIA GPU computers, video recording computers, compact embedded controllers, rugged computers, and more.