Skip to main content

GP-3000 GPU edge computing system

Cincoze, a professional manufacturer of embedded systems, has announced its new flagship GPU edge computing system, the GP-3000. Its crowning feature is an exclusive GPU Expansion Box that provides expansion for up to two high-end GPU graphics cards and creating a high-performance industrial-grade GPU computer. Brandon Chien, General Manager of Cincoze, stated that “We already know AI will drive innovation and expansion for industrial applications. The GP-3000 is Cincoze’s answer for intensive image processing and complex calculations, such as deep machine learning, autonomous driving, automated visual inspection, and mobile monitoring. As our latest flagship model, the GP-3000 multiplies edge computing efficiency, amplifies productivity and reliability, and accelerates AIoT automation.”

Extreme Computing Performance

The GP-3000’s extreme computing power starts with an 8th or 9th generation Intel® Xeon® or Core™ i3/i5/i7 (Coffee Lake and Coffee Lake-R) CPU, Intel® C246 chipset, and supports two sets of DDR4-2666 ECC/non-ECC SO-DIMM up to 64 GB and can support up to two 250 W high-end GPU graphics cards. With a total system power consumption of 720 W, it’s easy to meet and exceed high-efficiency application requirements. A precision heat dissipation and cooling design quickly wick away heat, keeping the focus squarely on the breathtaking performance of the GP-3000.

GPU Scalability

The GP-3000 has excellent high-efficiency edge computing capabilities and supports the installation of various GPU cards on the market. It can be expanded with up to two high-end full-length GPU cards (328 mm max.) using the exclusive GPU Expansion Box. The unique “Adjustable GPU Retainer” fixes the graphics card in place, providing extra protection against shaking loose in high-vibration environments. It has multiple built-in PCIe slots for various high-speed I/O cards or image capture cards for different applications. Future upgrades for higher-end or more GPU graphics cards are simple because you just replace the GPU Expansion Box. Multiple full-length card expansion of this caliber is truly a rare find.

Rich Application Functions

The GP-3000 redefines the standard for high-end GPU computers, with high-speed I/O and multiple functions. In addition to the standard five LAN ports and six USB 3.2 ports, the GP-3000 uses Cincoze’s exclusive CMI and CFM modular design, which offers expansion modules with eight Gigabit PoE, two USB 3.2, or dual 10 Gb/s LAN ports. Storage options include high-speed M.2 NVMe storage slots and four hot-swappable 2.5” HDD/SSD trays accessible through the front maintenance panel. Together, they meet large-capacity machine vision storage requirements and improve hard disk accessibility for convenient removal and replacement. The IGN module (power ignition sensing) can monitor the on-board battery voltage and set a delayed shutdown time to avoid damage to the system due to unstable current when starting or turning off the engine. This combination of diverse functions provides the flexibility to meet the requirements of different market applications.

Strong and Reliable

In the pursuit of ever-higher standards, the GP-3000 is MIL-STD-810G certified and promulgated by the US Department of Defense to qualify military equipment (pending). The GP-3000 features 9–48 VDC power input, is built for -40 to 70℃ temperature operation, and supports multiple installation methods, including wall, desktop, face-up, and 19”rack mount. The GP-3000 has E-mark and EN 50155 (EN 50121-3-2 only) certifications, so it is capable of withstanding the rigors of rail and vehicle applications as well as other harsh environments.

With computing power, scalability, functionality, and reliability baked into its design, the GP-3000 is the ultimate high-performance industrial-grade embedded GPU computer. The GP-3000 is the ideal platform for multiple verticals, including machine vision, automotive, image recognition, transportation, and any other application that requires high-performance GPU computing power.

Topics

Read more about:

Product, Embedded vision

Media Partners