The customer is considering introducing an image inspection system that uses AI, and the general-purpose PC does not have sufficient processing capacity, so we asked for a custom proposal. After testing with the prototype, it is planned to be introduced to the factory one after another.
Among the proposals from multiple options including the Xeon and Core X series, we finally adopted the configuration based on Core i9 that emphasizes high-speed processing with a single core.
The number of PCI-E support lanes of the CPU in this example is 16, and when multiple GPUs are installed, the PCI-E speed is x 1 per GPU (* 8), but even if you subtract that, GPU1 Since load balancing by inserting multiple cards was considered to be more effective, we confirmed that the PCI-E speed would be x2 operation, and then adopted the configuration with two GPUs inserted.
(* 1. The maximum number of PCI-E lanes that can be managed as a system is determined by the PCI-E controller built into the CPU.)
【Main Specifications】
CPU | Core i9 9900K (3.60GHz 8 core * TB time 5.00GHz) |
memory | 64GB (16GBx4)) |
storage | System: SSD 1TB high speed model (M.2) + data: HDD 2TB (S-ATA) |
video | NVIDIA Quadro RTX 6000 48GB x 2 * PCI-E operates x 8 |
network | Gigabit LAN x 1 |
Housing + power supply | Middle tower housing (width 233 x height 465 x depth 543 mm) + 1000W |
OS | Windows 10 Professional 64bit |
display | 21.5-inch wide display (DP connection possible) |
■ Click here for details and inquiries about this PC case
Example of a workstation for AI image inspection and analysis |