Congatec introduces the 12th generation of Intel Core processors on the COM-HPC Size-A and C modules »conga-HPC/cALS/P« and the COM-Express module »conga-TC670«. Client Size C modules are available with four different CPU versions, Client Size A and COM Express modules with three versions.
Particularly noteworthy is Intel's hybrid architecture with up to 14 cores/20 threads in BGA assembly and 16 cores/24 threads in the desktop variants (LGA assembly). IoT and edge applications benefit from up to 6 or 8 (BGA/LGA) optimized performance cores (P-cores) plus up to 8 low-power efficient cores (E-cores) and DDR5 memory support to accelerate multithreaded applications and run background tasks more efficiently. In addition, the BGA processors offer greatly improved graphics performance with up to 96 EUs of the integrated Intel Iris Xe GPU. In addition to high bandwidth and performance, Congatec's new flagship modules in COM-HPC client and COM Express Type 6 formats impress with dedicated AI engines that support Windows ML, Intel's distribution of the OpenVINO Toolkit, and Chrome Cross ML.
In addition, the modules offer increased bandwidth for connecting graphics processing units (GPUs) for maximum graphics and GPGPU-based AI performance. Compared to the BGA versions, these and all other peripherals benefit from doubled lane speeds as they also feature PCIe 5.0 in addition to PCIe 4.0. Furthermore, the desktop chipsets offer up to eight PCIe 3.0 lanes for additional connectivity. The mobile BGA versions also provide up to 16 PCIe 4.0 lanes via the CPU and up to eight PCIe 3.0 lanes via the chipset.
For example, use cases include edge computing and IoT gateways with multiple virtual machines for smart factories and process automation, AI-based quality inspection, and machine vision. Also, real-time collaborative robotics and autonomous logistics vehicles for warehousing and shipping. Typical outdoor applications include autonomous vehicles and mobile machines, video security and gateway applications in transportation and smart cities, and 5G cloudlets and edge devices that require AI-powered data processing.
The various AI workloads can be delegated to the P-cores, E-cores as well as the GPU execution units to handle even intensive edge AI workloads. Intel's deep learning boost technique leverages the various cores via Vector Neural Network Instructions (VNNI). In addition, the integrated graphics support AI-accelerated DP4a GPU instructions that even scale to dedicated GPUs. In addition, Intel Gaussian & Neural Accelerator 3.0 (Intel GNA 3.0) – which is Intel's integrated low-power AI accelerator - enables dynamic noise cancellation and speech recognition and can be activated in the processor's power-saving mode via a voice command.