Conquest Introduces COM-HPC and COM Express Computer Modules

congatec, the world's leading supplier of embedded modules and edge computing technology, is pleased to announce the launch of 10 new COM-HPC and COM Express computer modules based on 12th Generation Intel Core mobile and desktop processors (codenamed Alder Lake). Featuring Intel's latest high-performance cores, the new COM-HPC (sizes A and C) and COM Express Type 6 modules dramatically increase and improve the performance of embedded and edge computing systems. Most notably featuring Intel's innovative high-performance hybrid architecture that delivers up to 14 cores/20 threads on the BGA package size version and 16 cores/24 threads on the desktop version (LGA package size), the performance jumps dramatically by providing multitasking and scalability for next-generation IoT and edge applications. [1] Benefit from up to six or eight (BGA/LGA) performance-optimised cores (P cores) with up to eight low-power efficiency cores (E cores) and DDR5 memory support to accelerate multi-threaded applications and perform background tasks more efficiently.

Additionally, the 12th generation mobile BGA processor integrates up to 96 Intel Torch Xe GPU execution units for an estimated 129 per cent[2] improvement in graphics performance compared to 11th generation Intel Core processors, providing users with immersive experiences and faster processing of parallel workloads such as artificial intelligence (AI) algorithms.

For maximum performance for embedded customer use, the LGA processor-based modules deliver a 94 percent increase in graphics performance, nearly tripling their image classification inference performance and increasing throughput by 181 percent.[3] Additionally, these modules deliver up to 129 percent[4] in graphics performance for embedded customer use. [3] In addition, these modules provide a high-bandwidth, high-rate bus to communicate directly with the GPU for maximum graphics and GPGPU-based AI performance. Compared to the BGA version, these modules and other peripherals can provide double the channel speed between them, thanks to the ultra-fast PCIe 5.0 interface technology, which also leads to PCIe 4.0 lanes from the processor. In addition, desktop chipsets can provide up to 8x PCIe 3.0 lanes for additional connectivity, while the mobile BGA version provides 16x PCIe 4.0 lanes directly from the CPU and 8x PCIe 3.0 lanes from the chip.

These modules are used in a wide range of applications and are targeted at all places where high-end embedded and edge computing technology is required, such as edge computers and IoT gateways for smart factories and process automation with multiple virtual machines, AI-based quality inspection and industrial vision, real-time collaborative robotics, and automated logistic vehicles for warehouses and transport. Typical outdoor applications include self-driving vehicles and mobile devices, video security and gateway applications in transport and smart cities, as well as 5G cloud devices and edge computing that require AI computing.

‘Intel Thread Director leverages Intel's innovative Performance Hybrid Architecture, which combines powerful P-core performance with energy-efficient E-cores, assigning each work task to the right core for optimal performance. Selected processors are also suitable for Intel TCC and Time Sensitive Networking (TSN) hard real-time applications. They fully support Real-Time Systems' Hypervisor technology, an ideal platform for consolidating many different tasks on a single edge platform. As this applies equally to low-power and high-performance scenarios, highly sustainable designs can be achieved with this small ecological footprint.’ Christian Eder, Director of Marketing at Conquest, explains.

In addition to the highest bandwidth and performance, the new flagship COM-HPC Client (Client) and COM Express Type 6 modules are equipped with a dedicated AI engine supporting Windows ML, Intel Distribution of OpenVINO tool suite and Chrome Cross ML. Different AI workloads can be seamlessly delegated to P-cores (P-cores) and E-cores, as well as GPU execution units for the most intensive edge AI workloads. The built-in Intel Deep Learning Acceleration Technology leverages different cores through Volume Loss Neural Network Instructions (VNNI), and the integrated graphics processor supports AI-accelerated DP4a GPU instructions, which can even be scaled up to a dedicated GPU. in addition, Intel's lowest-power built-in AI accelerator, Intel Gaussian & Neural accelerator 3.0 (Intel GNA ) supports dynamic noise cancellation and speech recognition, and can even run when the processor is in a low-power state for voice wake-up commands.