Performance boost with ten embedded Alder Lake boards

New Products | January 5, 2022

Congatec's modules in COM-HPC Size A and C as well as COM Express Type 6 form factors use the latest Alder Lake hybrid processors to give a significant performance boost for embedded and edge computing designs.Read More

By Nick Flaherty

German board maker congatec has launched ten embedded boards based the 12th Generation Intel Core processors (Alder Lake) using the COM-HPC and COM Express Computer-on-Modules formats.

The modules in COM-HPC Size A and C as well as COM Express Type 6 form factors provide significant performance boost for embedded and edge computing systems.

Alder Lake is a key change with small, power optimised cores and larger cores for higher performance and multithreaded operation. The BGA package versions support up to 14 cores and 20 threads on BGA, with 16 cores/24 threads on the LGA packaged desktop variants for the COM-HPC (High Performance Computing) boards. These use up to 6 or 8 (BGA/LGA) Performance-cores (P-cores) plus up to 8 low power Efficient-cores (E-cores) and DDR5 memory support to accelerate multithreaded applications and execute background tasks more efficiently.

The mobile BGA processors also include up to 96 Execution Units of the integrated Intel Iris Xe GPU to boost embedded graphic performance and can also be used for machine learning and artificial intelligence (AI) algorithms

The graphics of the LGA processor based modules delivers now up to 94 percent faster performance and its image classification inference performance has nearly tripled with up to 181 percent higher throughput. In addition the modules provide higher bandwidth with the COM-HPC specification to connect discrete GPUs for maximum graphics and GPGPU based AI performance.

Compared to the BGA versions, these and all other peripherals benefit from doubled lane speed as they come with PCIe 5.0 interfaces in addition to PCIe 4.0 off the processor. The desktop chipsets also provide up to 8x PCIe 3.0 lanes for additional connectivity and the mobile BGA variants also offer up to 16x PCIe 4.0 lanes off the CPU and up to 8 PCIe 3.0 lanes off the chipset.

Boards using both the BGA and LGA variants are aimed at high end embedded and edge computer systems such as IoT gateways incorporating multiple virtual machines for smart factories and process automation, AI based quality inspection and industrial vision, real-time collaborative robotics, and autonomous logistics vehicles for warehouses and shipping. Typical outdoor applications include autonomous vehicles and mobile machines, video security and gateway applications in transportation and smart cities, as well as 5G cloudlets and edge devices requiring AI supported packet inspection.

“Leveraging Intel’s innovative performance hybrid architecture with impressive P – core performance in combination with power efficient E – cores Intel Thread Director assigns each workload to the proper cores for optimum performance,” said Christian Eder, Director Marketing at congatec and one of the architects of the COM-HPC specification.

“Selected processors are also suitable for hard real-time applications with Intel TCC and TSN. In combination with full support for Real-Time Systems’ hypervisor technology, they are the ideal platform to consolidate a multitude of different workloads on one single edge platform. As this applies to low-power and high-performance scenarios alike, it enables highly sustainable designs with a small ecological footprint,” he said.

The COM-HPC Client and COM Express Type 6 modules support Windows ML, Intel Distribution of OpenVINO toolkit and Chrome Cross ML. The different AI workloads can seamlessly be delegated to the P-cores, E-cores, as well as the GPU execution units to process even the most intensive edge AI workloads.

Built-in Intel Deep Learning uses the different cores via Vector Neural Network Instructions (VNNI), and the integrated graphics supports AI accelerated DP4a GPU instructions that can also be scaled to dedicated GPUs. Intel’s lowest power built-in AI accelerator, the Intel Gaussian & Neural Accelerator 3.0 (Intel GNA 3.0), enables dynamic noise suppression and speech recognition and can run while the processor is in low power states for wake-up voice commands.

Combining these features with support for Real-Time Systems’ hypervisor technology as well as OS support for Real-Time Linux and Wind River VxWorks, makes these modules a truly rounded ecosystem package to facilitate and accelerate the development of edge computing applications.

Next: Alder Lake variants

基于CONGA-TC670 COM Express的第12代Intel Core Mobile处理器类型6紧凑型模块（95毫米x 95毫米）和conga-hpc/calp com-hpc com-hpc客户端尺寸A模块（120 mm x 95 mm）可在以下配置

Processor	Cores/(P + E)	P-cores Freq [GHz]	E-cores Freq [GHz]	Threads	GPU Compute units	CPU Base power [W]
Intel Core i7 12800HE	14 (6+8)	2.4 / 4.6	1.8 / 3.5	20	96	45
Intel Core i5 12600HE	12 (4+8)	2.5 / 4.5	1.8 / 3.3	16	80	45
Intel Core i3 12300HE	8 (4+4)	1.9 / 4.3	1.5 / 3.3	12	48	45

12th Gen Intel Core desktop processor based conga-HPC/cALS COM-HPC Client Size C modules (120 mm x 160 mm) will be available in the following variants:

Processo	Cores/ (P + E)	P-cores Freq. [GHz]	E-cores Freq. [GHz]	Threads	GPU Compute Units	CPU base power [W]
Intel Core i9 12900E	16 (8+8)	2.3 / 5.0	1.7 / 3.8	24	32	65
Intel Core i7 12700E	12 (8+4)	2.1 / 4.8	1.6 / 3.6	20	32	65
Intel Core i5 12500E	6 (6+0)	2.9 / 4.5		12	32	65
Intel Core i3 12100E	4 (4+0)	3.2 / 4.2		8	24	60