General
Main memory	128 GB
Memory type	LPDDR5X
GPU	NVIDIA GB10 Blackwell iGPU
CPU architecture	10x ARM Cortex-X925 + 10x ARM Cortex-A725
M.2 slot	1x M-Key
M.2 interface	PCIe 5.0 x4
M.2 form factor	2242/2230
Operating system	Linux

The PNY DGX Spark is a mini-computer developed specifically for applications in the field of artificial intelligence and machine learning. It is based on the NVIDIA GB10 Blackwell GPU and features an integrated graphics unit (iGPU) optimized for high-performance computing. The system is equipped with 128 GB LPDDR5X RAM, enabling fast data processing and parallel computing. The CPU consists of a 20-core processor made up of ten ARM Cortex-X925 and ten ARM Cortex-A725 cores, achieving high computational performance while maintaining energy efficiency. The mini-PC is designed for use in data centers, research environments, or industrial applications where high computing power in a compact form factor is required.

Applications

The PNY DGX Spark is particularly suitable for tasks in the field of artificial intelligence, such as training neural networks, processing large data volumes, or running simulations. The integrated NVIDIA GB10 Blackwell GPU enables efficient parallel processing, which is essential for many AI workloads. Due to its compact design and high power density, the device is also suitable for edge computing applications, where computing power is needed directly at the point of data generation. The use of Linux as the operating system supports integration into existing IT infrastructures and enables the use of open-source tools.

Functions and Specifications in Detail

The system features an M.2 slot with M-Key connector, supporting PCIe 5.0 x4. This slot is designed for SSDs in 2242 or 2230 format, enabling flexible and fast storage expansion. The integrated GPU based on the Blackwell architecture provides high computational performance for graphics-intensive and compute-intensive applications. The use of LPDDR5X memory ensures high bandwidth and low latency, which is particularly beneficial for AI workloads. The CPU combination of Cortex-X925 and Cortex-A725 enables dynamic load distribution, with the powerful X925 cores handling demanding tasks and the energy-efficient A725 cores managing background processes.

Technical Explanation

The NVIDIA GB10 Blackwell GPU is an integrated graphics unit specifically developed for artificial intelligence processing. It enables the parallel execution of millions of calculations, which is crucial for training AI models. The LPDDR5X memory architecture offers high data rates and low power consumption, increasing the system's efficiency. The 20-core CPU (10 Cortex-X925 and 10 Cortex-A725) allows flexible usage: the X925 cores handle compute-intensive tasks, while the A725 cores manage less demanding processes, optimizing overall performance. The M.2 slot with PCIe 5.0 x4 provides high data transfer rates for SSDs, ensuring fast read and write operations.

Compatibility and Accessories

The PNY DGX Spark is designed for use with Linux systems, facilitating integration into existing IT infrastructures. The use of PCIe 5.0 x4 in the M.2 slot allows compatibility with SSDs supporting this interface.

Mounting and Installation

The PNY DGX Spark is mounted in a compact housing designed for use in data centers or industrial environments. Installation is performed via the pre-configured software environment based on Linux. System component configuration is carried out through the operating system environment, with no specific hardware installation steps required.

Operational Advantages

The PNY DGX Spark offers high computing performance in a compact form factor, making it particularly suitable for artificial intelligence applications. The combination of a powerful GPU, fast LPDDR5X memory, and a 20-core processor enables efficient processing of large data volumes. The use of Linux as the operating system supports integration into existing IT infrastructures and enables the use of open-source tools. The compact design reduces space requirements and heat generation, increasing operational reliability. The M.2 interface with PCIe 5.0 x4 enables fast storage expansion, enhancing the system's scalability.