Embedded MXM Modules
Based on NVIDIA’s and Intel‘s embedded GPUs, ADLINK’s Embedded MXM GPU modules help improve the performance of a broad range of tasks, including image processing and analysis, computing acceleration, and artificial intelligence (AI). The Embedded MXM GPU modules from ADLINK provide excellent performance, high data bandwidth, power efficiency, and longevity in a small hardware footprint, making them ideal for applications with limited space, weight, and power. The Embedded MXM GPU modules from ADLINK can improve system responsiveness and accuracy while lowering latency, making them perfect for applications such as industrial automation, healthcare, transportation, aerospace and defense, gaming, and telecom. To learn more about the Adlink Embedded MXM Modules, please visit here.
- NVIDIA RTX A500 GPU, 4GB GDDR6
- 2048 NVIDIA CUDA® cores/ 64 NVIDIA Tensor Cores/ 16 NVIDIA RT Cores
- 100 TOPS DENSE INT8 in inference
- 6.54 TFLOPS Peak FP32 performance
- NVIDIA CUDA X, RTX Software Enhancements
- 106 x 72 x 25mm (less than 200cc) and 250 gram in weight
- Auxiliary Power Input via USB PD
- 40Gb/s bandwidth for Data Transfer via Thunderbolt 3.0+ USB Type-C
- NVIDIA Ampere Architecture
- Standard MXM 3.1 Type A (82 x 70 mm)
- PCIe Gen 4 x4 interface
- 2048 CUDA® cores, 16 RT Cores, and 64 Tensor Cores
- 2/4GB GDDR6 memory, 64-bit
- 96GB/s maximal memory bandwidth
- 5-year availability
- NVIDIA Ampere Architecture
- Standard MXM 3.1 Type A (82 x 70 mm)
- PCIe Gen 4 x8 interface
- 2048 CUDA® cores, 16 RT Cores, and 64 Tensor Cores
- 7.4 TFLOPS peak FP32 performance
- 4GB GDDR6 memory, 128-bit
- 192GB/s maximal memory bandwidth
- 5-year availability
- NVIDIA Ampere Architecture
- Standard MXM 3.1 Type A (82 x 70 mm)
- PCIe Gen 4 x8 interface
- 2560 CUDA® cores, 20 RT Cores, and 80 Tensor Cores
- 9.3 TFLOPS peak FP32 performance
- 4GB/8GB GDDR6 memory, 128-bit
- 192GB/s maximal memory bandwidth
- 5-year availability
- NVIDIA Ampere Architecture
- Standard MXM 3.1 Type B (82 x 105 mm)
- PCIe Gen 4 x16 interface
- 5888 CUDA® cores, 40 RT Cores, and 160 Tensor Cores
- 17.8 TFLOPS peak FP32 performance
- 8GB/16GB GDDR6 memory, 256-bit
- 384GB/s maximal memory bandwidth
- 5-year availability
- NVIDIA Turing™ Architecture
- Standard MXM 3.1 Type A (82 x 70 mm)
- 896 CUDA cores
- 2.6 TFLOPS peak FP32 performance
- 4GB GDDR6 memory, 128-bit
- 192GB/s maximal memory bandwidth
- Support up to 4 DP 1.4a displays, 50W TGP
- 5-year availability
- NVIDIA Turing™ Architecture
- MXM 3.1 Type B form factor (82 x 105 mm)
- 1920 CUDA cores, 30 RT cores, and 240 Tensor cores
- 5.3 TFLOPS peak FP performance
- 6GB GDDR6 memory, 192-bit
- 336GB/s maximal memory bandwidth
- Support up to 4 DP 1.4a displays, 80W TGP
- 5-year availability
- NVIDIA Turing™ Architecture
- Standard MXM 3.1 Type B+ form factor (82 x 110mm)
- 3072 CUDA cores, 48 RT cores, and 384 Tensor cores
- 9.4 TFLOPS peak FP32 performance
- 16GB GDDR6 memory, 256-bit
- 448GB/s maximal memory bandwidth
- Support up to 4 DP 1.4b displays, 110W TGP
- 5-year availability
The footprint of the host system is reduced via ADLINK’s compact embedded MXM GPU modules, which are just one-fifth the size of PCIe graphics cards. The modules are made to resist the extreme temperatures, vibrations, and shocks that are common in an industrial setting. The embedded MXM GPU modules from ADLINK also have low power consumption, starting at just 20 watts.
Product Video
Applications
Applications of the embedded MXM graphics modules include:
- Healthcare: Accelerated image reconstruction for mobile X-ray, ultrasound and endoscopic systems
- Transportation: Real-time object detection on railways or airport runways to enhance transport safety
- Retail and Logistics: Navigation and route planning for autonomous drones and mobile robots (AMRs) to assist with last-mile delivery
- Aerospace and Defense: Time-sensitive and mission-critical command, control, communications, computers (C4) intelligence, surveillance and reconnaissance (ISR) applications
- Gaming: Immersive and stunning visuals experience for multi-display gaming machines
Consulting & Support
Do you have any further questions about Embedded MXM Modules? We will be delighted to answer any questions you may have. As Germany’s leading system integrator, we provide individual support to our customers in the field of Embedded MXM Modules and quickly find bespoke solutions.
Together with our technology partner Adlink Technology we find the best embedded solutions in the field of Embedded MXM Modules. Contact us directly now!