Skip to main content

SK1660S

NVIDIA GeForce® GTX1660SUPER MXM 3.1 Graphic Module, supporting 4 Channel outputs, and 6GB GDDR6 High-Speed Memory

  • Powered by NVIDIA GeForce GTX 1660 SUPER
  • NVIDIA CUDA technology with 1408 Cores
  • MXM 3.1 Type B
  • Support DirectX 12, OpenGL 4.6, Vulkan API
  • Support for 4 outputs
  • 192-bit width, 6GB , GDDR6 High-Speed Memory
  • Power Consumption: 120W
  • Technical Profile
  • Specifications
  • GPU
I. Introduction

MXM Graphic Solutions

Perfectron’s MXM modules are the compact, thinnest graphics module solution based on the industrial standard Mobile PCI Express Module (MXM) Version 3.1 specification, delivering the latest and cutting-edge GPU benefits for your embedded systems. Its superior graphics performance, GPU computing and video capabilities are the ideal solution for performance demanding systems such as digital signage, medical image, defense, aerospace applications, and AI computing.

NVIDIA CUDA Platform

4444444_0.jpg

Powered by NVIDIA, CUDA® is a parallel computing platform and programming model developed by NVIDIA for general computing on graphical processing units (GPUs). With CUDA, developers are able to dramatically speed up computing applications by harnessing the power of GPUs. In GPU-accelerated applications, the sequential part of the workload runs on the CPU – which is optimized for single-threaded performance – while the compute intensive portion of the application runs on thousands of GPU cores in parallel.

Perfectron’s Computing Accelerating CPU Module 

Perfectron’s high-performance GPU-based modules (OXY5740 - 7th Gen. EBX Platform) can provide exceptional processing power to applications which benefit from advanced parallel processing. Typical applications include image processing, video stabilization, filtering, terrain analytics, 3D visualization of Geospatial data, object recognition and tracking. The latest generation modules include tensor cores and other specialized circuitry which can be used for AI inference, where a model created by training a neural network can be used by a processing module in the field to analyze real world data.

圖片322.png

Mechanical
1660S

 

Block Diagram
2

 

 

II. GPU Roadmap
MXM GPU
III. Related Products

1. OXY5740A (Embedded Board)

OXY5740+SK221+MXM GPU

  • EBX Form factor
  • Intel® 7th Gen. Kabylake-H® Core™ i7-7820EQ processors  
  • Stackable with StackPC & FPE Expansion
  • Capable of connecting to various MXM GPU Module such as GTX-1660SUPER/GTX-1650 /GTX-1050Ti
  • Extended Temperature - 40°C to 85°C

2. SK513 (COM Express Type 6 carrier board) 

SK513+MXM GPU

  • COM Express Type 6 Support Up to Intel XEON E-2276ME, i7-9850H Processor
  • Standard MXM Type 3.1 Supporting NVIDIA GeForce® GPUs
  • PCI/104 Express Expansion Slot for Modular Open Structure
  • 2 x VGA, 1 x LVDS, 4 x COM, 6 x USB, 2 x miniPCIe, 1 x M.2, 2 x SATA Slots
  • Extreme Temperature Support -40~+85 degree
  • 9V-36V DC Input

3. SR800-X1 GPGPU Radar Subsystem  

SR800-X1
IP65 VMware GPGPU rugged workstation with NVIDIA NVIDIA 1050Ti CUDA 768/GTX 1650 CUDA896 /GTX 1660S CUDA1408 ) , Intel® Xeon® D-1577 Processor (16 Cores), 128GB DDR4 ECC RDIMM, MIL-DTL-38999 , MIL-STD 461/1275 Compliant
  • 16 Cores INTEL ® Xeon® D-1577  VMware Support
  • 128GB DDR4 ECC RDIM
  • Quadro P3000 MXM Support
  • Dual Removable 2.5” SSD Tray
  • MIL-STD 461/1275 18V~36V DC Input
  • M12/ DTL38999 Connector
IP65 VMware GPGPU rugged workstation
Operating Temperature
0 to 50
Graphic Module Information
 
GPU NVIDIA GTX 1660SUPER
Process Technology 12 nm
GPU Clock (max.) 1530 MHz
CUDA Cores 1408 CUDA
MXM type MXM 3.1, Type B
Bus Interface PCI Express® 3.0 (X16)
DirectX® Capability DirectX® 12
OpenGL™ OpenGL 4.6
OpenCL™ OpenCL 1.2
Memory Configuration, type 192-bit, 6 GB, GDDR6
G3D Mark Ref. Scores 12705
TDP 120W

1