Skip to main content

P R O D U C T S

NMP-350

Lowest Power and Cost End Point Accelerator

Example

nmp300 example

Markets

  • Automotive
  • AIoT/Sensors
  • Industry 4.0
  • Smart Appliances
  • Wearables

Applications

  • Driver Authentication, Digital Mirrors, and Personalization
  • Predictive Maintenance
  • Machine Automation
  • Health Monitoring

Specifications

  • Up to 1 TOPS
  • Up to 1 MB Local Memory
  • RISC-V/Arm Cortex-M 32-bit CPU
  • 3 x AXI4, 128b (Host, CPU & Data)
frameworks logo

NMP-550

Performance Efficiency Accelerator

Example

nmp500 example

Markets

  • Automotive
  • Mobile and AR/VR
  • Drones and Robotics
  • Security and Surveillance
  • Medical Devices

Applications

  • Driver Monitoring and Fleet Management
  • Image / Video Analytics and Super Resolution
  • Intruder, Safety, and Compliance

Specifications

  • Up to 6 TOPS
  • Up to 6 MB Local Memory
  • RISC-V/Arm Cortex-M or A 32-bit CPU
  • 3 x AXI4, 128b (Host, CPU & Data)
frameworks logo

NMP-750

Performance Accelerator for Edge Computing

Example

nmp700 example

Markets

  • Automotive, AMR, and UAVs
  • AR/VR
  • Smart Building/Factory/City
  • Security and Surveillance
  • Telecommunications

Applications

  • Mobility and Autonomous Control
  • Process, Building, and Factory
  • Automation
  • Multi-Camera Stream Processing
  • Spectral Efficiency and Energy Mgmt

Specifications

  • Up to 16 TOPS
  • Up to 16 MB Local Memory
  • RISC-V/Arm Cortex-R or A 32-bit CPU
  • 3 x AXI4, 128b (Host, CPU & Data)
frameworks logo

NeuroMosAIc Studio

Software Platform

NeuroMosAIc Studio

Converter

  • Quantization (1bit / FXP8 / FXP16)
  • Network Compression
  • Network Optimization
  • Network for NMP Generation
  • Precision Analysis
  • Precision Adjustment

Mapper

  • H/W Level Optimization
  • C Code Generation

Runtime Libraries

  • Optimized function for A.I.
  • Support edge training

Tools

  • NMP Compiler
  • NMP Simulator

NMP-Aware Training

  • Optimize on training stage
  • Optimize quantization