Neural Processing Unit (NPU) IP

TM

ENLIGHT

OIC

OIC

ENLIGHT

TM

Features a highly optimized network model compiler that reduces DRAM traffic from intermediate activation data by grouped layer partitioning and scheduling. ENLIGHT is easy to customize to different core sizes and performance for customers' targeted market applications and achieves significant efficiencies in size, power, performance, and DRAM bandwidth.

ENLIGHT Toolkit Overview Diagram.png

Performs various operations of deep neural networks such as convolution, pooling, and non-linear activation functions for the edge computing environments. It has a solid competitive edge in delivering unparallel compute density with energy efficiency (power, performance, and area)

Hardware Key Advantages

Mixed Precision (4-/8-bit) Computation

  • Higher efficiency in PPAs (power, performance, and area), DRAM bandwidth

Deep Neural Networks (DNN)-optimized Vector Engine

  • Better adaptation to future DNN changes

Scale-out with Multi-core

  • Even higher performance by parallel processing of DNN layers

Modern DNN Algorithm Support

  • Depth-wise convolution, feature pyramid network (FPN), swish/mish activation, etc. 

Software Key Advantages

High-level Inter-layer Optimization

  • Grouped layer partitioning and scheduling for reducing DRAM traffic from intermediate data​

 

DNN-layers Parallelization

  • Efficiently utilize multi-core resources for higher performance

  • Optimize data movements among cores

Aggressive Quantization

  • Maximize use of 4-bit computation capability

 
 

ENLIGHT Toolkit Overview

NN Converter

  • Converts a network file into internal network format (.enlight)

  • Supports ONNX (PyTorch), TF-Lite, and CFG (Darknet)

 

NN Quantizer

  • Generates quantized network: float to 4-/8-bit integer​

  • Supports per-layer quantization of activation and per-channel quantization of weight

 

​NN Simulator

  • Evaluates full precision network and quantized network

  • Estimates accuracy loss due to quantization

NN Compiler

  • Generates NPU handling code for target architecture and network

 

ENLIGHT Toolkit Applications

  • Person, Vehicle, Bike, Traffic Sign Detection

  • Parking Lot Vehicle Location Detection & Recognition
  • License Plate Detection & Recognition

  • Detection, Tracking, and Action Recognition for Surveillance

 

Deliverables

ENLIGHT Toolkit is available to all eligible companies with the following items:

  • RTL design for synthesis

  • User guide

  • Integration guide