Vendor: Andes Technology Corp. Category: Edge AI Accelerator

Deep learning accelerator

High performance-efficient edge deep learning accelerator for small LLM/VLM/VLA and CNN/RNN/LSTM

Contact Provider Request Datasheet

Overview

AndesAIRE® AnDLA® I370 is a deep learning accelerator (DLA) designed for high-performance, cost-sensitive edge and endpoint AI inference. It supports popular AI framework formats (TensorFlow Lite, PyTorch, ONNX), NN models (CNN/RNN/ViT/SLM) and datatype INT8/INT16.

AndesAIRE® AnDLA® I370 performs a wide range of neural network operations—convolution, fully connected, elementwise, activation, pooling, channel padding, upsampling, and concatenation—and leverages internal DMA and local memory to maximize hardware efficiency. Operator and layer fusion further improve execution efficiency.

Key configurable parameters of AndesAIRE® AnDLA® I370 include MAC count, local memory size, Programmable Tensor Operator Set Architecture, and DSU (divide/square), delivering flexible computing power up to 4 TOPS at 1 GHz for diverse applications.

Key features

  • Configurable MACs: 32, 64, 128, 256, 512, 1024, 2048 (INT8)
  • Maximum performance up to 4 TOPS at 1GHz
  • Data type: i8i8, i8i16, i16i16 (weight INT16, feature map INT16)
  • Configurable local memory: 16KB to 4MB
  • Multi-dimensional Direct Memory Access (DMA)
  • Bus interfaces: AHB-64b, AXI-64/128/256/512b
  • Neural network (NN) models:
    • Image and Video: MobileNet-v1/v2/v3, ResNet-8/50, Tiny
      YOLO v1/v2/v7, YOLO v3/v5/v8s, Inception-v2, EfficientNet-
      Lite, MobileNet-v1-SSD, BlazeFace, MCUNet-VWW2,
      GhostFaceNet, SqueezeNet v1.1, ShuffleNet-v2, PUNET, …
    • Speech/Voice and Audio: RNNoise, DS-CNN, Tiny
      Wav2letter, DTLN, BC-ResNet-8, U-Net, DeepFilterNet, …
    • ViT/SLM: MobileViT, Swin Transformer, BertTiny, YOLOv12,
      TinyLlama 2 110M
  • Operators: MatMul, MHA, FeedForward, Conv2D, Depthwise
    Conv, Pointwise Conv, Transpose Conv, Dilated Conv,
    Elementwise (Add, Add_const, Sub, Sub_const, Mult, Mult_const,
    input vector broadcasting to tensor), Fully Connected, Activation
    (Hardsigmoid, Hardswish, leaky ReLU, ReLU, ReLU6,
    ReLU_n1_to_1, MISH, Softplus, Logistic, Swish, Tanh, SiLU,
    GeLU, PReLU), Pooling (Average, Global Avg, Max), Upsampling,
    Concatenation, Split, Strided slice, D2S, S2D, SpaceToBatchND,
    BatchToSpaceND, Reshape, Transpose, Batch norm, Channel
    padding, GRU, LSTM, RNN, Tiny channel
  • Programmable Tensor Operator Set Architecture (configurable)
  • Sqrt, div (configurable)
  • Operator and layer fusion

Block Diagram

Click to open fullscreen

Applications

  • AIoT and TinyML on edges and endpoints
  • Smart cameras
  • Smart sensors
  • Sensor hubs
  • Wearables
  • TWS earbuds
  • Hearing aids
  • Smart home appliances
  • Smart cockpit
  • Robotics

Specifications

Identity

Part Number
AndesAIRE® AnDLA® I370
Vendor
Andes Technology Corp.
Type
Silicon IP
Application
Edge AI

Files

Product Sheet Download

Note: some files may require an NDA depending on provider policy.

Provider

Contact Andes Technology Corp.

Learn more about Andes' Deep learning accelerator IP core

Andes Technology Unveils AndesAIRE™ AnDLA™ I370: A Next-Generation Deep Learning Accelerator for Edge and Endpoint AI

The AndesAIRE™ AnDLA™ I370 supports industry-standard deep learning frameworks including TensorFlow Lite, PyTorch, and ONNX, making it easier for developers to deploy AI workloads across platforms. Capable of executing complex neural network operations such as convolution, fully connected, elementwise, pooling, activation, channel padding, upsampling, and more, the I370 integrates internal DMA and local memory and boosts execution efficiency through operator and layer fusion.

Frequently asked questions about Edge AI Accelerator IP cores

What is Deep learning accelerator?

Deep learning accelerator is a Edge AI Accelerator IP core from Andes Technology Corp. listed on Semi IP Hub.

How should engineers evaluate this Edge AI Accelerator?

Engineers should review the overview, key features, supported foundries and nodes, maturity, deliverables, and provider information before shortlisting this Edge AI Accelerator IP.

Can this semiconductor IP be compared with similar products?

Yes. Buyers can compare this product with similar semiconductor IP cores or IP families based on category, provider, process options, and structured technical specifications.

×
Semiconductor IP