Quadric favicon

Quadric
Flexibility of a Processor + Efficiency of an NPU Accelerator

Quadric
Contact for Pricing

Home: https://quadric.io

Social:
Visit Quadric

What is Quadric?

Quadric offers an advanced processor architecture specifically engineered for on-device artificial intelligence tasks. Its core product, the Chimera GPNPU (General-Purpose Neural Processing Unit), combines high machine learning inference performance with the capability to execute complex C++ code within a single processor. This unified approach eliminates the need for developers to partition code across multiple processor types, simplifying System-on-Chip (SoC) hardware design and accelerating the porting of new ML models.

The Chimera GPNPU architecture is designed for scalability, offering performance ranging from 1 to 864 TOPs (Tera Operations Per Second). It supports a wide array of machine learning networks, encompassing classical backbones, modern vision transformers, and large language models (LLMs). Quadric also provides safety-enhanced versions tailored for automotive applications and supports designers with the Chimera SDK and Quadric DevStudio for development and visualization.

Features

  • Unified Architecture: Handles matrix, vector, and scalar (control) code in one execution pipeline, eliminating code partitioning.
  • Broad Model Support: Runs diverse ML models including classical backbones, vision transformers, and Large Language Models (LLMs).
  • High Scalability: Offers performance options ranging from 1 to 864 TOPs.
  • C++ Code Execution: Runs complex C++ code alongside ML inference.
  • Automotive Ready: Provides safety-enhanced cores (ASIL-ready) for automotive applications.
  • Developer Tools: Includes Chimera SDK and Quadric DevStudio for design and development.

Use Cases

  • Developing and deploying AI models on edge devices.
  • Designing efficient System-on-Chip (SoC) solutions for AI applications.
  • Implementing advanced driver-assistance systems (ADAS) in automotive applications.
  • Running complex AI workloads like Vision Language Models (VLMs) on-device.
  • Accelerating the porting and deployment of new machine learning models.
  • Optimizing performance for applications requiring both ML inference and traditional C++ processing.

Related Queries

Helpful for people in the following professions

Related Tools:

View all Alternatives

Blogs:

Didn't find tool you were looking for?

Be as detailed as possible for better results