EdgeCortix MERA

Compiler and Software Framework

Supported Frameworks & Applications

Enabling software-defined heterogeneous AI inference acceleration

EdgeCortix MERA is the software companion to the EdgeCortix Dynamic Neural Accelerator IP (DNA IP), whether the hardware core sits in an FPGA, a custom SoC, or the EdgeCortix SAKURA-I Energy-efficient Edge AI Co-processor. MERA provides the entire stack for developing edge AI inference applications from modeling to deployment.

Runtime configuration
Dynamic grouping of DNA IP compute units
Fits many workflows
PyTorch, TensorFlow, ONNX, Apache TVM
Multi-model ready
50+ pre-trained models, or custom models
MERA integrates in AI inference workflows

Integrated in familiar AI inference workflows

MERA is a compiler and software tool kit enabling deep neural network graph compilation and AI inference using the Dynamic Neural Accelerator IP (DNA IP). It provides the necessary tools, APIs, code-generator and runtime needed to deploy a pre-trained deep neural network after a simple calibration and quantization step.

MERA offers software developers and data scientists familiar workflows for developing neural network models without dealing with details of chip-level hardware architecture.

  • Native support for PyTorch, TensorFlow, TensorFlow Lite, and ONNX
  • Python and C++ interfaces for workflow integration and customization
  • Post-training calibration and INT8 quantization of user-defined models
  • Supports stand-alone estimation of inference latency and throughput
  • Over 50 pre-defined AI inference applications ready for use
MERA leverages Apache TVM and MLIR functionality, and its front end is open-sourced for easy distribution to project teams.

Everything needed for AI inference development and deployment

  • Runtime configuration of DNA IP
  • Simulation and profiling for stand-alone use
  • Low-level IR (intermediate representation)
  • Host processor code generation
  • High-level graph partitioning
  • APIs for popular machine learning frameworks
  • Model calibration and quantization
  • Pre-trained applications in the Model Zoo
EdgeCortix MERA compiler and software framework provides the entire stack for developing edge AI inference applications

Given the tectonic shift in information processing at the edge, companies are now seeking near cloud level performance where data curation and AI driven decision making can happen together. Due to this shift, the market opportunity for the EdgeCortix solutions set is massive, driven by the practical business need across multiple sectors which require both low power and cost-efficient intelligent solutions. Given the exponential global growth in both data and devices, I am eager to support EdgeCortix in their endeavor to transform the edge AI market with an industry-leading IP portfolio that can deliver performance with orders of magnitude better energy efficiency and a lower total cost of ownership than existing solutions."

Akira Takata
Former CEO of MegaChips Corporation

Learn more about MERA