New MathWorks Hardware Support Package Automates Code Generation from MATLAB and Simulink Models to Qualcomm Hexagon Neural Processing Unit Architecture

MathWorks, the leading developer of mathematical computing software, today announced the availability of a hardware support package for the Qualcomm^® Hexagon™ Neural Processing Unit (NPU), the technology embedded within the Snapdragon^® family of processors. The MathWorks hardware support package automates code generation from MATLAB^® and Simulink^® models optimized explicitly for Qualcomm Technologies’ Hexagon NPU architecture to improve data accuracy and standards compliance while improving developer workflow productivity. Using MathWorks software and Model-Based Design empowers engineers to deploy production-quality C code for embedded applications without in-depth knowledge of the NPU. 

"Qualcomm Technologies’ mission is to empower engineers to create innovative products and applications that leverage the full potential of Snapdragon processors," said Vinesh Sukumar, Sr. Director of Product Management, Qualcomm Technologies, Inc. “As the demand for high-performance NPU applications continues to grow, the collaboration between MathWorks and Qualcomm Technologies enables developers to efficiently develop embedded applications optimized for the Hexagon NPU.”

The MathWorks hardware support package for the Hexagon NPU addresses several critical pain points in the development process, including compilation errors, the reliance on hand-written frameworks, manual numeric verification, and the challenge of tracing code back to the model. Engineers without in-depth knowledge of the Hexagon NPU and its toolchain can use the package to run Processor-In-the-Loop (PIL) simulations directly in Simulink leveraging the Hexagon simulator, ensuring that the optimized target code performs correctly, efficiently, and reliably before deploying on the actual hardware. 

"By harnessing the power of Model-Based Design and the advanced capabilities of the Hexagon NPU, we are making it easier for engineers to create high-quality NPU applications," said Jim Tung, MathWorks Fellow. “Our collaboration with Qualcomm Technologies is focused on removing manual processes in favor of highly automated workflows to help engineering organizations reduce time to market and ensure optimal performance.” 

MATLAB and Simulink’s automatic code generation capabilities are trusted worldwide by customers in automotive, aerospace, medical devices, and other industries. MathWorks software already optimizes code generation for various processors, including the ARM Cortex-M, ARM Cortex-A, ARM Cortex-R, Infineon AURIX TC3x and TC4x, and GPUs from ARM and NVIDIA.