Deploy AI with Simulink
summary
Shows a Simulink-to-hardware workflow that generates optimized C/C++ and deploys an ONNX model to Qualcomm Hexagon NPU and ADSP using the MathWorks hardware support package. Builds a smart‑speaker demo (pre/post audio DSP on ADSP, CNN on ENPU), converts ONNX→.EAI with LPAI, runs process‑in‑the‑loop simulations and task profiling, and reports real hardware-like results: plain C misses real‑time (≈250% load) while optimized + ENPU + QHL/HVX hits 17% CPU utilization on Snapdragon-class boards. Setup is three steps (Hexagon SDK, LPAI add‑on, ADB).
AI-written summary · about
chapters (9)
- 0:00AI deployment with Simulink
- 3:07Qualcomm’s AI edge vision
- 9:07Hexagon NPU workload offload
- 11:52Hexagon DSP hardware support
- 15:47Setup and getting started
- 18:56Smart speaker AI demo
- 23:09Hardware simulation and profiling
- 26:22Optimized C code generation
- 29:37Summary and resources
related talks
16mRealize your vision with a guide to building Windows on Snapdragon
14mAccelerating On-Device AI Development
17mOverview of Qualcomm Snapdragon Neural Processing Engine
29mQualcomm® AI Stack for developers and extension to on-device AI
15mBuilding Gen-AI experiences using LLM and LVM
26mHow to use Qualcomm AI Hub
41mAccelerate model deployment with
39mQualcomm Technologies and the power of Windows AI Q&A