Generating CUDA Code from MATLAB: Accelerating Embedded Vision and Deep Learning Algorithms on GPUs

GPU computing lets you accelerate your compute-intensive algorithms in areas such as deep learning, embedded vision, and radar. With GPU Coder™, you can accelerate your existing MATLAB^® algorithms by automatically generating optimized CUDA code for NVIDIA^® GPUs like NVIDIA DRIVE, Jetson, and Tesla^®.

Read this white paper to find out how you can generate optimized CUDA code from your algorithms developed in MATLAB:

• Prepare MATLAB code for GPU code generation
• Generate, test, and deploy the generated CUDA code on NVIDIA GPUs
• Optimize code for performance

The white paper shows the workflow through an example featuring deep learning–based traffic sign detection. It also provides GPU benchmarks for MATLAB performance compared with other deep learning frameworks.