Speeds up deep learning inference by XNUMXx compared to TensorFlow on NVIDIA GPUs
MathWorks has announced that MATLAB now offers integration with NVIDIA TensorRT via GPU Coder. This helps engineers and scientists develop new AI and deep learning models in MATLAB with the performance and efficiency needed to meet the growing demands of data centers, embedded applications, and automotive applications. MATLAB provides a comprehensive workflow for rapidly training, validating, and deploying deep learning models. Engineers can use GPU resources without additional programming, so they can focus on their applications instead of tuning performance. New NVIDIA TensorRT integration with GPU Coder enables deep learning models developed in MATLAB to run on NVIDIA GPUs with high performance and low latency.
Internal benchmark tests show that MATLAB-generated CUDA code combined with TensorRT can deploy Alexnet with 5x better performance than TensorFlow, and can deploy VGG-16 with 1,25x better performance than TensorFlow for deep learning inference . “Rapidly evolving imaging, voice, sensor and IoT technologies are pushing teams to explore AI solutions with higher performance and efficiency. Furthermore, deep learning models are becoming more and more complex. All of this puts immense pressure on engineers,” says David Rich, director at MathWorks. “Now, teams training deep learning models with MATLAB and NVIDIA GPUs can deploy real-time inference in any environment, from the cloud to the data center to embedded edge devices.” For more information on MATLAB for deep learning, visit: mathworks.com/solutions/deeplearning.html.
