NVIDIA GPUs

High-performance computing with MATLAB and NVIDIA GPUs for PCs, data centers, and embedded systems

Highlights

Accelerate MATLAB^® using NVIDIA GPUs with over 500 built-in functions though Parallel Computing Toolbox™
Generate CUDA code from MATLAB for workstations, data centers, and embedded systems with GPU Coder™
Generate NVIDIA TensorRT™ CUDA code for high-performance inference with GPU Coder
Access multiple GPUs on desktops, clusters, and clouds using MATLAB and MATLAB Parallel Server™
Learn more about NVIDIA GPU capabilities in MATLAB

Description

GPUs are specialized processors designed for mathematically intense and massively parallel computations. NVIDIA GPUs today power many of the top 500 fastest supercomputers, including the fastest petascale systems in the world. NVIDIA and MathWorks have collaborated to deliver the power of GPU computing in MATLAB. Learn more about MATLAB GPU Computing with NVIDIA GPUs from MathWorks.

The GPU computing capabilities in MATLAB enables users enjoy the performance of massively parallel computation without having to be an expert in CUDA, C, or Fortran. Users can also scale up to large GPU clusters using MATLAB Parallel Server.

GPU Coder enables users to generate optimized CUDA code from MATLAB for deep learning, embedded vision, and autonomous systems. The generated code automatically calls optimized NVIDIA CUDA libraries, including TensorRT, cuDNN, and cuBLAS, to run on NVIDIA GPUs with low latency and high-throughput.