Compute Unified Device Architecture is a programming model and parallel computing platform developed by the Nvidia Corporation and released in 2007. It runs on Windows, macOS and Linux, and improves computing performance through the use of a CUDA-enabled graphics processing unit (GPU). CUDA can work with applications that use the Direct3D and OpenGL APIs. Programming in Java, Python, MATLAB, C, C++ and Fortran is supported. GPU compilation is maintained by the Nvidia CUDA Compiler (NVCC), a proprietary compiler based on LLVM.
Enables study of the inner structure of an object to an accuracy of up to 1 micron by a non-destructive method. The software controls x-ray equipment which makes 360° snapshots of cross sections in increments of 1°. A 3D voxel model is then built. The reconstruction speed is increased by an expandable computing cluster with CUDA.
2002–2024