CUDA SDK4.2.9

Especially intended for C and C++ developers and programmers who need to build GPU-accelerated programs in a secure and intuitive environment, CUDA SDK proves to be an effective and steady solution that comprises various tools and features for debugging and optimizing the performance of their apps.

Guided performance analysis

Thanks to its CUDA-accelerated libraries, compiler directives and extensions that help Fortran and C/C++ programmers to render their projects and programs, CUDA SDK provides them also with a step-by-step wizard for a better understanding.On this manner, you are able to quickly identify the bottlenecks and boost the performance of your projects within seconds.

Extensive documentation and drop-in libraries

Since it includes a compiler for NVIDIA GPUs and different utilities that help you to debug your applications so you can improve their performance and responsiveness, CUDA SDK also provides you with programming guides, user manuals and API references.In addition to libraries, here referring to cuSPARE, NPP, cuFFT and cuBLAS, to name a few, CUDA SDK comes with support for computational interfaces such as C++ AMP and Microsoft’s DirectComputer, as well. What’s more, Python, Perl, Java, Ruby, Lua, MATLAB and IDL are also supported.This way, you are able to develop your own parallel applications and libraries by moving specific portions of your code to NVIDIA CPU.

A unified development environment optimized for MPI applications

Since CUDA SDK comes with Hyper-Q support for MPI (Message Passing Interface) processes, it simplifies the programming job consistently. What's more, it comes in handy especially for users who need to design and enhance the performance of their applications.