NVIDIA has announced the release of CUDA Toolkit 7.5, a parallel computing platform and programming model for enhancing high-performance computing by leveraging the power of the graphics processing unit (GPU). To take advantage of the CUDA features on requires any one or more of the CUDA-enabled GPUs.
CUDA eliminates the need for Assembly when programming on the GPU. The programmer can basically use C, C++ and Fortran code to access the GPU features directly.
With modern GPUs doing more than just Graphics processing, GPUs are being used for anything from medicine to finance to running server farms.
NVIDIA’s first GPU was released in 1999 and years later NVIDIA released CUDA the solution for general computing on GPUs. That was in 2006. Fast forward to 2015. NVIDIA now has a new version of CUDA.
What’s new in CUDA Toolkit 7.5?
CUDA 7.5 introduces a bunch of new features including:
- Bayer CFA-to-RGB color conversion routines
- n-dimensional Euclidean norm and the four-dimensional and n-dimensional Euclidean reciprocal norms
- cuSPARSE GEMVI routines for acceleration of natural language processing
- FP16 storage for up to twice as large datasets in GPU memory and optimized memory bandwidth
- Instruction-level profiling
- New cuFFT functions for the specification of transforms on arrays spanning more than 4G elements … and more
You can download CUDA Toolkit 7.5 here and get to use it in your new or existing GPUs computed projects.
Found this article interesting? Follow Brightwhiz on Facebook, Twitter, and YouTube to read and watch more content we post.