NVIDIA has announced that one of the most important Open Source compilers, LLVM, and is capable of supporting NVIDIA GPUs , and so developers can take advantage of this compiler for CUDA accelerate applications using NVIDIA graphics processors.
LLVM supports a large number of programming languages including C / C + +, Objective-C, Fortran, Ada, Haskell, Java, Python, Ruby, ActionScript, GLSL or Rust. In addition is the basis of the infrastructure itself uses NVIDIA CUDA architecture for C / C + + , and has also been adopted by companies like Apple, AMD or Adobe.
This could boostnotably the use of CUDA in different environments, but we must also bear in mind that there are other alternatives in the segment GPGPU: Stream, AMD alternative (formerly Close to Metal) is its traditional rival, while above the two is, at least in nuustro scene, I believe, OpenCL, which takes time trying to attract interest from the developer community.
See if this ad
a substantial change in the development of GPGPU capabilities, a field that was fashionable for some time but which not hear so much as we would like .
No related posts.