AccelerEyes has released version 1.0 of LibJacket, enabling programmers to achieve better GPU performance with less programming. Designed for use with any Cuda application, in the same way that the native Cublas and Cufft libraries are used, the library is available for C, C++, Fortran, and Python. It can also be used to avoid writing kernel code, for productivity and performance gains. LibJacket is suitable for all Cuda-capable GPUs, from laptops and workstations to high-end supercomputers, and high-level, matrix style code achieves low-level, down-to-metal speeds so that developers can quickly build high-performance applications. This high-level interface makes it easy to experiment and change various parts of algorithms without having to recode and tune from scratch. This new v1.0 release also incorporates the popular GFOR loop for running FOR loop iterations simultaneously on the GPU.
Robert Roe looks at research from the University of Alaska that is using HPC to change the way we look at the movement of ice sheets
Robert Roe talks to cooling experts to find out what innovation lies ahead for HPC users