Just days before the start of the firm's San Francisco located developer forum symposium, Intel has detailed the release of updates to two flagship development suites: Intel Cluster Studio XE 2013 and Intel Parallel Studio XE 2013.
These "integrated" suites are designed to help C++ and Fortran developers create applications for the latest Intel processors and coprocessors, including Haswell-EP, Broadwell and the range of Intel Xeon Phi coprocessors.
Updates include better compiler and library performance, additional Windows development support, and standards integration for OpenMP 4.0, MPI 2.x, Fortran 2003 & 2008, and C++ 11. The Intel Cluster Studio XE 2013 SP1 supports shared, distributed, and hybrid memory applications.
This release can "scale performance" — there is a low-latency MPI library and native profiling and tuning of MPI applications now includes node level analysis; it can "scale forward" — there is scope to scale up to 120K MPI Processes and it features high-performance advanced parallel programming models; plus it can also "scale efficiently" — there is rapid performance profiling for identification of application hotspots, plus threading and memory debugging finds errors and static misses at every node.
According to the Intel Developer Zone spec details pages, "Intel Cluster Studio XE 2013 meets the challenges facing HPC developers by providing, for the first time, a comprehensive suite of tools that enables developers to boost HPC application performance and reliability. It combines Intel's proven cluster tools with Intel's advanced threading/memory correctness analysis and performance profiling tools to enable scaling application development for today's and tomorrow's HPC cluster systems."
Intel further states that the impact of budget and schedule pressure makes it crucial to have the right tools and programming models to rapidly develop and deploy reliable HPC applications. Intel Cluster Studio XE has promised to provide "threading and correctness tools" for hybrid applications development and parallel programming models that are simple to adopt.