SIE brings the best compatibility within scientific computing clusters by introducing Intel PHI GPUs with Xeon processors.
SIE provides high-level solutions for scientific calculation, based on Intel Phi. This card represents a revolution in the world of mathematical coprocessors, due to its compatibility with x86. This makes it a great competitor to the GPU cards offered by NVDIA and ATI, since even accepting that they offer great features and a large number of cores, code migration is not trivial.
In fact, in GPU-based technologies, it means rewriting the programs again, even though there are CUDA-like libraries to help you migrate to the NVIDIA platform.
The advantage of the Intel Phi is that it is only a migration from x86 "some cores" to x86 "many cores" and this can be done in a trivial way using Intel's Fortran and C++ compilers. For the first time, we have SIE Ladón systems with 260 x86 cores, on a single machine, which allows us to reuse thousands of lines of scientific code, written in
these languages since 1990.
If you need more specific information on the ease of migration, we send you a European document of good practices on Intel Phi:
http://www.prace-project.eu/Best-Practice-Guide-Intel-Xeon-Phi-HTML?lang=en
