Skip to end of metadata
Go to start of metadata

This part of the guide gives a short introduction to compiling, running,  profiling and debugging an application that can utilize Intel Xeon Phi coprocessors installed on Taito.

Xeon Phi coprocessor are based on x86 technology but offer much more cores than standard CPUs. This makes them comparable to GPUs when looking at performance, but they can be programmed with a more traditional approach and require a less steep learning curve. These coprocessors are physically installed normal compute nodes but, differently from GPU accelerators, they run their own instance of GNU/Linux and can be used independently from the hosting node.

  • No labels