we are looking for either one post-doctoral researcher or PhD student to work with us on the CELERITY project.
The main goal is the design of a programming environment to develop energy- and performance- efficient, predictably scalable, and easy-to-program parallel applications targeting large-scale heterogeneous HPC clusters. The CELERITY environment will integrate a compilation part (based on LLVM), a distributed runtime system, and a set of advanced modeling approaches.
The programming model is based on SYCL, the compilation infrastructure will use LLVM on SPIR-V kernels.
For full details please see: http://www.aes.tu-berlin.de/menue/jobs/