Development and Tuning of Algorithms and Applications with OpenPower and NVIDIA GPUs
In this internship you will have the opportunity to participate in a leading edge technology project in the IBM Research & Development Lab in Boeblingen.
Our department is developing supercomputer and High Performance Computing (HPC) products. One of the topics we are focusing on is the recently announced OpenPower platform including NVIDIA GPU based accelerators. Our goal is to implement and optimize scientific as well as commercial algorithms on this highly parallel platform. At the time of your application we will identify an interesting and challenging hot topic in this environment for you to work on. Fluent English or German skills spoken and written required.
Aspects of this internship include:
– Work in a large, globally distributed team.
– Do parallel software development on the latest Power8 CPUs and NVIDIA GPUs with Cuda
– Identify and implement improvements for existing and new software
– Identify hardware bottlenecks and suggest improvements
– Study in a bachelor, master or diploma degree study course in computer science, mathematics, electrical engineering, physics or comparable
– Have very good programming skills in C and/or C++
– Be familiar with Linux shells
– Have background in and enthusiasm for system architectures
– Knowledge in scientific applications
– Knowledge in parallel computing concepts such as Cuda, Message Passing Interface, OpenMP, OpenCL, etc
– Programming in FORTRAN
– Knowledge in GPU architecture and programming
Weitere Informationen und Bewerbung:
Ausschreibung bis: 31.07.2017
Unternehmen: IBM Deutschland GmbH