High-Performance Systems for Next Generation Computing - overview
Abstract:
Cloud computing has progressed beyond simple application hosting to embrace new workloads such as deep learning, data analytics, container orchestration and serverless applications. Furthermore, on-demand computing is spreading from the data center to the edge of the network to create new organisational models such as edge-computing. This has motivated the exploration of new computing architectures, integration of specialised hardware (e.g. accelerators), and new software frameworks for simplifying management of these workloads (e.g. Tensorflow, Kubernetes). It's an exciting time to perform research in distributed systems.
The High-Performance Systems group in IBM Research-Ireland is at the forefront of research into next generation computing systems. We are exploring topics such as advancing the state-of-the-art of data centers through disaggregation, optimizing the performance of accelerator-based deep learning workloads, and composing and orchestrating intelligent-workflows. See our group site to get a flavour of our past and current projects.
Candidates will get the opportunity to solve problems such as improving I/O performance for cloud-based workloads, efficient workflow scheduling and execution across heterogenous clouds, and designing next-generation rack-scale/disaggregated Cloud data centers.
Required skills:
- Operating systems hands-on experience
- Computer architecture
- Programming skills: C, C++ and Python
- Distributed systems hands-on experience
- Experience with kubernetes is desirable.
- Experience with a deep learning framework (e.g. tensorflow) is desirable