Navigation and service

Advertising division: ER-C-1 - Physics of Nanoscale Systems
Reference number: D034/2018, Physics, informatics/computer science

Master thesis: Interactive supercomputing for terabyte-scale electron microscopy datasets

Fast pixelated direct electron detectors for scanning transmission electron microscopy (STEM) allow capturing the entire transmitted beam as a high-resolution detector frame for each scanned pixel.

Such a data set contains a wealth of information about the sample and allows to perform detailed analysis interactively with computational methods rather than on the microscope: a paradigm shift in electron microscopy.

We are developing high-performance computing strategies to handle and process the immense amounts of raw data that can reach 64 TB for one scan at full resolution, ideally responding to commands within seconds.

At this scale, efficient use of processors and storage is critical.

In this thesis, you can choose to work on a number of strategies to deal with such data sets, depending on your preferences and prior knowledge:

  • Optimized processing strategies to maximize throughput on multicore machines
  • Smart lossless compression algorithms that exploit hidden redundancies in the data
  • Optimized lossy data reduction that keeps a maximum of scientifically relevant information and eliminates noise
  • Parallelization of algorithms to run efficiently on clusters with both CPUs and GPUs

Depending on the problem you choose, your job can involve analysis of the bottlenecks and development of strategies to circumvent them, proof-of-principle implementations, and integration in existing software. Ideally, your work will find widespread application as part of an open source software infrastructure that we are developing for this application.

The Forschungszentrum Jülich is home to some of the most advanced electron microscopes in the world and houses several supercomputers that currently take ranks 22 and 29 on the Top500 list. Experts both in electron microscopy and supercomputing will guide you with your project.

Contact person:
Dr. Dieter Weber
Forschungszentrum Jülich
Ernst Ruska-Centre for Electron microscopy