Jülich: Master Thesis - Application of LLM in performance analysis in HPC

Cube, which is used as performance report explorer for Scalasca and Score-P, is a generic tool for displaying a multi-dimensional performance space consisting of following dimensions: performance metric, call path, system resources. Each dimension can be represented as a tree, where the non-leaf nodes can be collapsed or expanded to achieve the desired level of granularity. CUBEGUI is a graphical component, which allows the manual and visual exploring of the performance measurement result, the cube profile. Its plugin-based architecture promotes extensibility, allowing for the seamless integration of additional analysis capabilities through dedicated plugins. This kind of sophisticated data analysis is now being revolutionized by artificial intelligence. AI is increasingly demonstrating its power across a wide range of human activities, often yielding surprising and impressive outcomes. By analyzing vast amounts of text+data, AI models learn to understand the world in a profound way, extending beyond the mere memorization of facts. They begin to grasp the deeper meaning and connections between things, much like humans do, developing a nuanced comprehensive of how the world works. Join us now for this highly interesting topic.

Your Job:
This master`s thesis aims to investigate the potential of leveraging Large Language Model (LLM) capabilities for performance analysis. This will involve developing a prototype plugin for CubeGUI that implements this new approach and demonstrating its practical applicability.

Your tasks will include:

• Development of the method to describe performance data in form as a prompt for the LLW
• Demonstration of the possibility to detect typical performance issues such as load imbalance using LLM
• Developing a a CubeGUI plugin to provide the interface and the automation of the developed approach
• Investigation of possible limitations of the approach

Your Profile:

• Completed Bachelor`s degree and currently enrolled in a master`s program in computer science, mathematics or related field
• Advanced skills in parallel programming using C++
• Basic understanding of the functionality of LLM
• Interest in the High Performance Computing field
• Ability to communicate with colleagues and discuss various problems in the field
• Well-structured and systematic research approach

Please feel free to apply for the position even if you do not have all the required skills and knowledge. We may be able to teach you missing skills during your time with us.

Our Offer:

We work on the very latest issues that impact our society and are offering you the chance to actively help in shaping the change! We support you in your work with:

• An interesting and socially relevant topic for your thesis with future-oriented themes
• Ideal conditions for gaining practical experience alongside your studies
• Excellent technical equipment and the newest technology
• Qualified support through your scientific colleagues
• The chance to independently prepare and work on yours tasks
• Extensive company health management
• Ideal conditions for balancing work and private life, as well as a family-friendly corporate policy
• Flexible work (location) arrangements, e.g. remote work

We welcome applications from people with diverse backgrounds, e.g. in terms of age, gender, disability, sexual orientation / identity, and social, ethnic and religious origin. A diverse and inclusive working environment with equal opportunities in which everyone can realize their potential is important to us.