Master thesis: Multi gpu work splitting for a raytracing simulator
Om jobbet
Join our Team
About this opportunity:
This is an opportunity for a Master of Science student to work with state-of-art HPC simulators for next generation connectivity.
Connectivity technology has become not only ubiquitous, but also increasingly complex with the need to reduce lead times from idea to field test. This is enabled by simulators with high fidelity that can be used to evaluate new ideas early and to guide investment decisions
At the foundation of radio access simulation is radio wave propagation simulations in city environments. Our internal state-of-art propagation simulator makes use of GPU accelerated raytracing combined multithreading on CPU to deliver high quality results to our researchers. The jobs it runs are large and thus making this code performant is essential.
What you will do:
The objective of the Master thesis is to investigate low-level GPU multitasking for single and multi-gpu systems in the context of our simulator. The student will investigate the current state-of-art, both in academia and relevant open source projects. Analyzing the simulator, they will propose which state-of-art solution is the most promising and argue the case with a proof-of-concept implementation. Depending on the direction, adaptive task scheduling may also be in scope for this work.
You will be working in a research environment alongside experts in their respective field in a collaborative. The job will give you experience working with software R&D at a large company with a possibility to innovate, a great starting point for a career in this area.
The project is expected to be performed in Kista for a duration of 6 months, starting in 2026 Q1.
The skills you bring:
You should be a Master of Science student in Computer/Data Science, or similar. Courses in GPU programming, parallel programming, and algorithm analysis are highly desirable. Experience of wireless communication systems, and 5G are valuable merits as well as previous experience in working with radio frequency propagation simulators.
Excellent skills in parallel programing and parallel algorithm analysis
Good skills in CUDA and modern c++ programming
Experience working with large code bases
Good communications skills
You are a self-motivated and positive person
Strong analytical and problem-solving skills
Collaborative
About this opportunity:
This is an opportunity for a Master of Science student to work with state-of-art HPC simulators for next generation connectivity.
Connectivity technology has become not only ubiquitous, but also increasingly complex with the need to reduce lead times from idea to field test. This is enabled by simulators with high fidelity that can be used to evaluate new ideas early and to guide investment decisions
At the foundation of radio access simulation is radio wave propagation simulations in city environments. Our internal state-of-art propagation simulator makes use of GPU accelerated raytracing combined multithreading on CPU to deliver high quality results to our researchers. The jobs it runs are large and thus making this code performant is essential.
What you will do:
The objective of the Master thesis is to investigate low-level GPU multitasking for single and multi-gpu systems in the context of our simulator. The student will investigate the current state-of-art, both in academia and relevant open source projects. Analyzing the simulator, they will propose which state-of-art solution is the most promising and argue the case with a proof-of-concept implementation. Depending on the direction, adaptive task scheduling may also be in scope for this work.
You will be working in a research environment alongside experts in their respective field in a collaborative. The job will give you experience working with software R&D at a large company with a possibility to innovate, a great starting point for a career in this area.
The project is expected to be performed in Kista for a duration of 6 months, starting in 2026 Q1.
The skills you bring:
You should be a Master of Science student in Computer/Data Science, or similar. Courses in GPU programming, parallel programming, and algorithm analysis are highly desirable. Experience of wireless communication systems, and 5G are valuable merits as well as previous experience in working with radio frequency propagation simulators.
Excellent skills in parallel programing and parallel algorithm analysis
Good skills in CUDA and modern c++ programming
Experience working with large code bases
Good communications skills
You are a self-motivated and positive person
Strong analytical and problem-solving skills
Collaborative
Ericsson AB
FöretagEricsson AB
Visa alla jobb för Ericsson AB