GPU access for LiU-course: Project Course in Applied Physics, TFYB13
Abstract
TFYB13 is an undergraduate project course given for, primarily, civil engineering program students at the very end of their education. It offers a computational project run for 1-2 small student groups. The course runs all of the autumn term (which includes some time into January). The students doing the computational project in this course develop a molecular dynamics code and run it on a fairly large data set. Last year the students performed these runs on Sigma using allocations LiU-compute-2025-38 and LiU-gpu-2025-13.
This year we have 1 student group with 7 students.
For the current iteration of the project, it is useful with a mix of a small amount of CPU and GPU resources: GPU resources to be able to run machine-learned interatomic potentials. Since there are less GPU resources than CPU resources on Sigma, we've made this proposal a GPU round proposal. However, we'd be happy if 1x1000 CPUh/month also could be added to this project (alongside the 1x1000 GPUh/month we have filled into "resources" of this proposal). Alternatively, if we have to get the CPU resources through a separate CPU round, just tell us, and we will submit that.
The course is already running, so if granted, we'd be happy for the allocation starting as soon as possible.
