Having completed the elementary work on the dummy data in the previous few weeks, I was now ready for the working on the actual data ! Each file here consists of 100 timestamps, in each dimension, radial, azimuthal, axial have 96, 128, 512 modes. The size of a single file was about 30GB. The structure was similar to the dummy data we had used in the earlier weeks.
The first difficulty at this stage was obviously deciding the appropriate method and resource to run the computations on. Working on local setup or on Google Colab was very expensive and consuming and infeasible when done naively. We decided to move to the Oscar supercomputing facility at Brown, where the setup got delayed a bit causing the work to slack off a bit. Meanwhile
I would be getting the computations and plots done during this week after getting familiarized with handling the clusters. The work is now on the cusp of experimentation, the plots would be giving an insight into the coherent states in the motion on which we would build our theory.
These 5 weeks have brought us to the midway stage of the program, where the mentors and students evaluate each other. The Week6 would be the evaluation week. From next week we would begin with the latter half of the project which would involve completing the remaining work, consolidating the code, documentation, reports and other deliverables.