ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
Team Hackathon
2
Date/time: Friday 14 July (9:00 - 17:30)
3
Location: Intel Lab in the Computer Science and Technology Department
4
5
Please enter details below to pitch a project idea or join an existing team.
6
Please feel free to suggest ideas even if you do not wish to lead the team or are not able to participate.
7
Projects should be of a size such that reasonable progress can be made in the course of a day with a team of 4-5 people (e.g., initial prototype, demonstration of idea, initial data analysis and preliminary results).
8
9
Team number123456
10
Team name
11
Short titleDeLorean Datasets
The curious case of Fortran-Torch data transfer.
Coupling ML models to tneXtSIM-DGReprohackPorting a small serial Julia code to GPUNumba GPU in PySDM
12
Summary of ideapython xarray, python parallelization, shell parallelization , SLURM on Cambridge HPC (alternatively different more optimized libraries than xarray: ecCodes, cdo)Benchmarking fortran-pytorch-lib to understand why we still seem to be slower than forpy and hopefully speed up.Melt ponds play a crucial role in the Arctic's energy budget and climate as a whole. Despite this they are not represented/parametrised in almost all climate models around the world. neXtSIM-DG is a state of the art sea ice model being developed by the SASIP project as part of VESRI.Choose a relevant climate paper from https://www.reprohack.org/paper/ and undertake a reproducibility study / or find another paper that people want to reproduce.I have a small Julia code which solves the two-dimensioanl advection equation within a data structure which is used in spectral element codes. We have to develop some new data structures and replace part of the lopp structures.PySDM is a package for simulating the dynamics of population of particles. It already has some device specific code for CPU and GPU, we can add another GPU implementation written with Numba CUDA
13
Anticipated languages / frameworks / datasetsFortran Pytorch fortran-pytorch-lib benchmarking codesPython/Tensorflow/Pytorch/C++Depends on paperJulia, some Cuda stuffPython, Numba, CUDA programming model
14
Links to any resources / existing codeGithub with existing code: https://github.com/elkir/ph6-02_ens-extended_analysis/tree/main/src
Presentation explaining the dataset: https://docs.google.com/presentation/d/1T21QLVIppK29Z_OKH8CJEnnPvkMYY-TkoPQN_bbauHk/edit?usp=sharing
https://github.com/Cambridge-ICCS/fortran-pytorch-lib-benchmark
We propose to input a basic observational emulator of sea ice melt ponds into the neXtSIM-DG as a proof of concept, understanding and discovering the software challenges of coupling observational emulators into C++ models, and thus creating the world's first sea ice model that includes a data-driven emulator of melt ponds.https://www.reprohack.org/paper/https://github.com/OsKnoth/Advection2D.jl.githttps://github.com/open-atmos/PySDM/tree/main
15
16
Team lead (name and email)Petr DolezalJack Atkinson (jwa34@cam.ac.uk)Simon DriscollTBDOswald KnothOleksii Bulenok, a.bulenok@student.uj.edu.pl
17
18
Team members (name and email) - add yourself hereChris Edsall cje57@cam.ac.ukSophie Turner, st838@cam.ac.ukChris Kent (chris.kent@metoffice.gov.uk)Tom Meltzer <tdm39@cam.ac.uk>
19
Elliott Kasoar elliott.kasoar@stfc.ac.ukDavid Kamm, david.kamm@locean.ipsl.frIvo Pasmans, i.c.pasmans@reading.ac.uk"Paul Richmond
par62@cam.ac.uk"
Kacper Kornet (kk562@cam.ac.uk)
20
Tianzhang Cai (online)
jultcai@gmail.com		Felix Jochum, fjochum@iau.uni-frankfurt.deBen Orchard (bhgo2@cam.ac.uk)
21
Sam Madge,
sam.madge@informaticslab.co.uk		Simon Clifford
sjc306
22
23
Prizes1st2nd3rd
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100