Open iluise opened 1 month ago
Modify the atmorep_read_data interface to run with DASK arrays that can be easily parallelised. \ Bottleneck at the moment: we run out of memory when reconstructing more than 45000 local samples. This means ~ 20 dates in global forecast.
other possible alternative: https://developer.nvidia.com/blog/rapids-cudf-accelerates-pandas-nearly-150x-with-zero-code-changes/
Modify the atmorep_read_data interface to run with DASK arrays that can be easily parallelised. \ Bottleneck at the moment: we run out of memory when reconstructing more than 45000 local samples. This means ~ 20 dates in global forecast.