To start with, Raphaël CC's kernel can finish the computation of 5 nearest atoms' Coulomb interaction in just 8 hours for all molecules and for H, C, O, N, F atoms.
However, if you wanna compute 8 or 9 nearest atoms, this kernel will time out since there are more than 10 H and C atoms in some molecules. The biggest reason is that pandas uses only 1 CPU core when manipulating dataframes. So I made a parallel version of it, where you can do even more complicated computation like Yukawa potentials for all molecules in under 4 hours.
To use the output file, simple concat or merge it with the original structures, then merge it with the train and test dataframes, and you are good to go.
", "messages": [ { "id": "572272", "postDate": "07/10/2019 17:55:20", "content": "To start with, Raphaël CC's kernel can finish the computation of 5 nearest atoms' Coulomb interaction in just 8 hours for all molecules and for H, C, O, N, F atoms.
However, if you wanna compute 8 or 9 nearest atoms, this kernel will time out since there are more than 10 H and C atoms in some molecules. The biggest reason is that pandas uses only 1 CPU core when manipulating dataframes. So I made a parallel version of it, where you can do even more complicated computation like Yukawa potentials for all molecules in under 4 hours.
To use the output file, simple concat or merge it with the original structures, then merge it with the train and test dataframes, and you are good to go.
", "rawMarkdown": "To start with, [Raphaël CC's kernel](https://www.kaggle.com/daijin12/coulomb-interaction-high-perf-no-loop-almost) can finish the computation of 5 nearest atoms' Coulomb interaction in just 8 hours for all molecules and for `H`, `C`, `O`, `N`, `F` atoms. \n\nHowever, if you wanna compute 8 or 9 nearest atoms, this kernel will time out since there are more than 10 `H` and `C` atoms in some molecules. The biggest reason is that `pandas` uses only 1 CPU core when manipulating dataframes. So I made a [parallel version of it](https://www.kaggle.com/scaomath/parallelization-of-coulomb-yukawa-interaction), where you can do even more complicated computation like Yukawa potentials for all molecules in under 4 hours.\n\nTo use the output file, simple concat or merge it with the original structures, then merge it with the train and test dataframes, and you are good to go.", "votes": null }, { "id": "573983", "postDate": "07/13/2019 04:31:30", "content": "Thanks it helps to run faster and its good for old processors which have low single CPU performance but good multi CPU.
", "rawMarkdown": "Thanks it helps to run faster and its good for old processors which have low single CPU performance but good multi CPU.", "votes": null } ], "comments": [ { "id": 573983, "author_name": "vaishvik25", "author_url": "", "post_date": "07/13/2019 04:31:30", "content": "Thanks it helps to run faster and its good for old processors which have low single CPU performance but good multi CPU.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "572272": "To start with, [Raphaël CC's kernel](https://www.kaggle.com/daijin12/coulomb-interaction-high-perf-no-loop-almost) can finish the computation of 5 nearest atoms' Coulomb interaction in just 8 hours for all molecules and for `H`, `C`, `O`, `N`, `F` atoms. \n\nHowever, if you wanna compute 8 or 9 nearest atoms, this kernel will time out since there are more than 10 `H` and `C` atoms in some molecules. The biggest reason is that `pandas` uses only 1 CPU core when manipulating dataframes. So I made a [parallel version of it](https://www.kaggle.com/scaomath/parallelization-of-coulomb-yukawa-interaction), where you can do even more complicated computation like Yukawa potentials for all molecules in under 4 hours.\n\nTo use the output file, simple concat or merge it with the original structures, then merge it with the train and test dataframes, and you are good to go.", "573983": "Thanks it helps to run faster and its good for old processors which have low single CPU performance but good multi CPU." }, "source": "meta" }