The backend database currently maintains an []Entry with large tuples containing a large amount of repetitive information, namely the lifter and meet information. Regardless of how the data is stored in CSV files, for purposes of the server, it would be significantly more efficient to process the data into separate {Meet, Entry, Lifter} structs with links between them.
That would decrease the size of []Entry, which means that more Entry structs fit on a cache line, so the leaderboard logic gets faster. Searching for meets can also be over []Meet, which is a much more efficient representation.
The backend database currently maintains an
[]Entrywith large tuples containing a large amount of repetitive information, namely the lifter and meet information. Regardless of how the data is stored in CSV files, for purposes of the server, it would be significantly more efficient to process the data into separate{Meet, Entry, Lifter}structs with links between them.That would decrease the size of
[]Entry, which means that more Entry structs fit on a cache line, so the leaderboard logic gets faster. Searching for meets can also be over[]Meet, which is a much more efficient representation.