Currently, we can trim the tries that are contained in checkpoints, but it seems like it does not save much memory to do so while coming at a huge cost in terms of time. Using the mainnet-14 checkpoint, it takes 4 hours to trim the trie (rebuild it using extensions and such, instead of just having branches, basically).
We should investigate whether it's worth trimming the trie, because it seems like it does have an impact but that it is smaller than
initially expected (due to wrong testing parameters).
Description
Currently, we can trim the tries that are contained in checkpoints, but it seems like it does not save much memory to do so while coming at a huge cost in terms of time. Using the mainnet-14 checkpoint, it takes 4 hours to trim the trie (rebuild it using extensions and such, instead of just having branches, basically).
We should investigate whether it's worth trimming the trie, because it seems like it does have an impact but that it is smaller than initially expected (due to wrong testing parameters).