I try a lot of different experiments with git performance, some of them more hare-brained than others. The ones that succeed end up as real patches. But I hate for the ones that fail to die a quiet death. Then nobody learns what _doesn't_ work, and nobody has the opportunity to point out the spot where I made a stupid mistake that invalidates the whole result.
Here are two performance experiments that did _not_ turn out. They are in patch form, because I want to document exactly what I did (especially because it helps in the "stupid mistake" case, or if somebody wants to try building on my results). So without further ado: [1/2]: obj_hash: convert to a critbit tree [2/2]: use zstd zlib wrapper -Peff