GitHub conducted an experiment to assess the performance impact of various git clone options on both client and server environments, focusing on full, shallow, treeless, and blobless clones. Using three open-source repositories—jquery/jquery, apple/swift, and torvalds/linux—the study aimed to quantify differences in cloning and subsequent fetching operations. Results indicated that while shallow clones were the fastest for clients, they increased computational load on servers due to the lack of performance optimizations for shallow boundaries. Treeless and blobless clones reduced network data transfer but increased Git CPU costs, with blobless clones being more server-intensive. The experiment highlighted that clone and fetch performance varies significantly with repository size and structure, and recommended that developers might favor full clones for comprehensive access, while CI builds could benefit from shallow or treeless clones for efficiency. The study suggested that real-world applications might yield different results, and encouraged GitHub Enterprise Server admins to replicate these tests to better understand their specific workflows.