Huge performance improvement

This MR adds a simple optimization which improves performance by huge margins when performing operations while zoomed in. The cache grid works by storing tiles of fixed size using an LRU policy (least recently used). When zoomed in, querying a region in the cache grid would previously iterate through all possible cache cells even though only the last few cells would be required due to the LRU policy (e.g., if max size is 12 cells, then querying more than the last 12 cells does not make sense). This prevents the app from freezing in some scenarios.