A cache management system providing improved page latching methodology. A method providing access to data in a multi-threaded computing system comprises: providing a cache containing data pages and a mapping to pages in memory of the multi-threaded computing system; associating a latch with each page in cache to regulate access, the latch allowing multiple threads to share access to the page for reads and a single thread to obtain exclusive access to the page for writes; in response to a request from a first thread to read a particular page, determining whether the particular page is in cache without acquiring any synchronization object regulating access and without blocking access by other threads; if the particular page is in cache, reading the particular page unless another thread has exclusively latched the particular page; and otherwise, if the particular page is not in cache, bringing the page into cache.