Disclosed in some examples are techniques for handling parity data of a non-volatile memory device with limited cache memory. In certain examples, user data can be programmed into the non-volatile memory of the non-volatile memory device in data stripes, and parity information can be calculated for each individual data stripe within a limited capacity cache of the non-volatile memory device. The individual parity information can be swapped between a swap block of the non-volatile memory and the limited capacity cache as additional data stripes are programmed.