In data compression, inverting a BWT with large blocks or using Context Mixing to compress large blocks (which requires huge context maps). These 2 cases requir

In data compression, inverting a BWT with large blocks or using Context Mixing to compress large blocks (which requires huge context maps). These 2 cases require a lot of random memory accesses.