The Z160 implements the full OpenVG™ 1.1 pipeline in hardware and includes both geometry transformation and pixel processing. The core can directly render curved lines without intermediate software processing. The GPU also frees up the ARM CPU from unnecessary processing by handling many of its own transactions to memory. The core has a built in MMU and DMA capability to allow it to fetch data and commands from DRAM without CPU intervention. The drivers running on the CPU need to merely write commands to a circular queue in DRAM and the GPU does the rest. Power savings are seen by a number of clock gating features, including the separation of the basic 2D and vector graphics functionality. When UIs are only using BLT transfers, the vector portion will clock gate to power consumption of the core. The high quality rendering of the core is achieved by anti-aliasing techniques. The Z160 supports a 16x multi-sample AA when rendering lines with no performance penalty. The core also supports an additional 4x full scene AA ‘rotated grid super sampling’ technique on top of that.