Blitrix ARM Neon optimized 2D blitting engine

This demo showcases a software–based 2D accelerator for embedded User Interfaces utilizing an ARM Neon CPU. Blitrix enables a rich GUI experience without costly investment in accelerated hardware. It offers a perfect fit for touch-screen phones, TV-UIs (up to 1080p), and any display menus on devices.

Blitrix provides real-time compilation by stitcher architecture with the following benefits:

Speeds up blitting (bitmap rendering) operations
Removes the need for hand-optimizing fast paths for specific use cases
Enables smooth HW upgrades & variations
Utilizes ARM Neon instruction set, such as an ARM Cortex platform
Very fast - eliminates per pixel branching
Nearly memory-bound on most platforms
Tiny footprint (< 200 kB) and is easily extended and maintained
Designed to efficiently handle and perform multiple operations in a single pass. It can generate highly optimized code-paths for operations such as taking a YUV video surface as input, rotating and scaling it to fit the screen, as well as perform the conversion to RGB, all in a single pass.
Blitrix saves memory by avoiding the allocation of temporary buffers for the results of each individual operation. It also makes the operation faster by reducing unnecessary accesses to the temporary surfaces.