Smooth schedule of large-scale terrain visualization from external memory

The paper proposes a smooth schedule algorithm to render terrain data out-of-core. The terrain data is organized as a quad-tree with Z-order filling curve, which preserves the data continuity and improved the schedule efficiency. A schedule strategy basing memory allocation is designed to support the stable frame rate terrain rendering. Controllable schedule area, instead of the view frustum culling, is applied to achieve the balance between the in-core memory requirements and scheduling time. A pre-estimated approach is adopted to smooth the data exchange. The algorithm achieves a smooth schedule and successfully avoids the slow or jerky phenomena, usually caused by inefficient and unsmooth schedule. By compared with traditional methods, the algorithm proposed is feasible and efficient in out-of-core terrain visualization.