80 likes | 199 Views
Sort Last Parallel Rendering for Viewing Extremely Large Data Sets on Tile Displays (PVG 2001). Kenneth Moreland, Sandia National Labs. Summary. General idea - N processors running a T tile display generate T images, 1 for each tile
E N D
Sort Last Parallel Rendering for Viewing Extremely Large Data Sets on Tile Displays (PVG 2001) Kenneth Moreland, Sandia National Labs
Summary • General idea - N processors running a T tile display generate T images, 1 for each tile • Images are composite and displayed at the processors controlling the tiles • Polygons to be rendered at distributed on N processors and projection information is scattered, telling the processors which tiles their images should go to
4 strategies Four different strategies are described for composition 1. Serial (every node in charge of a tile generates T images for that tile and composes) – worst case algorithm. 2. Virtual trees 3. Tile, Split and Delegate 4. Reduce to a single tile
Virtual Trees • Compositing is done in several binary trees in parallel • The tiles done with compositing drop the computation and join other trees • The scheduling is done so that processors with least number of images to send act as receivers and vice versa • A disadvantage is that during the final stage of compositing, most processors are idle
Tile, Split and Delegate • Assign a processor to a section of a tile • More processors are assigned for tiles which require more image composition • A disadvantage is communication cost is high - O(N2)
Reduce to Single Tile • Images rendered at any processor are sent directly to a single processor (for each tile) • A binary swap algorithm is used to composite them • Communication time – O(N*T + NlogN) – more scalable
Optimization • Bucketing – do not distribute all polygons to everyone – tiles with very few polygons do not need rendering of all other polygons • Active pixel encoding • the blank pixels are termed inactive • RLE compression is used • free decoding (use a max depth value for inactive pixels
Optimization, Results • Floating viewport – render all polygons on one node and shift pixels when image straddles other tiles Results • Reduce strategy performs better and scales with a linear speedup