Thank you for your help, utkarsh. On Thu, May 14, 2009 at 8:19 AM, utkarsh shukla <utk.shukla@xxxxxxxxx> wrote: > I am not sure about this. Your work woulbe like to access the memory map and > find out where the tile is residing. Then Correspondingly you have to apply > the operation if it is CPU or GPU. This is an interesting architecture. Is the memory map a hash table where the GeglBuffer translates tiles into OpenGL textures? I guess we still have to provide a mechanism to specify whether or not a GeglBuffer should map tiles to the GPU if at all. This is useful in the case where a node has an attached GeglOperation that doesn't support OpenGL (i.e. doesn't implement a shader). We can let GPU-supporting GeglBuffers (meaning, buffers which can map tiles to the GPU) allocate GPU memory first and fall back to returning normal tiles when the allocation is unsuccessful. This is, however, a greedy approach and would prefer a better algorithm to decide whether or not a tile should be mapped to the GPU. > We need to talk upon the structure of the memory map.I also (not quiet sure) > was thinking that since the we are implementing the two level > parallelization so why not justcreate the memory map at the lowest level. I > mean if I put a rectangle into a GPU then I would write information like all > the tiles are in GPU instead of the rectangle in GPU. they can be grouped > but i feelIn the unified memory the tiles must be written. It would be > better if someone looks into this logic. I'm not sure if I'm getting you right. But I'm going to stab in the dark, anyway. :) What about with rectangles that don't fall on tile-boundaries? By that, I mean that a certain rectangle may start somewhere in the middle of the first row/column of tiles and/or end in the middle of the last row/column of tiles, like so: +-----+-----+-----+ | +-+-----+--+ | | |R|RRRRR|RR| | +---+-+-----+--+--+ | |R|RRRRR|RR| | | |R|RRRRR|RR| | +---+-+-----+--+--+ | |R|RRRRR|RR| | | +-+-----+--+ | +-----+-----+-----+ (the rectangle enclosing the Rs is the requested rectangle of pixels) Kind regards, Daerd _______________________________________________ Gegl-developer mailing list Gegl-developer@xxxxxxxxxxxxxxxxxxxxxx https://lists.XCF.Berkeley.EDU/mailman/listinfo/gegl-developer
