Graphic processing apparatus and method of performing graphics pipeline
A graphic processing apparatus and a method of performing a graphics pipeline in the graphic processing apparatus are provided. The method of performing a graphics pipeline in a graphic processing apparatus includes binning to generate a bounding box bitstream corresponding to a drawcall requiring tessellation, and in response to a bounding box allocated by the bounding box bitstream being identified in a current tile to be processed, rendering the current tile by performing selective tessellation on drawcalls corresponding to the identified bounding box.
1. A method of performing a graphics pipeline in a graphic processing apparatus, the method comprising:
determining, by a binning pipeline, whether each of a plurality of input drawcalls is a first-type drawcall requiring tessellation or a second-type drawcall requiring no tessellation;
performing, by the binning pipeline, tile binning to identify a tile list covering a bounding box corresponding to the first-type drawcall;
generating, by the binning pipeline, a bounding box bitstream, the bounding box bitstream comprising identification information indicating the first-type drawcall, identification information indicating the second-type drawcall, and information about the tile list; and
performing, by a rendering pipeline, a tessellation, wherein the tessellation is only performed on tiles including the first-type drawcall,
wherein when a first tile includes the first-type drawcall, primitives of the first tile are tessellated, and when a second tile includes the first-type drawcall and the same primitives as the first tile, the second tile is not tessellated.
2. The method of claim 1 , further comprising generating, by the binning pipeline, a drawcall bitstream comprising, for each of the tiles, information indicating whether at least one of the input drawcalls is present in the tile, and a primitive bitstream comprising, for each of the tiles, information indicating a number of primitives included in the input drawcalls.
3. The method of claim 1 , wherein the tessellation is deferred from being performed in the binning pipeline.
4. The method of claim 1 , wherein the first-type drawcall is processed by a first sub rendering pipeline including a tessellation stage, and the second-type drawcall is processed by a second sub rendering pipeline not including the tessellation stage.
5. The method of claim 1 , further comprising:
scheduling, by the rendering pipeline, a rendering order of the tiles based on the bounding box bitstream, such that a higher priority is assigned to a tile covering the bounding box; and
performing, by the rendering pipeline, rendering on a per-tile basis according to the scheduled rendering order.
6. The method of claim 1 , further comprising deleting data associated with tessellated primitives from a memory, when tessellated primitives for the first-type drawcall are not used in other tiles.
7. A non-transitory computer-readable recording medium comprising computer readable code to implement the method of claim 1 .
8. A computing apparatus for performing a graphics pipeline, the computing apparatus comprising:
a graphic processing unit (GPU) configured to:
determine, by a binning pipeline, whether each of a plurality of input drawcalls is a first-type drawcall requiring tessellation or a second-type drawcall requiring no tessellation;
perform, by the binning pipeline, tile binning to identify a tile list covering a bounding box corresponding to the first-type drawcall;
generate, by the binning pipeline, a bounding box bitstream, the bounding box bitstream comprising identification information indicating the first-type drawcall, identification information indicating the second-type drawcall, and information about the tile list; and
perform, by a rendering pipeline, a tessellation, wherein the tessellation is only performed on each tile including the first-type drawcall,
wherein when a first tile includes the first-type drawcall, primitives of the first tile are tessellated, and when a second tile includes the first-type drawcall and the same primitives as the first tile, the second tile is not tessellated; and
a memory configured to store the bounding box bitstream generated in the binning pipeline and to provide the stored bounding box bitstream to the rendering pipeline.
9. The computing apparatus of claim 8 , wherein; the GPU is further configured to generate a drawcall bitstream comprising, for each of the tiles, information indicating whether at least one of the input drawcalls is present in the tile, and a primitive bitstream comprising, for each of the tiles, information indicating a number of primitives included in the input drawcalls.
10. The apparatus of claim 8 , wherein the GPU is further configured to defer the tessellation from being performed in the binning pipeline.
11. The apparatus of claim 8 , wherein the first-type drawcall is processed by a first sub rendering pipeline including a tessellation stage, and the second-type drawcall is processed by a second sub rendering pipeline not including the tessellation stage.
12. The apparatus of claim 8 , wherein the GPU is further configured to:
schedule a rendering order of the tiles based on the bounding box bitstream, such that a higher priority is assigned to a tile covering the bounding box, and
perform rendering on a per-tile basis according to the scheduled rendering order.
13. The apparatus of claim 8 , wherein the memory is further configured to delete data associated with tessellated primitives while the tessellated primitives for the first-type drawcall are not used in other tiles.