Pipeline delay elimination with parallel two level primitive batch binning
A technique for rendering is provided. The technique includes for a set of primitives processed in a coarse binning pass, outputting early draw data to an early draw buffer; while processing the set of primitives in the coarse binning pass, processing the early draw data in a fine binning pass; and processing remaining primitives of the set of primitives in the fine binning pass.
1. A system comprising:
a world-space pipeline;
a screen-space pipeline;
an early draw buffer; and
a two-level primitive batch binner configured to:
for a set of primitives processed through the world-space pipeline in a coarse binning pass, output early draw data to the early draw buffer;
while processing the set of primitives in the coarse binning pass, in the world-space pipeline, processing the early draw data in a fine binning pass though the world-space pipeline; and
processing remaining primitives of the set of primitives in the fine binning pass in the world-space pipeline and in the screen-space pipeline.
2. The system of claim 1 , wherein outputting the early draw data comprises identifying early draw primitives of the set of primitives that overlap a first coarse binning tile and outputting the early draw primitives to the early draw buffer.
3. A system, comprising:
an early draw buffer; and
a two-level primitive batch binner configured to:
for a set of primitives processed through a world-space pipeline in a coarse binning pass, output early draw data to an early draw buffer;
while processing the set of primitives in the coarse binning pass, in the world-space pipeline, processing the early draw data in a fine binning pass through the world-space pipeline; and
processing remaining primitives of the set of primitives in the fine binning pass in the world-space pipeline and in a screen-space pipeline.
4. The system of claim 3 , wherein outputting the early draw data comprises identifying early draw primitives of the set of primitives that overlap a first coarse binning tile and outputting the early draw primitives to the early draw buffer.
5. The system of claim 3 , wherein processing the early draw data in the fine binning pass includes identifying which fine binning tiles are overlapped by the early draw data.
6. The system of claim 5 , wherein the two-level primitive batch binner is further configured to process the early draw data in fine binning tile order.
7. The system of claim 3 , wherein the two-level primitive batch binner is further configured to store a guard token in the early draw buffer that indicates the end of the early draw data.
8. The system of claim 7 , wherein the two-level primitive batch binner is further configured to process the remaining primitives in the fine binning pass after detecting the guard token in the early draw buffer.
9. The system of claim 3 , further comprising:
a plurality of rendering pipeline configured to perform the fine binning pass and the coarse binning pass concurrently.
10. The system of claim 3 , wherein outputting the early draw data to the early draw buffer occurs in a manner that allows reading the early draw data without any other data from the coarse binning pass.
11. The system of claim 3 , wherein reading cannot be performed for results of the coarse binning pass, other than the early draw data, when stored in a coarse binning buffer, until all results of the coarse binning pass are output to the coarse binning buffer.
12. A method for rendering, comprising:
for a set of primitives processed through a world-space pipeline in a coarse binning pass, outputting early draw data to an early draw buffer;
while processing the set of primitives in the coarse binning pass, in the world-space pipeline, processing the early draw data in a fine binning pass through the world-space pipeline; and
processing remaining primitives of the set of primitives in the fine binning pass in the world-space pipeline and in a screen-space pipeline.
13. The method of claim 12 , wherein outputting the early draw data comprises identifying early draw primitives of the set of primitives that overlap a first coarse binning tile and outputting the early draw primitives to the early draw buffer.
14. The method of claim 12 , wherein processing the early draw data in the fine binning pass includes identifying which fine binning tiles are overlapped by the early draw data.
15. The method of claim 14 , further comprising processing the early draw data in fine binning tile order.
16. The method of claim 12 , further comprising storing a guard token in the early draw buffer that indicates the end of the early draw data.
17. The method of claim 16 , further comprising processing the remaining primitives in the fine binning pass after detecting the guard token in the early draw buffer.
18. The method of claim 12 , further comprising:
performing the fine binning pass and the coarse binning pass concurrently across a plurality of rendering pipelines.
19. The method of claim 12 , wherein outputting the early draw data to the early draw buffer occurs in a manner that allows reading the early draw data without any other data from the coarse binning pass.
20. The method of claim 12 , wherein reading cannot be performed for results of the coarse binning pass, other than the early draw data, when stored in a coarse binning buffer, until all results of the coarse binning pass are output to the coarse binning buffer.