Graphics processors configured to interleave passes from different graphics tiles
When performing tile-based rendering a first, pre-pass operation in which primitives in a sequence of primitives for a tile are processed to determine visibility information for the sequence of primitives, the visibility information being usable to determine whether or not fragments for a primitive in the sequence of primitives should subsequently be processed further for the render output, is performed. Thereafter a second, main pass operation is performed in which the further processing of fragments for primitives that were processed during the first, pre-pass operation is controlled based on the determined visibility information for the sequence of primitives, such that for fragments for which the visibility information indicates that the fragments should not be processed further for the render output some or all of the processing during the second, main pass is omitted. Passes from different tiles can be interleaved.
1 . A method of operating a graphics processor to render sequences of primitives, wherein the graphics processor is configured to render a sequence of primitives for a render output by:
performing a first, pre-pass operation in which primitives in the sequence of primitives are processed into respective sets of one or more fragments, each fragment corresponding to one or more sampling positions within the render output, and in which first, pre-pass operation fragments are processed to determine visibility information for the sequence of primitives; and
thereafter performing a second, main pass operation in which primitives that were processed by the first, pre-pass operation are processed again to generate rendered output data for the render output, and in which the processing is controlled using the determined visibility information for the sequence of primitives;
the method comprising:
for a plurality of sequences of primitives to be rendered separately:
scheduling the rendering for the plurality of sequences of primitives such that the first, pre-pass operation for a first one of the plurality of sequences of primitives is performed between the respective first, pre-pass and corresponding second, main pass operations for a second one of the plurality of sequences of primitives,
wherein the scheduling is performed according to a desired scheduling operation in which after an initial set of two or more pre-pass operations are performed, the graphics processor then attempts to alternate between first, pre-pass operations and second, main pass operations for different sequences of primitives, and
wherein when the desired scheduling operation dictates that a pre-pass operation for a sequence of primitives should be performed, but there is no pre-pass operation available to be scheduled, whereas there is a main pass operation for another sequence of primitives that is ready to be performed as its corresponding pre-pass operation has already completed,
the method comprises:
determining whether one or more conditions are met to perform the main pass operation immediately.
2 . The method of claim 1 , wherein for a processing job comprising rendering a plurality of sequences of primitives, the method comprises initially scheduling two or more first, pre-pass operations for a respective two or more different sequences of primitives before performing the second, main pass operation for the first sequence of primitives for the processing job.
3 . The method of claim 2 , comprising thereafter alternating first, pre-pass operations and second, main pass operation for different sequences of primitives.
4 . The method of claim 1 , wherein determining whether one or more conditions are met to perform the main pass operation immediately comprises waiting for a time period to elapse, and if a pre-pass operation is not available to be scheduled within the time period, the method comprises performing the main pass operation immediately, whereas if a pre-pass operation is received within the time period, the method comprises scheduling the received pre-pass operation to be performed according to the desired scheduling operation.
5 . The method of claim 1 , wherein determining whether one or more conditions are met to perform the main pass operation immediately comprises checking the progress of a pre-pass operation that is currently being performed, and if the pre-pass operation has progressed beyond a threshold processing stage, the method comprises performing the main pass operation immediately, whereas if the pre-pass operation has not progressed beyond the threshold processing stage, the method comprises waiting for a pre-pass operation to become available and scheduling the pre-pass operation to be performed according to the desired scheduling operation.
6 . The method of claim 5 , wherein the threshold processing stage comprises a primitive classification stage that is performed prior to the processing of primitives into their associated fragments.
7 . The method of claim 1 , wherein when the desired scheduling operation is interrupted so that a main pass operation can be performed immediately, the method comprises re-starting the desired scheduling operation by issuing a set of two or more pre-pass operations back to back.
8 . The method of claim 1 , wherein the graphics processor is selectively operable either to interleave respective first, pre-pass and second, main pass operations for different sequences of primitives or to process different sequences of primitives in a strictly back to back manner, and wherein the method comprises an initial step of determining that the scheduling operation should be performed in an interleaved manner, wherein the scheduling of the rendering for the plurality of sequences of primitives such that the first, pre-pass operation for a first one of the plurality of sequences of primitives is performed between the respective first, pre-pass and corresponding second, main pass operations for a second one of the plurality of sequences of primitives is performed in response to determining that the scheduling operation should be performed in an interleaved manner.
9 . The method of claim 1 , wherein the graphics processor is configured to perform tile-based rendering in which a render output is subdivided into a plurality of tiles for rendering, and wherein the plurality of sequences of primitives correspond to respective rendering tiles.
10 . A graphics processor that is operable to render sequences of primitives, the graphics processor comprising a rendering circuit, and configured to:
for a sequence of primitives to be rendered for a render output:
cause the rendering circuit to:
perform a first, pre-pass operation in which primitives in the sequence of primitives are processed into respective sets of one or more fragments, each fragment corresponding to one or more sampling positions within the render output, and in which first, pre-pass operation fragments are processed to determine visibility information for the sequence of primitives; and
thereafter perform a second, main pass operation in which primitives that were processed by the first, pre-pass operation are processed again to generate rendered output data for the render output, and in which the processing is controlled using the determined visibility information for the sequence of primitives;
the graphics processing further comprising a scheduling circuit that is configured to schedule the rendering for different sequences of primitives by the rendering circuit, wherein the scheduling circuit is operable to interleave respective first, pre-pass and second, main pass operations for different sequences of primitives,
wherein the scheduling is performed according to a desired scheduling operation in which after an initial set of two or more pre-pass operations are performed, the graphics processor then attempts to alternate between first, pre-pass operations and second, main pass operations for different sequences of primitives, and
wherein when the desired scheduling operation dictates that a pre-pass operation for a sequence of primitives should be performed, but there is no pre-pass operation available to be scheduled, whereas there is a main pass operation for another sequence of primitives that is ready to be performed as its corresponding pre-pass operation has already completed,
the scheduling circuit is configured to:
determine whether one or more conditions are met to perform the main pass operation immediately.
11 . The graphics processor of claim 10 , wherein for a processing job comprising rendering a plurality of sequences of primitives, the scheduling circuit is configured to initially schedule two or more first, pre-pass operations for a respective two or more different sequences of primitives before performing the second, main pass operation for the first sequence of primitives for the processing job.
12 . The graphics processor of claim 11 , wherein the scheduling circuit is configured to thereafter alternate scheduling first, pre-pass operations and second, main pass operation for different sequences of primitives.
13 . The graphics processor of claim 10 , wherein determining whether one or more conditions are met to perform the main pass operation immediately comprises waiting for a time period to elapse, and if a pre-pass operation is not available to be scheduled within the time period, the scheduling circuit is configured to perform the main pass operation immediately, whereas if a pre-pass operation is available within the time period, the scheduling circuit is configured to schedule the received pre-pass operation to be performed according to the desired scheduling operation.
14 . The graphics processor of claim 10 , wherein determining whether one or more conditions are met to perform the main pass operation immediately comprises checking the progress of a pre-pass operation that is currently being performed, and if the pre-pass operation has progressed beyond a threshold processing stage, the scheduling circuit is configured to perform the main pass operation immediately, whereas if the pre-pass operation has not progressed beyond the threshold processing stage, the scheduling circuit is configured to wait for a pre-pass operation to become available and to schedule the pre-pass operation to be performed according to the desired scheduling operation.
15 . The graphics processor of claim 10 , wherein when a main pass operation is performed immediately, such that the desired scheduling operation is interrupted, the scheduling circuit is configured to re-start the desired scheduling operation by issuing a set of two or more pre-pass operations back to back.
16 . The graphics processor of claim 10 , wherein the scheduling circuit is selectively operable either to interleave respective first, pre-pass and second, main pass operations for different sequences of primitives or to process different sequences of primitives in a strictly back to back manner.
17 . The graphics processor of claim 10 , wherein the graphics processor is configured to perform tile-based rendering in which a render output is subdivided into a plurality of tiles for rendering, and wherein the plurality of sequences of primitives correspond to respective rendering tiles.
18 . A non-transitory computer program product comprising a set of instructions that when executed by one or more data processor will cause the one or more processor to perform a method of operating a graphics processor to render sequences of primitives for a render output, wherein the graphics processor is configured to render a sequence of primitives by:
performing a first, pre-pass operation in which primitives in the sequence of primitives are processed into respective sets of one or more fragments, each fragment corresponding to one or more sampling positions within the render output, and in which first, pre-pass operation fragments are processed to determine visibility information for the sequence of primitives; and
thereafter performing a second, main pass operation in which primitives that were processed by the first, pre-pass operation are processed again to generate rendered output data for the render output, and in which the processing is controlled using the determined visibility information for the sequence of primitives;
the method comprising:
for a plurality of sequences of primitives to be rendered separately:
scheduling the rendering for the plurality of sequences of primitives such that the first, pre-pass operation for a first one of the plurality of sequences of primitives is performed between the respective first, pre-pass and corresponding second, main pass operations for a second one of the plurality of sequences of primitives,
wherein the scheduling is performed according to a desired scheduling operation in which after an initial set of two or more pre-pass operations are performed, the graphics processor then attempts to alternate between first, pre-pass operations and second, main pass operations for different sequences of primitives, and
wherein when the desired scheduling operation dictates that a pre-pass operation for a sequence of primitives should be performed, but there is no pre-pass operation available to be scheduled, whereas there is a main pass operation for another sequence of primitives that is ready to be performed as its corresponding pre-pass operation has already completed,
the method comprises:
determining whether one or more conditions are met to perform the main pass operation immediately.