Scheduling cache traffic in a tile-based architecture

A computing device for processing graphics data, the computing device comprising:

• a cache unit;
• an external memory unit;
• and
• a graphics processing pipeline comprising:
• a first screen-space pipeline;
• and
• a first tiling unit that:
• determines that a first set of primitives overlaps a first cache tile,
• generates a prefetch command for a second cache tile upon determining that a second set of primitives overlaps the second cache tile, wherein the prefetch command for the second cache tile causes the cache unit to fetch data associated with the second set of primitives from an external memory unit,
• transmits the prefetch command for the second cache tile to the first screen-space pipeline for processing,
• after transmitting the prefetch command for the second cache tile to the first screen-space pipeline, transmits the first set of primitives to the first screen-space pipeline for processing, wherein the first set of primitives is processed by the first screen-space pipeline before the second set of primitives is processed by the first screen-space pipeline, and
• after transmitting the first set of primitives to the first screen-space pipeline but before transmitting the second set of primitives to the first screen-space pipeline, transmits a first flush command for the first cache tile to the first screen-space pipeline.

A method for processing graphics data, the method comprising:

• determining that a first set of primitives overlaps a first cache tile,
• generating a prefetch command for a second cache tile upon determining that a second set of primitives overlaps the second cache tile, wherein the prefetch command for the second cache tile causes the cache unit to fetch data associated with the second set of primitives from an external memory unit,
• transmitting the prefetch command for the second cache tile to the first screen-space pipeline for processing,
• after transmitting the prefetch command for the second cache tile to the first screen-space pipeline, transmitting the first set of primitives to the first screen-space pipeline for processing, wherein the first set of primitives is processed by the first screen-space pipeline before the second set of primitives is processed by the first screen-space pipeline, and
• after transmitting the first set of primitives to the first screen-space pipeline but before transmitting the second set of primitives to the first screen-space pipeline, transmitting a first flush command for the first cache tile to the first screen-space pipeline.

A tile-based system for processing graphics data, the system comprising:

• a first screen-space pipeline;
• a cache unit;
• and
• a first tiling unit that:
• determines that a first set of primitives overlaps a first cache tile,
• generates a first prefetch command upon determining that a second set of primitives overlaps a second cache tile, wherein the first prefetch command causes the cache unit to fetch data from an external memory unit for the second cache tile,
• transmits the first set of primitives and the first prefetch command to the first screen-space pipeline for processing,
• after transmitting the first set of primitives to the first screen-space pipeline but before transmitting the second set of primitives to the first screen-space pipeline, transmits a first flush command to the screen-space pipeline for processing with the first set of primitives, wherein the first flush command is configured to cause the cache unit to flush data associated with the first cache tile,
• transmits the second set of primitives to the first screen-space pipeline for processing.