# US 2022/0036501 A1 · Hierarchical tiling in a graphics processing system using chain sorting of primitives

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0036501 A1
- Document ID: US20220036501A1
- Status: Published application
- Kind: A1
- Application number: 17/504,813
- Filed: 2021-10-19
- Published: 2022-02-03
- Earliest claimed: 2019-09-19
- Assignees: Imagination Technologies Limited
- Applicants: Imagination Technologies Limited
- Inventors: Robert Brigg; Lorenzo Belli
- Web page: https://patentgrep.com/patent/US20220036501A1

## CPC

- G06T 1/60: Memory management
- G06T 11/40: Filling planar surfaces by adding surface attributes, e.g. adding colours or textures
- G06T 1/20: Processor architectures; Processor configuration, e.g. pipelining
- G06T 15/005: {General purpose rendering architectures}
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Tiling engines and methods for hierarchically tiling a plurality of primitives. A chain of sorting units includes a top level sorting unit followed by lower level sorting units, the top level sorting unit determining which of a plurality of regions of a render space each of the plurality of primitives at least partially falls within. For each region a primitive at least partially falls within an identifier of that primitive is stored in a queue. Each of the lower level sorting units selects queues of a preceding sorting unit in the chain to process, and determines which of a plurality of sub-regions of the region associated with that queue each of the primitives at least partially falls within. For each such sub-region an identifier of that primitive is stored in a queue of the lower level sorting unit that is associated with that sub-region. An output unit outputs the primitives identified in the queues of the last lower level sorting unit in the chain on a queue by queue basis.

## Claims

### Claim 1

A method of storing transformed primitives in a geometry processing phase implemented by a graphics processing system, the method comprising:
maintaining a primitive block for each of a plurality of macro regions of a render space;
receiving a transformed primitive in the render space;
determining which of the plurality of macro regions the transformed primitive at least partially falls within;
adding the transformed primitive to at least one primitive block associated with a macro region that the transformed primitive at least partially falls within;
for each of the at least one primitive block that the transformed primitive is added to, determining whether the primitive block is full; and
in response to determining that a primitive block is full, storing the primitive block in a local transformed geometry buffer of the graphics processing system for use in a rasterization phase implemented by the graphics processing system, and starting a new primitive block for the macro region associated with that primitive block.

### Claim 2 (depends on 1)

The method of claim 1, further comprising, if the transformed primitive at least partially falls within at least two macro regions, selecting a transformed primitive block associated with one of the at least two macro regions to add the transformed primitive to.

### Claim 3 (depends on 1)

The method of claim 1, wherein the transformed primitive is added to the primitive block associated with each of the macro regions that the transformed primitive at least partially falls within.

### Claim 4 (depends on 1)

The method of claim 1, further comprising maintaining a reference counter for each primitive block stored in the local transformed geometry buffer that indicates whether the primitive block can be removed from the local transformed geometry buffer.

### Claim 5 (depends on 1)

The method of claim 1, wherein a primitive block associated with a macro region is determined to be full when:
that primitive block has a maximum size,
that primitive block comprises a maximum number of transformed primitives, and/or
each transformed primitive is associated with one or more vertices and that primitive block comprises a maximum number of vertices.

### Claim 6 (depends on 1)

The method claim 1, wherein a memory of the local transformed geometry buffer is divided into a plurality of chunks and a primitive block is only stored in a same chunk of the local transformed geometry buffer as another primitive block when the primitive block and the other primitive block are associated with a same macro region of the plurality of macro regions.

### Claim 7 (depends on 6)

The method of claim 6, further comprising maintaining a reference counter for each chunk that identifies whether data in that chunk can be overwritten.

### Claim 8 (depends on 1)

The method of claim 1, wherein:
a memory of the local transformed geometry buffer is divided into a plurality of physical chunks of size K;
a virtual memory is divided into a plurality of virtual memory chunks of size (M+1)*K, each virtual memory chunk being mapped to (M+1) physical chunks, wherein a maximum size of a primitive block is M*K wherein M is an integer greater than or equal to two; and
storing a primitive block in the local transformed geometry buffer comprises writing the primitive block to a virtual memory chunk only if the primitive block starts in a first K-sized block of the virtual memory chunk.

### Claim 9 (depends on 1)

The method of claim 1, wherein the rendering space is divided into a plurality of tiles and each macro region comprises at least two of the plurality of tiles.

### Claim 10 (depends on 1)

The method of claim 1, further comprising, prior to storing the primitive block in the local transformed geometry buffer, determining if there is enough free space in the local transformed geometry buffer to store the primitive block, and only storing the primitive block in the local transformed geometry buffer if it is determined that there is enough free space.

### Claim 11 (depends on 1)

The method of claim 1, wherein adding a transformed primitive to a primitive block comprises adding transformed geometry data related to the transformed primitive to the primitive block.

### Claim 12 (depends on 11)

The method of claim 11, wherein the transformed geometry data related to a transformed primitive comprises transformed vertex information for each of a plurality of vertices.

### Claim 13 (depends on 1)

The method of claim 1, wherein the graphics processing system is a hybrid rendering graphics processing system.

### Claim 14

A storing unit for use in a graphics processing system, the storing unit comprising logic configured to, in a graphics processing phase implemented by a graphics processing system:
maintain a primitive block for each of a plurality of macro regions of a render space;
receive a transformed primitive in render space;
determine which of the plurality of macro regions the transformed primitive at least partially falls within;
add the transformed primitive to at least one primitive block associated with a macro region that the primitive at least partially falls within;
for each of the at least one primitive block that the transformed primitive is added to, determine whether the primitive block is full; and
in response to determining that a primitive block is full, store the primitive block in a local transformed geometry buffer of the graphics processing system for use in a rasterization phase implemented by the graphics processing system and start a new primitive block for the macro region associated with that primitive block.

### Claim 15 (depends on 14)

A tiling engine for use in a graphics processing system, the tiling engine comprising:
the storing unit as set forth in claim 14; and
a chain of sorting units configured to sort transformed primitives into successively smaller regions of the render space using a set of queues, each sorting unit in the chain configured to sort transformed primitives into a plurality of regions by storing in a queue associated with each of the plurality of regions, information identifying the primitive blocks that comprise the transformed primitives that at least partially fall within that region.

### Claim 16 (depends on 15)

The tiling engine of claim 15, wherein the information identifying a primitive block comprises information identifying a location of the primitive block in the local transformed geometry buffer.

### Claim 17 (depends on 15)

The tiling engine of claim 15, wherein the chain of sorting units is configured to remove information identifying a primitive block from a queue when the information identifying the primitive block is added to one or more other queues or when the information identifying a primitive block is sent to rasterization logic of the graphics processing unit for processing a tile, and the tiling engine is configured to maintain a reference counter for each primitive block stored in the local transformed geometry buffer that indicates when that primitive block can be removed from the local transformed geometry buffer by:
incrementing the reference counter for a primitive block when information identifying that primitive block is added to a queue;
decrementing the reference counter for a primitive block when information identifying the primitive block is removed from a queue;
incrementing the reference counter for a primitive block when information identifying the primitive block is sent to the rasterization logic for processing a tile; and
decrementing a reference counter for a primitive block when the rasterization logic completes processing of the primitive block for a tile.

### Claim 18 (depends on 14)

A graphics processing system comprising the storing unit as set forth in claim 14.

### Claim 19 (depends on 1)

A non-transitory computer readable storage medium having stored thereon computer readable instructions that, when executed at a computer system, cause the computer system to perform the method as set forth in claim 1.

### Claim 20 (depends on 14)

A non-transitory computer readable storage medium having stored thereon a computer readable dataset description of the storing unit as set forth in claim 14 that, when processed in an integrated circuit manufacturing system, causes the integrated circuit manufacturing system to manufacture an integrated circuit embodying the storing unit.
