High vertex count geometry work distribution for multi-tile GPUs

Patent metadata

Field	Value
Publication number	US-10733693-B2
Application number	US-201816208715-A
Country	US
Kind code	B2
Filing date	Dec 4, 2018
Priority date	Dec 4, 2018
Publication date	Aug 4, 2020
Grant date	Aug 4, 2020

Abstract

Embodiments described herein provide data processing device comprising a processor, a memory, and a large draw monitor comprising a processing unit to determine whether a vertex count for a graphics workload exceeds a threshold value, and in response to a determination that the vertex count for the graphics workload exceeds the threshold value, to divide the graphics workload over graphics processing units instantiated on multiple separate tiles. Other embodiments may be described and claimed.

First claim

What is claimed is:

1. A general purpose graphics processing device, comprising:
a general purpose graphics processing compute block to process a workload including graphics or compute operations, the general purpose graphics processing compute block comprising at least a first graphics processing unit tile and a second graphics processing unit tile;
a geometry utilization monitor one at least one of the first graphics processing unit tile or the second graphics processing unit tile to track a geometry processing workload on the at least one of the first graphics processing unit tile or the second graphics processing unit tile; and
a large draw monitor comprising a processing unit to:
determine whether a vertex count for a graphics workload exceeds a threshold value; and
in response to a determination that the vertex count for the graphics workload exceeds the threshold value, to divide the graphics workload over a plurality of graphics processing units instantiated on the at least one of the first graphics processing unit tile or second graphics processing unit tile based at least in part on the geometry processing workload.

2. The general purpose graphics processing of claim 1 , the processing unit to:
set the threshold value to an initial threshold value based upon a graphics processing capability of the graphics processing compute block.

3. The general purpose graphics processing device of claim 2 , the processing unit to:
determine a current processing capacity of the graphics processing compute block; and
set the threshold value to an operating threshold value based upon the current processing capacity of the graphics processing compute block.

4. The general purpose graphics processing device of claim 2 , the processing unit to:
receive a workload indicator from a plurality of the graphics processing units in the graphics processing compute block; and
use the workload indicator to assign draw commands to graphics processing units.

5. The general purpose graphics processing device of claim 1 , the processing unit to:
determine whether a geometry type in the graphics workload has primitives which are fully specified; and
in response to a determination that the a geometry type in the graphics workload has primitives which are fully specified, to:
determine a number of graphics processing units to which the graphics workload is to be assigned;
generate a number of draw commands equal to the number of graphics processing units to which the workload is to be assigned; and
submit a draw command to each of the graphics processing units to which the workload is to be assigned.

6. The general purpose graphics processing device of claim 1 , the processing unit to:
determine whether a geometry type in the graphics workload has primitives which are fully specified; and
in response to a determination that the a geometry type in the graphics workload has primitives which are not fully specified, to:
insert one or more indices for vertexes in the graphics workload;
determine a number of graphics processing units to which the graphics workload is to be assigned;
generate a number of draw commands equal to the number of graphics processing units to which the workload is to be assigned; and
submit a draw command to each of the graphics processing units to which the workload is to be assigned.

7. A heterogeneous data processing system comprising:
a general purpose processor;
a general purpose graphics processing device comprising a general purpose graphics processing compute block to process a workload including graphics or compute operations, the general purpose graphics processing compute block comprising at least a first graphics processing unit tile and a second graphics processing unit tile;
a geometry utilization monitor one at least one of the first graphics processing unit tile or the second graphics processing unit tile to track a geometry processing workload on the at least one of the first graphics processing unit tile or the second graphics processing unit tile; and
a memory; and
a large draw monitor comprising a processing unit to:
determine whether a vertex count for a graphics workload exceeds a threshold value; and
in response to a determination that the vertex count for the graphics workload exceeds the threshold value, to divide the graphics workload over a plurality of graphics processing units instantiated on the at least one of the first graphics processing unit tile or second graphics processing unit tile based at least in part on the geometry processing workload.

8. The heterogeneous data processing device of claim 7 , the processing unit to:
set the threshold value to an initial threshold value based upon a graphics processing capability of the graphics processing compute block.

9. The heterogeneous data processing device of claim 8 , the processing unit to:
determine a current processing capacity of the graphics processing compute block; and
set the threshold value to an operating threshold value based upon the current processing capacity of the graphics processing compute block.

10. The heterogeneous data processing device of claim 8 , the processing unit to:
receive a workload indicator from a plurality of the graphics processing units in the graphics processing compute block; and
use the workload indicator to assign draw commands to graphics processing units.

11. The heterogeneous data processing device of claim 7 , the processing unit to:
determine whether a geometry type in the graphics workload has primitives which are fully specified; and
in response to a determination that the a geometry type in the graphics workload has primitives which are fully specified, to:
determine a number of graphics processing units to which the graphics workload is to be assigned;
generate a number of draw commands equal to the number of graphics processing units to which the workload is to be assigned; and
submit a draw command to each of the graphics processing units to which the workload is to be assigned.

12. The heterogeneous data processing device of claim 7 , the processing unit to:
determine whether a geometry type in the graphics workload has primitives which are fully specified; and
in response to a determination that the a geometry type in the graphics workload has primitives which are not fully specified, to:
inserting one or more indices for vertexes in the graphics workload;
determine a number of graphics processing units to which the graphics workload is to be assigned;
generate a number of draw commands equal to the number of graphics processing units to which the workload is to be assigned; and
submit a draw command to each of the graphics processing units to which the workload is to be assigned.

13. A method, comprising:
receiving a graphics workload including graphics or compute operations on a general purpose graphics processing device comprising:
a general purpose graphics processing compute block to process a workload including graphics or compute operations, the general purpose graphics processing compute block comprising at least a first graphics processing unit tile and a second graphics processing unit tile; and
a geometry utilization monitor one at least one of the first graphics processing unit tile or the second graphics processing unit tile to track a geometry processing workload on the at least one of the first graphics processing unit tile or the second graphics processing unit tile; and

in a large draw monitor comprising a processing unit:
determining whether a vertex count for a graphics workload exceeds a threshold value; and
in response to a d

Assignees

• Intel Corp

Inventors

• Schluessler Travis
• Waters Zack
• Apodaca Michael
• Surprise Jason
• Doyle Peter

Classifications

• G06T1/20 Primary

  Processor architectures; Processor configuration, e.g. pipelining · CPC title

• G06F9/505

  considering the load · CPC title

• G06F9/5044 Primary

  considering hardware capabilities · CPC title