# US 2021/0334450 A1 · Systems and methods for implementing tile-level predication within a machine perception and dense algorithm integrated circuit

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

- Number: US 2021/0334450 A1
- Document ID: US20210334450A1
- Status: Published application
- Kind: A1
- Application number: 17/365,555
- Filed: 2021-07-01
- Published: 2021-10-28
- Earliest claimed: 2020-01-06
- Assignees: quadric.io, Inc.
- Applicants: quadric.io, Inc.
- Inventors: Nigel Drego; Mrinalini Ravichandran; Aman Sikka; Daniel Firu; Veerbhan Kheterpal
- Web page: https://patentgrep.com/patent/US20210334450A1

## CPC

- G06F 30/398: Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM] (optical proximity correction [OPC] design processes G03F1/36)
- G06F 15/8023: {Two dimensional arrays, e.g. mesh, torus}
- G06F 30/27: using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
- G06F 30/32: Circuit design at the digital level (reconfigurable circuits G06F30/34)
- G06F 30/337: Design optimisation
- G06F 30/373: Design optimisation
- G06F 15/8007: {single instruction multiple data [SIMD] multiprocessors}
- G06F 2115/10: Processors
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Systems and methods of implementing tile-level predication of a computing tile of an integrated circuit includes identifying a plurality of distinct predicate state values for each of a plurality of distinct processing cores of the computing tile; calculating one or more summed predicate state values for an entirety of the plurality of distinct processing cores based on performing a summation operation of the plurality of distinct predicate state values; propagating the one or more summed predicate state values to an instructions generating circuit of the integrated circuit; and identifying, by the instructions generating circuit, a tile-level predication for the computing tile based on input of the one or more summed predicate state values.

## Claims

### Claim 1

A method of implementing tile-level predication of a computing tile of an integrated circuit, the method comprising:
calculating a summed predicate state value for the plurality of distinct processing cores of the computing tile based on a plurality of distinct predicate state values associated with the plurality of distinct processing cores; and
identifying, by the instructions generating circuit, a tile-level predication for the computing tile based on an input of the summed predicate state value.

### Claim 2 (depends on 1)

The method according to claim 1, further comprising:
identifying the plurality of distinct predicate state values that includes:
identifying a top of stack value of a distinct predicate stack associated with each of the plurality of distinct processing cores of the computing tile based on an evaluation an associated predicate condition; and
identifying a bit value of a distinct enablement bit associated with each of the plurality of distinct processing cores of the computing tile.

### Claim 3 (depends on 2)

The method according to claim 2, wherein
the identifying the plurality of distinct predicate state values further includes:
identifying a loop depth of a nested loop body;
incrementing a loop depth counter based on the identified loop depth;
tallying a subset of processing cores of the plurality of distinct processing cores that are in an active state for executing the nested loop body; and
identifying predicate state values of the subset of processing cores separately from the plurality of distinct predicate state values a remainder of the plurality of distinct processing cores that are in an inactive state for executing the nest loop body.

### Claim 4 (depends on 1)

The method according to claim 1, further comprising:
executing a branch target instruction to branch around a target instruction that avoids a target set of instructions accessible to the plurality of distinct processing cores based on identifying the tile-level predication.

### Claim 5 (depends on 1)

The method according to claim 1, wherein
the plurality of distinct predicate state values includes:
(i) a plurality of distinct top of stack values for a plurality of distinct predicate stacks of the plurality of distinct processing cores of the computing tile; and
(ii) a plurality of distinct enablement bit values of the plurality of distinct processing cores of the computing tile.
