# US 2026/0219849 A1 · Control of storage aliasing via automatic application of artificial dependences during program compilation

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

- Number: US 2026/0219849 A1
- Document ID: US20260219849A1
- Status: Published application
- Kind: A1
- Application number: 19/446,257
- Filed: 2026-01-12
- Published: 2026-07-30
- Earliest claimed: 2022-04-15
- Assignees: NVIDIA Corporation
- Applicants: NVIDIA Corporation
- Inventors: Malay SANGHI; Duane MERRILL
- Web page: https://patentgrep.com/patent/US20260219849A1

## CPC

- G06F 8/41: Compilation
- G06F 8/441: {Register allocation; Assignment of physical memory space to logical memory space}
- G06F 8/445: {Exploiting fine grain parallelism, i.e. parallelism at instruction level (run-time instruction scheduling G06F9/3836)}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

In various examples, systems and methods are disclosed relating to aliasing control of program variables in storage via automatic application of artificial dependences during program compilation. In some implementations, a system can include a detector to automatically detect a pattern, based at least on a structure of data flow in a source program, indicative of sequences of dependent operations, where the sequences are independent from one another. The system can determine a storage aliasing preference for whether to allow the compiler to allocate the program variables of the respective sequences to the same processor storage locations, or to prevent the compiler from doing so. The system can assign one or more annotations to the source program indicative of one or more artificial dependences for a compiler to respect when performing program transformations prior to the allocation of program variables.

## Claims

### Claim 1

A system comprising:
one more processors to:
determine, based at least on a data flow of a plurality of operations of a source program, an artificial dependence among at least two operations of the source program, the at least two operations each comprising a load operation, a compute operation, or a store operation;
apply, to the source program, an annotation indicating the artificial dependence; and
allocate, using a compiler and according to the annotation, each operation of the at least two operations to at least one of a first storage location of a parallel processing system or a second storage location of the parallel processing system.

### Claim 2 (depends on 1)

The system of claim 1, wherein the one or more processors are to:
determine a modified order of the at least two operations according to the artificial dependence; and
allocate each operation the at least one of the first storage location or the second storage location according to the modified order.

### Claim 3 (depends on 1)

The system of claim 1, wherein the artificial dependence comprises a compiler-readable label that the one or more processors are to insert into the source program.

### Claim 4 (depends on 1)

The system of claim 1, wherein the plurality of operations comprises one or more matrix-matrix multiplication operations.

### Claim 5 (depends on 1)

The system of claim 1, wherein the plurality of operations comprises one or more operations to configure a neural network.

### Claim 6 (depends on 1)

The system of claim 1, wherein the annotation is representative of a pseudo-dependence between the at least two operations to prevent the compiler from causing aliasing of the at least two operations.

### Claim 7 (depends on 1)

The system of claim 1, wherein the parallel processing system comprises a graphics processing unit (GPU).

### Claim 8 (depends on 1)

The system of claim 1, wherein the one or more processors are to assign the annotation to a position in a representation of the source program, the position corresponding to at least one of a register footprint, a register pressure, an instruction level parallelism, or an occupancy represented by an order of the plurality of operations in the source program.

### Claim 9 (depends on 1)

The system of claim 1, wherein the one or more processors are to assign the annotation to the source program based at least on an expected execution time of the plurality of operations.

### Claim 10 (depends on 1)

The system of claim 1, wherein the one or more processors are to assign the annotation to the source program to modify an amount of time between execution of a first operation of the at least two operations and execution of a second operation between the at least two operations.

### Claim 11 (depends on 1)

The system of claim 1, wherein the one or more processors are to detect the data flow based at least on a true dependence, a flow dependence, or an order dependence of the plurality of operations.

### Claim 12 (depends on 1)

The system of claim 1, wherein the compiler is to generate machine code by modifying an order of the plurality of operations as indicated in the source program.

### Claim 13 (depends on 1)

The system of claim 1, wherein the system is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing digital twin operations;
a system for performing light transport simulation;
a system for performing generative artificial intelligence (AI) operations;
a system for performing collaborative content creation for three-dimensional (3D) assets;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a system implemented using a language model;
a system for performing conversational AI operations;
a system for generating synthetic data;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.

### Claim 14

A processor comprising:
one more circuits to:
detect, based at least on a data flow between a first operation of a source program and a second operation of the source program, an artificial dependence among the first operation and the second operation, the first operation and the second operation each comprising a load operation, a compute operation, or a store operation;
apply, to the source program, an annotation indicating the artificial dependence; and
allocate, using a compiler and according to the annotation, the first operation and the second operation to at least one of a first storage location of a parallel processing system or a second storage location of the parallel processing system.

### Claim 15 (depends on 14)

The processor of claim 14, wherein the annotation is representative of a pseudo-dependence to prevent the compiler from aliasing the first operation and the second operation.

### Claim 16 (depends on 14)

The processor of claim 14, wherein the one or more circuits are to assign the annotation to a position in a representation of the source program, the position corresponding to at least one of a register footprint, a register pressure, an instruction level parallelism, or an occupancy represented by an order of the first operation and the second operation in the source program.

### Claim 17 (depends on 14)

The processor of claim 14, wherein the one or more circuits are to assign the annotation to the source program to modify an amount of time between execution of the first operation and execution of the second operation.

### Claim 18 (depends on 14)

The processor of claim 14, wherein the processor is comprised in at least one of:
a system for performing simulation operations;
a system for performing digital twin operations;
a system for performing generative AI operations;
a system for performing light transport simulation;
a system for performing collaborative content creation for 3D assets;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system implemented using a language model;
a system for performing conversational AI operations;
a system for generating synthetic data;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources.

### Claim 19

A method, comprising:
determining, based at least on a data flow of a plurality of operations of a source program, an artificial dependence among at least two operations of the source program, the at least two operations each comprising a load operation, a compute operation, or a store operation;
applying, to the source program, an annotation indicating the artificial dependence; and
allocating, using a compiler and according to the annotation, each operation of the at least two operations to at least one of a first storage location of a parallel processing system or a second storage location of the parallel processing system.

### Claim 20 (depends on 19)

The method of claim 19, comprising assigning the annotation to the source program based at least on an expected execution time of the plurality of operations.
