# US 2017/0286198 A9 · Fault buffer for tracking page faults in unified virtual memory system

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

- Number: US 2017/0286198 A9
- Document ID: US20170286198A9
- Status: Published application
- Kind: A9
- Application number: 14/055,345
- Filed: 2013-10-16
- Published: 2017-10-05
- Earliest claimed: 2013-03-14
- Assignees: NVIDIA Corporation
- Applicants: NVIDIA Corporation
- Inventors: Jerome F. DULUK, JR.; Cameron BUSCHARDT; Sherry CHEUNG; James Leroy DEMING; Samuel H. DUNCAN; Lucien DUNNING; Robert GEORGE; Arvind GOPALAKRISHNAN; Mark HAIRGROVE; Chenghuan JIA; John MASHEY
- Web page: https://patentgrep.com/patent/US20170286198A9

## CPC

- G06F 12/1009: using page tables, e.g. page table structures
- G06F 11/073: {in a memory management context, e.g. virtual memory or cache management (memory management G06F12/00; testing of static memory units G11C29/00)}
- G06F 11/0793: {Remedial or corrective actions (recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40)}
- G06F 12/08: in hierarchically structured memory systems, e.g. virtual memory systems
- G06F 12/1072: Decentralised address translation, e.g. in distributed shared memory systems
- G06F 12/109: for multiple virtual address spaces, e.g. segmentation (G06F12/1036 takes precedence)
- G06F 12/12: Replacement control
- G06F 12/10: Address translation
- G06F 2212/1016: Performance improvement
- G06F 2212/2542: Non-uniform memory access [NUMA] architecture
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system for managing virtual memory. The system includes a first processing unit configured to execute a first operation that references a first virtual memory address. The system also includes a first memory management unit (MMU) associated with the first processing unit and configured to generate a first page fault upon determining that a first page table that is stored in a first memory unit associated with the first processing unit does not include a mapping corresponding to the first virtual memory address. The system further includes a first copy engine associated with the first processing unit. The first copy engine is configured to read a first command queue to determine a first mapping that corresponds to the first virtual memory address and is included in a first page state directory. The first copy engine is also configured to update the first page table to include the first mapping.

## Claims

### Claim 1

A system for tracking page faults, the system comprising:
a first processing entity configured to:
execute a first instruction that is associated with a first virtual memory address,
determine that a first page table associated with the first processing entity does not include a first mapping associated with the first virtual memory address, and
transmit a first fault buffer entry to a fault buffer in response to determining that that first page table does not include the first mapping,
wherein the fault buffer is configured to store a plurality of fault buffer entries received from at least the first processing entity.

### Claim 2 (depends on 1)

The system of claim 1, wherein the first processing entity includes a memory management unit that is configured to determine that the first page table does not include the first mapping and, in response, to transmit the first fault buffer entry to the fault buffer.

### Claim 3 (depends on 2)

The system of claim 2, wherein:
the first processing entity further includes a first streaming multiprocessor and a second streaming multiprocessor;
the first streaming multiprocessor is configured to execute the first instruction;
the second streaming multiprocessor is configured to execute a second instruction that is associated with a second virtual memory address; and
the memory management unit is configured to determine that the first page table does not include a second mapping associated with the second virtual memory address, and, in response, to transmit a second page fault to the fault buffer.

### Claim 4 (depends on 3)

The system of claim 3, wherein:
the first streaming multiprocessor is configured to stall in response to the memory management unit determining that the first page table does not include the first mapping; and
the second streaming multiprocessor is configured to stall in response to the memory management unit determining that the first page table does not include the second mapping.

### Claim 5 (depends on 4)

The system of claim 4, wherein the first streaming multiprocessor is further configured to resume execution and attempt to execute the first instruction again, and the second streaming multiprocessor is further configured to resume execution and attempt to execute the second instruction again.

### Claim 6 (depends on 1)

The system of claim 1, further comprising a second processing entity configured to execute a first page fault sequence to resolve a page fault associated with the first fault buffer entry.

### Claim 7 (depends on 6)

The system of claim 6, further comprising a command queue, wherein executing the first page fault sequence includes transmitting one or more commands to the command queue, the one or more commands being configured to cause the first processing entity to execute a first set of operations associated with the first page fault sequence.

### Claim 8 (depends on 7)

The system of claim 7, wherein the first processing entity further includes a copy engine configured to execute the first set of operations.

### Claim 9 (depends on 8)

The system of claim 8, wherein the copy engine is configured to execute a first operation included in the first set of operations to modify the first page table to include the first mapping.

### Claim 10

A computing device for tracking page faults, the parallel processing unit comprising:
a first processing entity configured to:
execute a first instruction that is associated with a first virtual memory address,
determine that a first page table associated with the first processing entity does not include a first mapping associated with the first virtual memory address, and
transmit a first fault buffer entry to a fault buffer in response to determining that that first page table does not include the first mapping,
wherein the fault buffer is configured to store a plurality of fault buffer entries received from at least the first processing entity.

### Claim 11 (depends on 10)

The computing device of claim 10, wherein the first processing entity includes a memory management unit that is configured to determine that the first page table does not include the first mapping and, in response, to transmit the first fault buffer entry to the fault buffer.

### Claim 12 (depends on 11)

The computing device of claim 11, wherein:
the first processing entity further includes a first streaming multiprocessor and a second streaming multiprocessor;
the first streaming multiprocessor is configured to execute the first instruction;
the second streaming multiprocessor is configured to execute a second instruction that is associated with a second virtual memory address; and
the memory management unit is configured to determine that the first page table does not include a second mapping associated with the second virtual memory address, and, in response, to transmit a second page fault to the fault buffer.

### Claim 13 (depends on 12)

The computing device of claim 12, wherein:
the first streaming multiprocessor is configured to stall in response to the memory management unit determining that the first page table does not include the first mapping; and
the second streaming multiprocessor is configured to stall in response to the memory management unit determining that the first page table does not include the second mapping.

### Claim 14 (depends on 13)

The computing device of claim 13, wherein the first streaming multiprocessor is further configured to resume execution and attempt to execute the first instruction again, and the second streaming multiprocessor is further configured to resume execution and attempt to execute the second instruction again.

### Claim 15 (depends on 10)

The computing device of claim 10, further comprising a command queue configured to store one or more commands associated with a page fault sequence.

### Claim 16 (depends on 15)

The computing device of claim 15, wherein the first processing entity further includes a copy engine configured to execute the one or more commands.

### Claim 17

A method for tracking page faults, the method comprising:
executing a first instruction that is associated with a first virtual memory address;
determining that a first page table associated with a first processing entity does not include a first mapping associated with the first virtual memory address; and
transmitting a first fault buffer entry to a fault buffer in response to determining that that first page table does not include the first mapping,
wherein the fault buffer is configured to store a plurality of fault buffer entries received from at least the first processing entity.

### Claim 18 (depends on 17)

The method of claim 17, further comprising executing a page fault sequence to resolve a page fault associated with the first fault buffer entry.

### Claim 19 (depends on 18)

The method of claim 18, wherein executing the first page fault sequence includes transmitting one or more commands to a command queue, the one or more commands being configured to cause the first processing entity to execute a first set of operations associated with the first page fault sequence.

### Claim 20 (depends on 19)

The method of claim 19, further comprising executing a first operation included in the first set of operations to modify the first page table to include the first mapping.
