# US 2018/0025465 A1 · Gpu accelerated address translation for graphics virtualization

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

- Number: US 2018/0025465 A1
- Document ID: US20180025465A1
- Status: Published application
- Kind: A1
- Application number: 15/615,809
- Filed: 2017-06-06
- Published: 2018-01-25
- Earliest claimed: 2014-08-21
- Assignees: Intel Corporation
- Applicants: Intel Corporation
- Inventors: YUNBIAO LIN; JIANGHONG DU
- Web page: https://patentgrep.com/patent/US20180025465A1

## CPC

- G06T 1/60: Memory management
- G06F 9/45533: {Hypervisors; Virtual machine monitors}
- G06F 12/1027: using associative or pseudo-associative address translation means, e.g. translation look-aside buffer [TLB]
- G09G 5/363: {Graphics controllers}
- G09G 5/39: Control of the bit-mapped memory
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G09G: ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION (arrangements for transferring data between digital computers and displays G06F3/14; static indicating arrangements comprising an association of a number of separate sources or light control cells G09F9/00; static indicating arrangements comprising integral associations of a number of light sources H01J, H01K, H10H, H05B33/12; scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission, details thereof H04N1/00)

## Abstract

In accordance with embodiments disclosed herein, there are provided methods, systems, mechanisms, techniques, and apparatuses for implementing GPU (Graphics Processing Unit) accelerated address translation for graphics virtualization. In one embodiment, such a system includes a main memory having a plurality of machine physical addresses; a graphics processor unit having graphics memory therein; an address translation service integrated with the graphics processor unit; a hypervisor to manage one or more guest machines; wherein the hypervisor is to configure a lookup table within the graphics memory of the graphics processor unit; and further wherein the address translation service of the graphics processor unit is to translate a guest physical address for one of the one or more guest machines to a corresponding machine physical address within the main memory. Such a graphics processor unit may be implemented separate from a system, for example, embodied within a silicon integrated circuit.

## Claims

### Claim 1

A system comprising:
a main memory having a plurality of machine physical addresses;
a graphics processor unit having graphics memory therein;
an address translation service integrated with the graphics processor unit;
a hypervisor to manage one or more guest machines;
wherein the hypervisor to configure a lookup table within the graphics memory of the graphics processor unit; and
wherein the address translation service of the graphics processor unit to translate a guest physical address for one of the one or more guest machines to a corresponding machine physical address within the main memory.

### Claim 2 (depends on 1)

The system of claim 1, further comprising:
one or more graphics virtualization registers;
wherein the one or more graphics virtualization registers inform the graphics processor unit that one of the guest machines have been assigned to the graphics processor unit by the hypervisor.

### Claim 3 (depends on 2)

The system of claim 2, wherein the one or more graphics virtualization registers further identify where the lookup table is located for the assigned guest machine.

### Claim 4 (depends on 1)

The system of claim 1, further comprising:
a graphics virtualization determination component integrated within the graphics processor unit;
wherein the graphics virtualization determination component is to determine whether or not the graphics processor unit is operating within a virtualization environment on behalf of one of the guest machines;
wherein the graphics processor unit to engage the address translation service when operating within the virtualization environment; and
wherein the system to implement logical memory address to guest physical address translation without engaging the address translation service of the graphics processor unit when the graphics processor unit is not operating within the virtualization environment.

### Claim 5 (depends on 1)

The system of claim 1, wherein the hypervisor to engage the address translation service of the graphics processor unit by:
configuring one or more graphics virtualization registers to inform the graphics processor unit that one of the guest machines have been assigned to the graphics processor unit by the hypervisor; and
passing a Guest physical Frame Number (GFN) to the graphics processor unit for translation by the address translation service to a Machine Frame Number (MFN), wherein the GFN represents the guest physical address for one of the one or more guest machines and wherein the MFN represents the machine physical addresses within the main memory corresponding to the guest physical address.

### Claim 6 (depends on 1)

The system of claim 1, wherein:
an address translation service integrated with the graphics processor unit retrieves Guest physical Frame Number (GFN) to Machine Frame Number (MFN) mappings from the lookup table in the graphics memory.

### Claim 7 (depends on 6)

The system of claim 6, wherein:
the address translation service retrieves the GFN to MFN mappings through a Translation Lookaside Buffer (TLB) integrated within the graphics processor unit; and
wherein the TLB provides an indexable cache on behalf of the address translation service of the graphics processor unit.

### Claim 8 (depends on 1)

The system of claim 1:
wherein the hypervisor comprises a virtual machine manager; and
wherein the one or more guest machines comprise one or more virtual machines.

### Claim 9 (depends on 1)

The system of claim 1, further comprising:
a separate and distinct central processor unit communicably interfaced with the graphics processor unit via a system bus;
wherein the central processing unit lacks dedicated hardware circuitry to perform address translation of guest physical addresses to machine physical addresses.

### Claim 10 (depends on 1)

The system of claim 1, wherein the graphics processor unit comprises a silicon integrated circuit type graphics processor unit.

### Claim 11 (depends on 1)

The system of claim 1, wherein the graphics processor unit comprises a microprocessor within a tablet computing device or a smart phone or one of a plurality of microprocessors integrated within the tablet computing device or the smartphone.

### Claim 12 (depends on 11)

The system of claim 11, wherein the tablet computing device or the smartphone comprises a separate and distinct central processing unit communicatively interfaced with the graphics processor unit within the tablet computing device or the smartphone; and
wherein the central processing unit lacks dedicated hardware circuitry to perform address translation of guest physical addresses to machine physical addresses.

### Claim 13 (depends on 1)

The system of claim 1:
wherein the graphics memory comprises shared graphics memory; and
wherein the hypervisor to configure the lookup table within the graphics memory of the graphics processor unit comprises the hypervisor to write the lookup table into the shared graphics memory of the graphics processor unit.

### Claim 14 (depends on 1)

The system of claim 1:
wherein the hypervisor to configure the lookup table within the graphics memory of the graphics processor unit comprises the hypervisor to instruct the graphics processor unit to retrieve and store the lookup table;
wherein the graphics processor unit to issue a Direct Memory Access (DMA) request to fetch the lookup table; and
wherein the graphics processor unit to store the lookup table in the graphics memory.

### Claim 15 (depends on 1)

The system of claim 1:
wherein the hypervisor to issue a lookup table update to the graphics processor unit; and
wherein the graphics processor unit to responsively update the lookup table in the graphics memory.

### Claim 16

A method comprising:
managing, via a hypervisor, one or more guest machines on a system;
configuring a lookup table with a mapping of guest physical addresses for the one or more guest machines to corresponding machine physical addresses;
transferring the lookup table to a graphics memory of a graphics processor unit;
receiving an access request to a main memory of the system from one of the guest machines, wherein the access request specifies a guest physical address;
engaging an address translation service internal to the graphics processor unit by passing the guest physical address to the graphics processor unit; and
translating, via the address translation service of the graphics processor unit, the guest physical address to a corresponding machine physical address.

### Claim 17 (depends on 16)

The method of claim 16, wherein the graphics processor unit comprises one or more graphics virtualization registers;
wherein the hypervisor configures the one or more graphics virtualization registers to inform the graphics processor unit that one of the guest machines have been assigned to the graphics processor unit by the hypervisor; and
wherein the one or more graphics virtualization registers further identify where the lookup table is located for the assigned guest machine.

### Claim 18 (depends on 17)

The method of claim 17, further comprising:
passing a Guest physical Frame Number (GFN) from the hypervisor to the graphics processor unit requesting translation of the GFN by the address translation service to a Machine Frame Number (MFN), wherein the GFN represents the guest physical address for one of the one or more guest machines and wherein the MFN represents the machine physical addresses within the main memory corresponding to the guest physical address specified by the access request.

### Claim 19 (depends on 16)

The method of claim 16, wherein translating the guest physical address to a corresponding machine physical address comprises the address translation service of the graphics processor unit performing the translating based on a Guest physical Frame Number (GFN) to Machine Frame Number (MFN) mapping from the lookup table in the graphics memory.

### Claim 20 (depends on 19)

The method of claim 19, wherein:
the address translation service retrieves the GFN to MFN mapping through a Translation Lookaside Buffer (TLB) integrated within the graphics processor unit; and
wherein the TLB provides an indexable cache on behalf of the address translation service of the graphics processor unit.
