IP Library › Granted Patent US 11,573,898
Granted Patent B2
US 11,573,898 · App. 16/995,411 · Granted Feb 7, 2023

System and method for facilitating hybrid hardware-managed and software-managed cache coherency for distributed computing

Inventors: Randy Passint (Chippewa Falls, WI); Paul Frank (Chippewa Falls, WI); Russell L. Nicol (Chippewa Falls, WI); Thomas McGee (Chippewa Falls, WI); Michael Woodacre (Winchester, GB)
Assignee: Hewlett Packard Enterprise Development LP
G06F12/0831G06F13/4027G06F2212/271G06F2212/62
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 11,573,898
App. No.
16/995,411
Granted
Feb 7, 2023
Kind
B2
Abstract

A node controller is provided to include a first interface to interface with one or more processors, a second interface including a plurality of ports to interface with node controllers within a base node and other nodes in the cache-coherent interconnect network. The node controller can further include a third interface to interface with a first plurality of memory devices and a cache coherence management logic. The cache coherence management logic can maintain, based on a first circuitry, hardware-managed cache coherency in the cache-coherent interconnect network. The cache coherence management logic can further facilitate, based on a second circuitry, software-managed cache coherency in the cache-coherent interconnect network.

Claims (54)

1. A node controller, comprising:

a first interface to interface with one or more processors;

a second interface to interface with other node controllers in a cache-coherent interconnect network;

a third interface to interface with a first plurality of memory devices;

a cache coherence management logic to:

maintain, based on a first circuitry, hardware-managed cache coherency in the cache-coherent interconnect network; and

facilitate, based on a second circuitry, software-managed cache coherency in the cache-coherent interconnect network;

wherein the cache coherence management logic is further to:

operate the first circuitry, in response to determining that a memory access request is destined to a second plurality of memory devices coupled to the one or more processors to which the node controller is coupled, wherein the second plurality of memory devices correspond to a hardware-managed cache-coherent space; and

operate the second circuitry, in response to determining that a memory access request is destined to the first plurality of memory devices coupled to the node controller, wherein the node controller is coupled to the one or more processors that are coupled to the second plurality of memory devices, and wherein the first plurality of memory devices correspond to a software-managed cache-coherent space.

2. The node controller of claim 1 , wherein the cache coherence management logic is further to:

operate the first circuitry, in response to determining that a memory access request is destined to a hardware-managed cache coherent space in the cache-coherent interconnect network; and

operate the second circuitry, in response to determining that a memory access request is destined to a software-managed cache coherent space in the cache-coherent interconnect network.

3. The node controller of claim 1 , wherein the first interface facilitates Ultra Path Interconnect (UPI) links for coupling with the one or more processors; and

wherein the second interface facilitates a plurality of Non-uniform Memory Access (NUMA) links for coupling with other node controllers in the cache-coherent interconnect network.

4. The node controller of claim 1 , wherein the third interface comprises one or more of:

a Gen-Z interface;

an InfiniBand interface; and

an Ethernet interface.

5. The node controller of claim 1 , wherein the first circuitry implements a directory-based hardware-managed cache coherence protocol.

6. The node controller of claim 1 , wherein the second circuitry facilitates a software-managed cache coherence protocol, and wherein the one or more processors implement the software-managed cache coherence protocol.

7. A system comprising:

a plurality of nodes in a cache-coherent interconnect network, wherein a respective node in a first subset of the nodes includes:

a group of processors;

a first plurality of memory groups with each memory group coupled to a processor; and

a first plurality of node controllers with each node controller coupled to one or more processors;

wherein a respective node in a second subset of the nodes includes:

a second plurality of memory groups; and

a second plurality of node controllers with each node controller coupled to a respective memory group in the second plurality of memory groups; and

wherein each node controller of the first plurality of node controllers and the second plurality of node controllers comprises a cache coherence management logic, which includes:

a first circuitry to maintain hardware-managed cache coherency in the cache-coherent interconnect network; and

a second circuitry to facilitate software-managed cache coherency in the cache-coherent interconnect network;

wherein the cache coherence management logic is further to:

operate the first circuitry, in response to determining that a memory access request is destined to the first plurality of memory groups, wherein first plurality of memory groups corresponds to a hardware-managed cache-coherent space; and

operate the second circuitry, in response to determining that a memory access request is destined to a third plurality of memory devices coupled to the node controller, wherein the node controller is coupled to the group of processors that are coupled to the first plurality of memory groups, and wherein the third plurality of memory devices corresponds to a software-managed cache-coherent space.

8. The system of claim 7 , wherein the cache coherence management logic is further to:

operate the first circuitry, in response to determining that a memory access request is destined to a hardware-managed cache coherent space in the cache-coherent interconnect network; and

operate the second circuitry, in response to determining that a memory access request is destined to a software-managed cache coherent space in the cache-coherent interconnect network.

9. The system of claim 7 , wherein the cache coherence management logic is further to:

operate the first circuitry, in response to determining that a memory access request is destined to the second plurality memory groups, wherein the second plurality memory groups corresponds to a hardware-managed cache-coherent space.

10. The system of claim 7 , wherein the respective node controller further includes:

a first interface to interface with the group of processors;

a second interface to interface with other node controllers in the cache-coherent interconnect network; and

a third interface to interface with the second plurality of memory groups.

11. The system of claim 10 , wherein the first interface facilitates Ultra Path Interconnect (UPI) links for coupling with the group of processors; and

wherein the second interface facilitates a plurality of Non-uniform Memory Access (NUMA) links for coupling with other node controllers in the cache-coherent interconnect network.

12. The system of claim 10 , wherein the third interface comprises one or more of:

a Gen-Z interface;

an InfiniBand interface; and

a Ethernet interface.

13. The system of claim 7 , wherein each memory group in the first plurality of memory groups corresponds to a group of Double Data Rate (DDR) Dual In-Line Memory Modules (DIMMs).

14. The system of claim 7 , wherein the first circuitry implements a directory-based hardware-managed cache coherence protocol;

wherein the second circuitry supports a software-managed cache coherence protocol; and

wherein the one or more processors implement the software-managed cache coherence protocol.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 17, 2020
From: PASSINT, RANDY; FRANK, PAUL; NICOL, RUSSELL L.; MCGEE, THOMAS; WOODACRE, MICHAEL
To: HEWLETT PACKARD ENTERPRISE DEVELOPMENT LP
Reel/Frame 053516/0484 →
Continuity (1)
Related Publication 20220050780A1 · Feb 17, 2022
Cited By (6)
US 12,259,816 US 12,326,813 US 12,360,937 US 12,386,751 US 12,475,050 US 12,596,648