Granted Patent B2

US 12,293,227 · App. 17/254,195 · Granted May 6, 2025

Inventors: Amir Roozbeh (Stockholm, SE); Alireza Farshin (Stockholm, SE); Dejan Kostic (Solna, SE); Gerald Q. Maguire, Jr. (Stockholm, SE)

Assignee: TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

G06F9/5016G06F12/0811G06F2212/608

Abstract

Claims (29)

1. A method performed by a memory allocator in a computer system for allocating memory to applications, wherein the computer system comprises at least one computer server comprising a plurality of central processing unit (CPU) cores, the plurality of CPU cores comprising a first CPU core and a second CPU core, and the computer server further comprising at least a first memory unit and a second memory unit having different data access times, and wherein the second memory unit is divided into a plurality of slices, wherein the plurality of slices of the second memory unit comprises a first slice and a second slice, the first slice is associated with the first CPU core, and the second slice is associated with the second CPU core, the method comprising:

receiving a request from an application for allocating memory to the application, wherein the application is running on the first CPU core;

in response to the request, selecting a memory block from the first memory unit, wherein the memory block comprises a set of memory chunks;

determining whether at least T number of the memory chunks included in the selected memory block of the first memory unit are associated with the first slice of the second memory unit, where T is a predetermined threshold;

in response to determining that at least T number of the memory chunks included in the selected memory block of the first memory unit are associated with the first slice of the second memory unit, allocating the selected memory block to the requesting application.

2. The method of claim 1 , further comprising monitoring a selection algorithm that controls which memory chunks in the first memory unit are loaded into a slice of the second memory unit; and

determining which memory chunks of the first memory unit are associated with the first slice of the second memory unit.

3. The method of claim 2 , wherein the selection algorithm uses a hash function.

4. The method of claim 2 , wherein the monitoring of the selection algorithm is done by a memory allocation enhancer unit.

5. The method of claim 1 , further comprising, for each CPU core in the plurality of CPU cores, determining a data access time between the CPU core and each slice of the second memory unit.

6. The method of claim 5 , further comprising associating each CPU core with at least one slice of the second memory unit based on the determined data access times.

7. The method of claim 1 , wherein

the first memory unit is a dynamic random-access memory (DRAM), and

the second memory unit is a last level cache (LLC).

8. A memory allocator in a computer system for allocating memory to applications, wherein the computer system comprises at least one computer server comprising a plurality of central processing unit (CPU) cores, the plurality of CPU cores comprising a first CPU core and a second CPU core, and the computer server further comprises at least a first memory unit and a second memory unit having different data access times, and wherein the second memory unit is divided into a plurality of slices, wherein the plurality of slices of the second memory unit comprises a first slice and a second slice, the first slice is associated with the first CPU core, and the second slice is associated with the second CPU core, further wherein the computer system is further comprising a processor and a memory containing instructions which, when executed by the processor instructs the memory allocator to perform a method comprising:

receiving a request from an application for allocating memory to the application, wherein the application is running on the first CPU core;

in response to the request, selecting a memory block from the first memory unit, wherein the memory block comprises a set of memory chunks;

determining whether at least T number of the memory chunks included in the selected memory block of the first memory unit are associated with the first slice of the second memory unit, where T is a predetermined threshold;

in response to determining that at least T number of the memory chunks included in the selected memory block of the first memory unit are associated with the first slice of the second memory unit, allocating the selected memory block to the requesting application.

9. A Key Value Storage data base comprising the memory allocator of claim 8 .

10. The memory allocator of claim 8 , further comprising monitoring a selection algorithm that controls which memory chunks in the first memory unit are loaded into a slice of the second memory unit; and

determining which memory chunks of the first memory unit are associated with the first slice of the second memory unit.

11. The memory allocator of claim 10 , wherein the selection algorithm uses a hash function.

12. The memory allocator of claim 10 , wherein the monitoring of the selection algorithm is done by a memory allocation enhancer unit.

13. The memory allocator of claim 8 , further comprising, for each CPU core in the plurality of CPU cores, determining a data access time between the CPU core and each slice of the second memory unit.

14. The memory allocator of claim 13 , further comprising associating each CPU core with at least one slice of the second memory unit based on the determined data access times.

15. The memory allocator of claim 8 , wherein

the first memory unit is a dynamic random-access memory (DRAM), and

the second memory unit is a last level cache (LLC).

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (12)