Method and apparatus for memory allocation in a multi-node system
View Patent ↗According to at least one example embodiment, a multi-chip system includes multiple chip devices configured to communicate to each other and share resources. According to at least one example embodiment, a method of memory allocation in the multi-chip system comprises managing, by each of one or more free-pool allocator (FPA) coprocessors in the multi-chip system, a corresponding list of pools of free-buffer pointers. Based on the one or more lists of free-buffer pointers managed by the one or more FPA coprocessors, a memory allocator (MA) hardware component allocates a free buffer, associated with a chip device of the multiple chip devices, to data associated with a work item. According to at least one aspect, the data associated with the work item represents a data packet.
1. A multi-chip system comprising:
multiple chip devices, a first chip device of the multiple chip devices includes a memory allocator (MA) hardware component; and
one or more free-pool allocator (FPA) coprocessors, each associated with a corresponding chip device, and each configured to manage a corresponding list of pools of free-buffer pointers,
the MA hardware component configured to:
allocate a free buffer, associated with a chip device of the multiple chip devices, to data associated with a work item based on the one or more lists of free-buffer pointers managed by the one or more FPA coprocessors.
2. The multi-chip system as recited in claim 1 , wherein the MA hardware component is further configured to maintain a list including all pools of free-buffer pointers managed by the one or more FPA coprocessors.
3. The multi-chip system as recited in claim 1 , wherein a single FPA coprocessor is configured to manage a single list including all pools of free-buffer pointers associated with buffers in the multiple chip devices.
4. The multi-chip system as recited in claim 1 , wherein each chip device includes a corresponding FPA coprocessor.
5. The multi-chip system as recited in claim 4 , wherein for each FPA coprocessor, the free-buffer pointers, in the corresponding list of pools, point to buffers residing in the same chip device as the FPA coprocessor.
6. The multi-chip system as recited in claim 1 , wherein in allocating a free buffer to the data associated with the work item, the MA hardware component is further configured to:
send a request for a free-buffer pointer to a FPA coprocessor of the one or more FPA coprocessors;
receive a response including the free-buffer pointer; and
cause the data associated with the work item to be stored in a buffer pointed to by the free-buffer pointer received.
7. The multi-chip system as recited in claim 6 , wherein the free buffer allocated is freed upon the free-buffer pointer returned to the FPA coprocessor.
8. The multi-chip system as recited in claim 6 , wherein the FPA coprocessor of the one or more coprocessors resides in a second chip device, of the multiple chip devices, distinct from the first chip device.
9. The multi-chip system as recited in claim 8 , wherein the request for the free-buffer pointer and the response including the free-buffer pointer are exchanged between the first and second chip devices through communication channels designated for cross-chip communications between coprocessors.
10. The multi-chip system as recited in claim 1 , wherein the work item is assigned to a second chip device, of the multiple chip device, for processing.
11. The multi-chip system as recited in claim 10 , wherein the MA hardware component is further configured to allocate a free buffer, residing in the second chip device, to the data associated with the work item.
12. The multi-chip system as recited in claim 1 , wherein the data associated with the work item represents a data packet.
13. The multi-chip system as recited in claim 1 , wherein the MA hardware component includes a core processor or a coprocessor.
14. A method of memory allocation in a multi-chip system including multiple chip devices, the method comprising:
managing, by each of one or more free-pool allocator (FPA) coprocessors in the multi-chip system, a corresponding list of pools of free-buffer pointers; and
allocating, by a memory allocator (MA) hardware component, a free buffer, associated with a chip device of the multiple chip devices, to data associated with the work item based on the one or more lists of free-buffer pointers managed by the one or more FPA coprocessors.
15. The method as recited in claim 14 further comprising maintaining, by the MA hardware component, a list including all pools of free-buffer pointers managed by the one or more FPA coprocessors.
16. The method as recited in claim 14 , wherein a single FPA coprocessor, in the multi-chip system, is managing a single list including all pools of free-buffer pointers associated with buffers in the multiple chip devices.
17. The method as recited in claim 14 , wherein each chip device includes a corresponding FPA coprocessor.
18. The method as recited in claim 17 , wherein for each FPA coprocessor, the free-buffer pointers, in the corresponding list of pools, point to buffers residing in the same chip device as the FPA coprocessor.
19. The method as recited in claim 14 , wherein allocating a free buffer to the data associated with the work item includes:
sending a request, by the MA hardware component, for a free-buffer pointer to a FPA coprocessor of the one or more FPA coprocessors;
receiving, by the MA hardware component, a response including the free-buffer pointer; and
causing, by the MA hardware component, the data associated with the work item to be stored in a buffer pointed to by the free-buffer pointer received.
20. The method as recited in claim 19 , wherein the free buffer allocated is freed upon the free-buffer pointer returned to the FPA coprocessor.
21. The method as recited in claim 19 , wherein the FPA coprocessor of the one or more coprocessors resides in a second chip device, of the multiple chip devices, distinct from the first chip device.
22. The method system as recited in claim 21 , wherein the request for the free-buffer pointer and the response including the free-buffer pointer are exchanged between the first and second chip devices through communication channels designated for cross-chip communications between coprocessors.
23. The method as recited in claim 14 , wherein the work item is assigned to a second chip device, of the multiple chip device, for processing.
24. The method as recited in claim 23 , wherein allocating a free buffer includes allocating a free buffer, residing in the second chip device, to the data associated with the work item.
25. The method as recited in claim 14 , wherein the data associated with the work item represents a data packet.
26. The method as recited in claim 14 , wherein the MA hardware component includes a core processor or a coprocessor.