MANAGING AN OUT-OF-ORDER ASYNCHRONOUS HETEROGENEOUS REMOTE DIRECT MEMORY ACCESS (RDMA) MESSAGE QUEUE

• US 20100106948A1
• Filed: 10/24/2008
• Published: 04/29/2010
• Est. Priority Date: 10/24/2008
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Dale, Jason N., Bellows, Gregory Howard
• Granted Patent: US 8,904,064 B2
• Time in Patent Office: Days
• US Class Current: 712/225
• CPC Class Codes: G06F 13/28 using burst mode transfer, e.g. direct memory access DMA, cycle steal G06F13/32 takes precedence

Abstract

A system and method operable to manage a message queue is provided. This management may involve out-of-order asynchronous heterogeneous remote direct memory access (RDMA) to the message queue. This system includes a pair of processing devices, a primary processing device and an additional processing device, a memory in storage location and a data bus coupled to the processing devices. The processing devices cooperate to process queue data within a shared message queue wherein when an individual processing device successfully accesses queue data the queue data is locked for the exclusive use of the processing device. When the processing device acquires the queue data, the queue data is locked and the queue data acquired by the acquiring processing device includes the queue data for both the primary processing device and additional processing device such that the processing device has all queue data necessary to process the data and return processed queue data.

Claims

1. A system, comprising:

a primary processing device;

at least one additional processing device;

a memory storage location; and

a data bus coupled to the primary processing device, the at least one additional processing device, and the memory storage location, wherein;

the primary processing device and the at least one additional processing device cooperate to process queue data within a shared message queue; and

the primary processing device and the at least one additional processing device access the queue data within the shared message queue via remote direct memory access (RDMA).

2. The system of claim 1, wherein:

a primary memory domain is associated with the primary processing device, and the shared message queue resides within the primary memory domain;

at least one additional memory domain is associated with the at least one additional processing device;

a lock is associated with the queue data such that when a processing device, the primary processing device or the at least one additional processing device, locks queue data, the processing device acquires the queue data.

4. The system of claim 1, wherein the primary processing device comprises:

a power processing unit (PPU); and

a main memory location.

5. The system of claim 1, wherein the at least one additional processing device comprises:

a synergistic processing unit (SPU); and

a local memory location.

6. The system of claim 1, wherein a processing device, the primary processing device or the at least one additional processing device:

checks to see if queue data is locked;

performs an atomic GET to retrieve and lock the queue data when the queue data is not locked;

operate on the queue data to produce updated queue data; and

performs an atomic PUT to return the updated queue data.

7. The system of claim 1, wherein another processing device cannot operate on the queue data when the queue data is locked.

3. The system of claim 3, wherein the queue data acquired by the processing device when the queue data is locked comprises all data necessary to manage the queue.

16. The system of claim 3, wherein the queue data acquired by the processing device when the queue data is locked comprises primary processing device queue data and additional processing device queue data necessary to manage the queue.

8. A method comprising:

determining, by a cooperating processing device, if a shared queue data to be processed by the cooperating processing device is locked;

acquiring a shared queue lock when the shared queue data is not locked;

exclusively acquiring, by the cooperating processing device, the shared queue data during the lock;

processing the shared queue data, by the cooperating processing device, to produce updated queue data;

outputting the updated queue data and releasing the shared queue lock.

9. The method of claim 8, further comprising wherein the cooperating processing device is selected from:

a primary processing device; and

at least one additional processing device.

10. The method of claim 8, wherein:

acquiring a shared queue lock when the shared queue data is not locked; and

exclusively acquiring, by the cooperating processing device, the shared queue data during the lock, comprises an atomic GET; and

outputting the updated queue data and releasing the shared queue lock comprises an atomic PUT.

11. The method of claim 8, wherein acquiring a shared queue lock when the shared queue data is not locked comprises:

fetching the queue data from main memory;

setting a lock flag to locked; and

writing the queue data with a locked lock flag to main memory.

12. The method of claim 8, wherein the queue data comprises actual data to be processed.

13. The method of claim 8, wherein the queue data comprises an address and size associated with actual data to be processed.

14. A system comprising:

a primary processing device wherein a primary memory domain is associated with the primary processing device;

an additional processing device wherein an additional memory domain is associated with the additional processing device;

a memory storage location, wherein the primary memory domain and additional memory domain are mapped to locations within the memory storage;

a data bus coupled to the primary processing device, the at least one additional processing device, and the memory storage location, wherein;

the primary processing device and the additional processing device cooperate to process queue data within a shared message queue;

the primary processing device and the additional processing device access the queue data within the shared message queue via direct memory access (DMA);

15. The system of claim 14, wherein:

a lock is associated with the queue data such that when a processing device, the primary processing device or the at least one additional processing device, locks queue data, the processing device acquires the queue data.

17. The system of claim 14, wherein the primary processing device comprises:

a power processing unit (PPU); and

a main memory location.

18. The system of claim 14, wherein the at least one additional processing device comprises:

a synergistic processing unit (SPU); and

a local memory location.

19. The system of claim 14, wherein a processing device, the primary processing device or the at least one additional processing device:

checks to see if queue data is locked;

performs an atomic GET to retrieve and lock the queue data when the queue data is not locked;

operate on the queue data to produce updated queue data; and

performs an atomic PUT to return the updated queue data.

20. The system of claim 14, wherein another processing device cannot operate on the queue data when the queue data is locked.