Scalable low bandwidth multicast handling in mixed core systems

• US 7,302,691 B2
• Filed: 05/10/2002
• Issued: 11/27/2007
• Est. Priority Date: 05/10/2002
• Status: Active Grant

• Current Assignee: META Platforms Technologies LLC (Meta Platforms, Inc. (f/k/a Facebook, Inc.))
• Original Assignee: Sonics Incorporated (Meta Platforms, Inc. (f/k/a Facebook, Inc.))
• Inventors: Weber, Wolf-Dietrich, Chou, Chien-Chun, Masri, Nabil N., Wingard, Drew Eric
• Primary Examiner(s): WON, MICHAEL YOUNG
• Application Number:
  US10/144,073
• Publication Number:
  US 20030212743A1
• Time in Patent Office: 2,027 Days
• Field of Search: 709/220, 709/223, 709/245, 709/248, 709/201, 709/204, 709/238, 709/203, 709/202, 709/232, 709/215, 398/66, 455/519, 455/3.02, 455/500, 710/100, 719/311, 719/317, 719/319, 719/313
• US Class Current: 719/319
• CPC Class Codes: H04L 12/185 with management of multicast group membership

Abstract

Embodiments of methods and apparatuses for multicast handling in mixed core systems have been described. A method for multicast handling in mixed core systems includes configuring broadcast group registers located in targets. The method also includes receiving a request to create a broadcast group and creating the broadcast group. Finally, the method includes transmitting the broadcast group to targets with broadcast group registers that correspond to the broadcast group.

Claims

1. A method comprising:

creating a request from an initiator in a network having a broadcast group, wherein broadcast group registers associated with a first plurality of agents located on a chip are configured to store an associated value for the broadcast group and each agent is associated with at least one target;

sending the request to the first plurality of agents in the network, wherein the network, including the initiator and the first plurality of agents and associated targets, is located within a single device;

extracting for each agent a current value of the broadcast group from the request;

analyzing for each agent the current value of the broadcast group and a previously stored value of the broadcast group register associated with each agent; and

determining for each agent whether to accept the request for each agent based on the current value of the broadcast group and the previously stored value of the broadcast group register associated with each agent, wherein at least one agent ignores the request based on the current value of the broadcast group not matching the previously stored value of the broadcast group register associated with the at least one agent such that the at least one agent does not receive the request or at least one agent accepts the request based on the current value of the broadcast group matching the previously stored value of the broadcast group register associated with the at least one agent such that the at least one agent does receive the request.

2. The method of claim 1 wherein the broadcast group is encoded in an address.

3. The method of claim 2 wherein the encoding in the address is in a group of contiguous address bits, and wherein each agent is associated with a single target.

4. A processing system comprising a processor, which when executing a set of instructions performs the method of claim 1.

5. A machine-readable medium having stored thereon instructions, which when executed by the machine, to cause the machine to perform the method of claim 1.

10. The method of claim 1 further comprising:

receiving a multicast; and

determining if the multicast is intended for a core;

and if so, then transmitting the multicast within the core.

11. The method of claim 10 wherein transmitting the multicast within the core further comprises cascaded multicasting within the core.

12. A processing system comprising a processor, which when executing a set of instructions, to perform the method of claim 10.

13. A machine-readable medium having stored thereon instructions, which, when executed by a machine, to cause the machine to perform the method of claim 10.

14. The method of claim 1 further comprising:

receiving a broadcast command from a first address space; and

mapping the broadcast command to at least one of a write command, and a second address space.

15. The method of claim 14 further comprising mapping the write to a third address space.

16. The method of claim 14 wherein the second address space is a plurality of different address spaces.

17. A processing system comprising a processor, which when executing a set of instructions, to perform the method of claim 15.

18. A machine-readable medium having stored thereon instructions, which, when executed by a machine, to cause the machine to perform the method of claim 14.

19. The method of claim 1 further comprising:

defining the broadcast group;

initializing the broadcast group registers in the plurality of targets;

(a) creating a request;

(b) choosing a first group of broadcast registers from the initialized broadcast group of registers;

(c) sending the request to the first group; and

repeating (a) through (c) if other requests are pending.

20. A processing system comprising a processor, which when executing a set of instructions, to perform the method of claim 19.

21. A machine-readable medium having stored thereon instructions, which when executed by a machine, to cause the machine to perform the method of claim 19.

22. The method of claim 1 further comprising:

extracting the broadcast group from the request;

expanding the broadcast group into a bit vector;

comparing the bit vector with a previously stored broadcast group register value;

determining if there is a match between the bit vector and the group register previously stored value; and

accepting the request if the bit vector and the group register previously stored value match.

23. The method of claim 22 further comprising:

examining the accepted request with respect to a core capability; and

determining if the core capability requires an operation selected from the group consisting;

of mapping attributes of the core into the accepted request to match the core, creating a cascaded broadcast, mapping the accepted request into a write command, mapping the accepted request into a different address space.

24. The method of claim 1 further comprising:

receiving a nobackpressure thread; and

receiving a backpressure thread.

25. The method of claim 24 further comprising receiving the threads in a single queue.

26. The method of claim 25 wherein the threads may be received in any order.

27. The method of claim 24 further comprising receiving the nobackpressure thread in a nobackpressure queue, and receiving the backpressure thread in a backpressure queue.

28. The method of claim 27 wherein the threads may be received in any order.

29. The method of claim 24 wherein the receiving further comprises receiving over a single interface.

30. The method of claim 24 further comprising a connection identification to identify the threads.

31. The method of claim 24 further comprising receiving a plurality of nobackpressure threads.

32. The method of claim 31 wherein two or more of the plurality of nobackpressure threads are received in a single queue.

33. The method of claim 24 further comprising receiving a burst of requests through a nobackpressure queue.

34. The method of claim 24 further comprising receiving multiple bursts of requests and accepting them into a backpressure queue and nobackpressure queue.

35. The method of claim 1 further comprising:

sending a nobackpressure thread; and

sending a backpressure thread.

36. The method of claim 35 wherein the sending further comprises sending over a single interface.

37. The method of claim 1 further comprising:

configuring broadcast group registers associated with a second plurality of targets located on at least two chips;

creating a request having a broadcast group; and

broadcasting the request to the second plurality of targets located on at least two chips.

38. The method of claim 1 further comprising:

determining for each target if a match occurs between a value of the broadcast group and a previously stored value of a broadcast group register associated with each target.

6. An apparatus comprising:

means for creating a request from an initiator in a network having a broadcast group, wherein broadcast group registers associated with a first plurality of agents located on a chip are configured to store an associated value for the broadcast group and each agent is associated with at least one target;

means for sending the request to the first plurality of agents in the network, wherein the network, including the initiator and the first plurality of agents and associated targets, is located within a single device;

means for extracting for each agent a current value of the broadcast group from the request;

means for analyzing for each agent the current value of the broadcast group and a previously stored value of the broadcast group register associated with each agent; and

means for determining for each agent whether to accept the request for each agent based on the current value of the broadcast group and the previously stored value of the broadcast group register associated with each agent, wherein at least one agent ignores the request based on the current value of the broadcast group not matching the previously stored value of the broadcast group register associated with the at least one agent such that the at least one agent does not receive the request or at least one agent accepts the request based on the current value of the broadcast group matching the previously stored value of the broadcast group register associated with the at least one agent such that the at least one agent does receive the request.

7. The apparatus of claim 6 further comprising means for encoding the broadcast group.

8. The apparatus of claim 7 further comprising means for encoding the broadcast group in an address.

9. A machine-readable medium having stored thereon information including data and instructions, which when executed by the machine to cause the machine to generate a representation of the apparatus of claim 6.