公开(公告)号：GB2522653A

公开(公告)日：2015-08-05

申请号：GB201401669

申请日：2014-01-31

Applicant: IBM

Inventor： AUERNHAMMER FLORIAN ALEXANDER ,  MCDONALD JOSEPH GERALD

IPC: G06F13/40

Abstract: A bridge for coupling a requesting interconnect and a serving interconnect connected to a number of coherent units in a computer system is disclosed. The bridge includes N machines 131-13N, a tracking entity 121, a first arbiter 122, and a second arbiter 123. Each of the N machines is configured to handle requests from the requesting interconnect and to handle allocated responses from the serving interconnect. Each of the N machines has an allocated local count. The tracking entity is configured to track priorities of the N machines for issuing received requests to the serving interconnect and for issuing received responses to the requesting interconnect based on the N local counts. The first arbiter is configured to arbitrate the received requests to be issued to the serving interconnect based on the tracked priorities. Further, the second arbiter is configured to arbitrate the received responses to he issued to the requesting interconnect based on the tracked priorities. The tracking entity is configured to track the priorities of the N machines for handling requests based on the N local counts and a status of each of the N machines. The invention is said to improve the ordering for requests while supporting retries.

公开(公告)号：GB2525577A

公开(公告)日：2015-11-04

申请号：GB201401670

申请日：2014-01-31

Applicant: IBM

Inventor： AUERNHAMMER FLORIAN ALEXANDER ,  MCDONALD JOSEPH GERALD

IPC: G06F13/40

Abstract: A bridge 100 for coupling a requesting interconnect and a serving interconnect connected to a number of coherent units in a computer system is proposed. The bridge includes a first interface 111 and an issuer 120. The first interface is configured to receive a read request and a number of write requests that the read request depends on from the requesting interconnect. The issuer is configured to issue the received number of dependent write requests to the serving interconnect. The issuer also issues the received read request to the serving interconnect after issuing the number of dependent write requests and before receiving write acknowledgments for the issued dependent write requests from the serving interconnect. The bridge may further comprise a plurality of read machines 131-13N for handling read requests received at the first interface and a plurality of write machines 141-14M for handling write requests received at the first interface. The read and write machine can be controlled by the issuer. The first interface can be coupled to the requesting interconnect and a second interface can be connected to a serving interconnect, the interfaces each having a plurality of buffers 111-114, 151-154. The invention reduces latency and improves throughput.