Configuration based cache coherency protocol selection

Patent metadata

Field	Value
Publication number	US-2016147658-A1
Application number	US-201414548988-A
Country	US
Kind code	A1
Filing date	Nov 20, 2014
Priority date	Nov 20, 2014
Publication date	May 26, 2016
Grant date	—

Abstract

Topology of clusters of processors of a computer configuration, configured to support any of a plurality of cache coherency protocols, is discovered at initialization time to determine which one of the plurality of cache coherency protocols is to be used to handle coherency requests of the configuration

First claim

1 . A computer system for selecting and performing one of a plurality of cache coherency protocols, the computer system comprising:
a first node of a configuration comprising one or more communicatively coupled nodes, each node comprising one or more communicatively coupled computer elements, each computer element consisting of at least one of a cluster or a storage controller (SC) function, each cluster comprising one or more processors, the SC function comprising a shared cache, the shared cache being shared by processors of each cluster, the first node configured to perform a method, said method comprising: initiating communication, by each element of the configuration, to other elements of the configuration to determine configuration topology; based on a result of the initiated communication, storing a coherency value by an element of the first node; based on the coherency value, selecting, by each element of the first node, a cache coherency protocol of the plurality of cache coherency protocols to be used by the first node; and using, by the first node, the selected cache coherency protocol to handle cache coherency requests.

2 . The computer system according to claim 1 , wherein the configuration topology consists of the first node, wherein the plurality of cache coherency protocols comprise a first protocol that does not use an SC function to maintain cache coherency of the first node and a second protocol that uses an SC function to maintain cache coherency of the first node.

3 . The computer system according to claim 1 , wherein the configuration topology consists of the first node, the using the selected cache coherency protocol further comprising accessing an inclusive directory of the SC function of the first node to handle the cache coherency requests.

4 . The computer system according to claim 1 , wherein the configuration topology further comprises one or more second nodes, the first node communicatively coupled to said one or more second nodes, the using the selected cache coherency protocol further comprises:
accessing, by the first node, an inclusive directory of the SC function of the first node to handle the cache coherency requests; based on determining, by the first node, that the cache coherency operation can be performed solely by the first node, performing the cache coherency operation, by the first node, without accessing caches of the one or more second nodes; and based on determining, by the first node, that the cache coherency operation must access said one or more second nodes, performing the cache coherency operation by accessing caches of said one or more second nodes.

5 . The computer system according to claim 1 , wherein the plurality of cache coherency protocols comprise a third protocol and a fourth protocol for maintaining cache coherency of the node wherein the third protocol is the MESI protocol and the fourth protocol is the MOESI protocol.

6 . The computer system according to claim 2 , wherein the configuration topology further comprises one or more second nodes, the first node communicatively coupled to said one or more second nodes, wherein the SC function of the first node determines that one of the first protocol and second protocol is to be used in local cache coherency operations within the first node and that a different one of the first protocol and second protocol is to be used in global cache coherency operations involving said one or more second nodes.

7 . The computer system according to claim 1 , wherein the configuration topology further comprises one or more second nodes, the first node communicatively coupled to said one or more second nodes, wherein a first plurality of nodes are communicatively coupled by an S-Bus, the first plurality of nodes communicatively coupled to other nodes of the plurality of nodes by one or more A-Buses wherein coherency operations are broadcast to all nodes by way of the S-Bus and-the one or more A-Buses.

8 . A computer program product for selecting and performing one of a plurality of cache coherency protocols in a computer system comprising a first node of a configuration comprising one or more communicatively coupled nodes, each node comprising one or more communicatively coupled computer elements, each computer element consisting of at least one of a cluster or a storage controller (SC) function, each cluster comprising one or more processors, the SC function comprising a shared cache, the shared cache being shared by processors of each cluster, the computer program product comprising:
a computer readable storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising: initiating communication, by each element of the configuration, to other elements of the configuration to determine configuration topology; based on a result of the initiated communication, storing a coherency value, by an element of the first node; based on the coherency value, selecting, by each element of the first node, a cache coherency protocol of the plurality of cache coherency protocols to be used by the first node; and using, by the first node, the selected cache coherency protocol to handle cache coherency requests.

9 . The computer program product according to claim 8 , wherein the configuration topology consists of the first node, wherein the plurality of cache coherency protocols comprise a first protocol that does not use an SC function to maintain cache coherency of the first node and a second protocol that uses an SC function to maintain cache coherency of the first node.

10 . The computer program product according to claim 8 , wherein the configuration topology consists of the first node, wherein the using the selected cache coherency further comprises accessing an inclusive directory of the SC function of the first node to handle cache coherency requests.

11 . The computer program product according to claim 8 , wherein the configuration topology further comprises one or more second nodes, the first node communicatively coupled to said one or more second nodes, the using the selected cache coherency protocol further comprises:
accessing, by the first node, an inclusive directory of the SC function of the first node to handle the cache coherency requests; based on determining, by the first node, that the cache coherency operation can be performed solely by the first node, performing the cache coherency operation, by the first node, without accessing caches of the one or more second nodes; and based on determining, by the first node, that the cache coherency operation must access said one or more second nodes, performing the cache coherency operation by accessing caches of said one or more second nodes.

12 . The computer program product according to claim 8 , wherein the plurality of cache coherency protocols comprise a third protocol and a fourth protocol for maintaining cache coherency of the node wherein the third protocol is the MESI protocol and the fourth protocol is the MOESI protocol.

13 . The computer program product according to claim 9 , wherein the configuration topology further comprises one or more second nodes, the first node communicatively coupled to said one or more second nodes, wherein the SC function of the first node determines that one of the first protocol and second protocol is to be used in local cache coherency operations within the first node and that a different one of the first protocol and second protocol is to be used in global cache coherency operations involving said one or more second nodes.

14 . The computer program product according

Assignees

• IBM
• IBM

Inventors

• Ambroladze Ekaterina M
• Berger Deanna P
• Fee Michael
• O'Neill Jr Arthur J
• Sonnelitter Iii Robert J

Classifications

• G06F12/0815 Primary

  Cache consistency protocols · CPC title

• G06F12/0817

  using directory methods · CPC title

• G06F12/084

  with a shared cache · CPC title

• G06F12/0831

  using a bus scheme, e.g. with bus monitoring or watching means · CPC title

• G06F2212/1016

  Performance improvement · CPC title