WO2015027806A1

WO2015027806A1 - Read and write processing method and device for memory data

Info

Publication number: WO2015027806A1
Application number: PCT/CN2014/083873
Authority: WO
Inventors: 董建波; 张立新; 侯锐; 李晔
Original assignee: 华为技术有限公司
Priority date: 2013-08-26
Filing date: 2014-08-07
Publication date: 2015-03-05
Also published as: CN104424105A; CN104424105B

Abstract

A read and write processing method and device for memory data, which relate to the technical field of computer systems, and solve the problem in the prior art that the transmission efficiency is low due to the fact that access operations of a plurality of memory data with smaller data amount waste more overhead on a data request packet. The method comprises: a source node acquiring various request packets to be sent; parsing destination node identifiers of the various request packets to be sent, and judging whether the destination node identifiers of the various request packets to be sent are the same or not; packaging the at least two request packets to be sent with the same destination node identifier into a merged request packet by respectively taking them as sub-request packets; and sending the merged request packet to a destination node corresponding to the destination node identifier, the merged request packet being used for indicating the destination node to conduct a read and/or write operation on the memory data of the destination node. The method is applied to the process of read and write processing of the memory data.

Description

The present invention claims to be submitted to the Chinese Patent Office on August 26, 2013, and the application number is 201310376301.X, and the invention name is "a method and device for reading and writing memory data, Priority of the Chinese Patent Application, the entire contents of which is incorporated herein by reference.

Technical field

The present invention relates to the field of computer system technologies, and in particular, to a method and apparatus for reading and writing memory data.

Background technique

In a multi-node computer system, the efficiency of data interaction between local and remote nodes directly affects the '1' energy of the entire system. Remote Direct Memory Access (RDMA) is the rapid movement of memory data from a system to a system. In the remote system memory, without affecting the operating system, for example, without the CPU intervention, the memory data of one node is directly transferred to the physical memory of another node, thereby reducing the impact of data transmission on the processing components. Improve system performance.

In RDMA read and write operations, the data exchange granularity between memory and Cache is a Cache Line with a size of 64 Bytes (bytes) or 128 Bytes, if each Cache Line's data is packed into a request. Packets are sent out. This small-scale data transfer has a certain amount of link bandwidth waste. For example, the packet header of PCIE (Peripheral Component Interface Express) may have 16 Bytes, and if Payload (effective) The payload is 16 bytes, and the link bandwidth utilization is only 50%. Therefore, multiple remote data access operations with small data volume waste more overhead on the data request packet, resulting in inefficient transmission.

Summary of the invention Embodiments of the present invention provide a method and apparatus for reading and writing memory data, which improves link utilization efficiency.

In order to achieve the above object, embodiments of the present invention use the following technical solutions:

In a first aspect, the present invention provides a method for reading and writing memory data, including: acquiring each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet;

Parsing the destination node identifiers of the to-be-sent request packets, and determining whether the destination node identifiers of the respective to-be-sent request packets are the same;

The at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet, and the merge request packet includes a common header and at least two sub-request packets, where the common header is included. The merging packet identifier, the sub-request packet number identifier, and the source node identifier, where the merging packet identifier is used to indicate that the request packet is a combined packet of the at least two sub-request packet combinations, and the sub-request packet number identifier is used by And indicating the number of the sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information; and sending the merge request packet to the destination node corresponding to the destination node identifier, where The merge request packet is used to instruct the destination node to perform read and/or write operations on the memory data of the destination node.

In a first possible implementation manner of the first aspect, the packaging into a merge request packet includes:

And setting the merged packet identifier to a predetermined value, where the predetermined value indicates that the request packet is a merged package of the at least two sub-request packet combinations;

Obtaining the number of sub-request packets, and writing the number into the sub-request packet number identifier. With reference to the first aspect or the first possible implementation manner of the first aspect, in a second possible implementation manner, the parsing the target node identifiers of the to-be-sent request packets, and determining the to-be-sent requests Whether the destination node identifier of the packet is the same includes:

Obtaining the destination node identifier carried in each request packet to be sent;

Determining whether the destination node identifiers of the to-be-sent request packets are the same; The packaging into a merge request package includes:

Extracting the destination node identifier in the sub-request packet, and placing the destination node identifier in a common header of the merge request packet.

With reference to the first aspect, or the first possible implementation manner of the first aspect, in a third possible implementation manner, the parsing the target node identifiers of the to-be-sent request packets, and determining the to-be-sent requests Whether the destination node identifier of the packet is the same includes:

Obtaining a memory address of the destination node carried in each request packet to be sent;

And searching, according to the destination node memory address, the destination node identifier of the destination node's memory address in the address attribution table, where the address attribution table is a correspondence table between the destination node memory address and the destination node identifier;

Determining whether the destination node identifiers of the respective to-be-sent request packets are the same.

In conjunction with the second possible implementation or the third possible implementation of the first aspect, in a fourth possible implementation, the sub-request packet includes a write request sub-packet and/or a read request sub-package, The request sub-packet header includes a write request identifier that identifies the write request sub-packet, and the write request information in the write request sub-package includes data to be written to the destination node and a memory address of the destination node to be written. The read request sub-packet header includes a read request identifier that identifies the read request sub-packet, and the read request information in the read request sub-package includes a memory address of the destination node where the data to be read is located, and the method further Includes:

Receiving a write request response message sent by the destination node, where the write request response message includes a write request identifier corresponding to the write request sub-package with a successful write request; and/or,

Receiving a read request response message sent by the destination node, where the read request response message includes a read request identifier corresponding to the read request sub-package with a successful read request, and a read payload, where the payload is Read the data corresponding to the remote memory address in the request subpackage.

In combination with the first aspect or the first possible implementation or the second possible implementation or the third possible implementation or the fourth possible implementation of the first aspect, in a fifth possible implementation manner And parsing the destination node identifiers of the to-be-sent request packets, and determining whether the destination node identifiers of the to-be-sent request packets are the same include: Determining, according to the destination node identifier corresponding to each to-be-sent request packet, a routing port corresponding to the destination node identifier in the routing table, where the routing table includes a routing port corresponding to the destination node identifier, where the routing port is Determining a port on the path of the sending request packet to the destination node;

Determining whether the routing ports of the respective to-be-sent request packets are the same.

In a second aspect, the present invention provides a method for reading and writing memory data, including: receiving a request packet;

Determining whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a merge request packet of at least two sub-request packet combinations, where the sub-request package includes: a read request sub-package, and/ Or write a request sub-package;

When the request packet carries the merged package identifier, the request packet is determined to be a merge request packet, where the merge request packet includes a common header and a sub-request packet, where the common header includes the merge packet identifier and the number of sub-request packets. The identifier of the sub-request packet is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information;

And extracting, according to the number of the sub-request packets and the sub-request packet headers carried in the merge request packet, a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet;

Corresponding sub-request packets are processed according to each of the sub-request packet headers and sub-request information. In a first possible implementation manner of the second aspect, the processing, by the sub-request packet header and sub-request information, processing the corresponding sub-request packet includes:

Writing the data to be written into the destination according to the write request identifier in the write request sub-packet header, the data to be written in the write request information of the write request sub-packet, and the memory address of the destination node to be written Node memory address; and/or,

Reading, according to the read request identifier in the read request sub-packet header, and the destination node memory address of the read request information of the read request sub-packet, the destination node memory address is read in the destination node memory address Read the data. With reference to the first possible implementation manner of the second aspect, in a second possible implementation manner, after the extracting the sub-request packet corresponding to the number of sub-request packets in the merge request packet , Also includes:

The source node identifier is written into a sub-request packet header of each sub-request packet such that the destination node sends a response message to the source node.

With the second possible implementation of the second aspect, in a third possible implementation, the sending, by the destination node, the response message to the source node includes:

Determining whether the source node identifiers corresponding to the respective response messages to be sent are the same;

The response message to be sent with the same source node identifier is packaged as a sub-response message packet, and the merge response message packet includes a write request sub-response message and/or a read request sub-response message.

Sending the merge response message packet to a source node corresponding to the source node identifier.

In a fourth possible implementation manner of the second aspect, when the destination node is a routing port, extracting, in the extracting the request packet, a number of sub-request packets corresponding to the number of the sub-request packets After that, it also includes:

The source node identifier in the common header of the merge request packet is extracted and written in the header of each subrequest packet.

In a third aspect, the present invention provides an apparatus for reading and writing memory data, including: an obtaining unit, configured to acquire each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet;

a determining unit, configured to parse the destination node identifiers of the to-be-sent request packets, and determine whether the destination node identifiers of the to-be-sent request packets are the same;

a processing unit, configured to package the at least two to-be-sent request packets having the same destination node identifier as a sub-request packet, and the merge request packet includes a common header and at least two sub-request packets. The common packet header includes a merge packet identifier, a sub-request packet number identifier, and a source node identifier, where the merge packet identifier is used to indicate that the request packet is a merge packet of the at least two sub-request packet combinations, and the sub-request Packet number identifier is used to indicate a number of sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information;

a sending unit, configured to send the merge request packet to a destination node corresponding to the destination node identifier, where the merge request packet is used to instruct the destination node to perform read and/or write operations on the memory data of the destination node .

In a first possible implementation manner of the third aspect, the processing unit includes: a first identifier module, configured to set the merged packet identifier to a predetermined value, where the predetermined value indicates that the request packet is a merged package of at least two sub-request package combinations;

The second identifier module is configured to obtain the number of the sub-request packets, and write the number into the sub-request packet number identifier.

With reference to the third aspect or the first possible implementation manner of the third aspect, in a second possible implementation manner, the determining unit includes:

a first obtaining module, configured to obtain a destination node identifier carried in each to-be-sent request packet; a first determining module, configured to determine whether the destination node identifiers of the to-be-sent request packets are the same;

The processing unit further includes:

And a third identifier module, configured to extract a destination node identifier in the sub-request packet, and place the destination node identifier in a common header of the merge request packet.

With reference to the third aspect or the first possible implementation manner of the third aspect, in a third possible implementation, the determining unit includes:

a second acquiring module, configured to obtain a memory address of the destination node carried in each request packet to be sent;

a first search module, configured to search, according to the destination node memory address, a destination node identifier of the destination node memory address in the address attribution table, where the address attribution table is a correspondence between the destination node memory address and the destination node identifier Table

The second determining module is configured to determine whether the destination node identifiers of the to-be-sent request packets are the same. With reference to the second possible implementation manner of the third aspect, or the third possible implementation manner, in a fourth possible implementation manner, the sub-request packet included in the merge request packet obtained by the processing unit includes writing a request sub-packet and/or a read request sub-package, the write request sub-packet header includes a write request identifier for identifying the write request sub-packet, and the write request information in the write request sub-package includes a write request information to be written to the destination node The data and the destination node memory address to be written, the read request sub-packet header includes a read request identifier identifying the read request sub-packet, and the read request information in the read request sub-package includes the data to be read The destination node memory address, the device further includes:

a receiving unit, configured to receive a write request response message sent by the destination node, where the write request response message includes a write request identifier corresponding to the write request sub-package with a successful write request; and/or, a read request response message sent by the destination node, where the read request response message includes a read request identifier corresponding to the read request sub-package with a successful read request, and a read payload, where the payload is the read requester The data corresponding to the remote memory address in the package.

In combination with the third aspect or the first possible implementation or the second possible implementation or the third possible implementation or the fourth possible implementation of the third aspect, in a fifth possible implementation manner The determining unit further includes:

a second search module, configured to search, according to the destination node identifier corresponding to each to-be-sent request packet, a routing port corresponding to the destination node identifier in a routing table, where the routing table includes a routing port corresponding to the destination node identifier The routing port is a port on the path of the to-be-sent request packet to the destination node;

The third determining module is configured to determine whether the routing ports of the to-be-sent request packets are the same.

A fourth aspect of the present invention provides a memory data read/write processing apparatus, including: a receiving unit, configured to receive a request packet;

a determining unit, configured to determine whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a combined request packet of at least two sub-request packet combinations, The sub-request package includes: a read request sub-package, and/or a write request sub-package;

a determining unit, configured to: when the request packet carries a merged package identifier, determine that the request packet is a merge request packet, where the merge request packet includes a common header and a subrequest package, where the common header includes a merge packet identifier, a sub-request packet number identifier and a source node identifier, where the sub-request packet number identifier is used to indicate a number of sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information;

An extracting unit, configured to extract, according to the number of sub-request packets and the sub-request packet header carried in the merge request packet, a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet ;

And a processing unit, configured to process the corresponding sub-request packet according to each of the sub-request packet header and sub-request information.

In a first possible implementation manner of the fourth aspect, the processing unit includes: a first processing module, configured to identify, according to a write request identifier in the write request sub-packet header, the write request information of the write request sub-packet The data to be written and the memory address of the destination node to be written, and the data to be written is written into the memory address of the destination node;

a second processing module, configured to: according to the read request identifier in the read request sub-packet header, and the memory address of the destination node where the data to be read in the read request information of the read request sub-package is located, in the destination node memory The data to be read is read in the address.

With reference to the first possible implementation of the fourth aspect, in a second possible implementation, the device further includes:

And an identifier writing unit, configured to write the source node identifier into a sub-request packet header of each sub-request packet, so that the destination node sends a response message to the source node.

With reference to the second possible implementation manner of the fourth aspect, in a third possible implementation manner, the determining unit is further configured to determine whether the source node identifiers corresponding to the respective response messages to be sent are the same;

The processing unit is further configured to package the to-be-sent response message with the same source node identifier as a sub-response message, into a merge response message packet, and the merge response message The packet includes a write request sub-response message and/or a read request sub-response message;

The device also includes:

And a sending unit, configured to send the merge response message packet to a source node corresponding to the source node identifier.

In a fourth possible implementation manner of the fourth aspect, when the destination node is a routing port, the processing unit is further configured to extract a source node identifier in a common header of the merge request packet, and write each The header of the sub-request package.

The method and device for reading and writing memory data provided by the embodiment of the present invention, the source node acquiring each to-be-sent request packet, the to-be-sent request packet includes: a read request packet, and/or a write request packet, and then parsing the And determining, by the target node identifiers of the to-be-sent request packets, the destination node identifiers of the respective to-be-sent request packets are the same, and the at least two to-be-sent request packets having the same destination node identifier are respectively packaged as sub-request packets. a merge request packet, the merge request packet is sent to the destination node corresponding to the destination node identifier, and the merge request packet is used to instruct the destination node to perform read and/or write operations on the memory data of the destination node. . The embodiment of the present invention solves the problem that the memory data read/write operation with a small amount of data is small in the prior art, which wastes more overhead on the data request packet, resulting in low transmission efficiency and improves the utilization efficiency of the link.

DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief description of the drawings used in the embodiments or the prior art description will be briefly made. Obviously, the drawings in the following description It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any creative work.

1 is a flow chart of a method for reading and writing memory data according to an embodiment of the present invention;

2 is a schematic diagram of a read request packet format; 3 is a schematic diagram of a write request packet format;

Figure 4 is a schematic diagram of another read request packet format;

Figure 5 is a schematic diagram of another write request packet format;

FIG. 6 is a schematic diagram of a merge write request packet format according to an embodiment of the present invention; FIG. 7 is a schematic diagram of a merge read request packet format according to an embodiment of the present invention; FIG. 8 is a schematic diagram of a merge provided by an embodiment of the present invention; FIG. 9 is a schematic diagram of another merge write request packet format according to an embodiment of the present invention; FIG. 10 is a schematic diagram of another merge read request packet format according to an embodiment of the present invention; 1 is a schematic diagram of another merging read and write request packet format provided by an embodiment of the present invention; FIG. 1 is a flowchart of another method for reading and writing memory data according to an embodiment of the present invention;

FIG. 13 is a flow chart of still another method for reading and writing memory data according to an embodiment of the present invention;

FIG. 14 is a structural diagram of an apparatus for reading and writing memory data according to an embodiment of the present invention;

FIG. 15 is a structural diagram of a processing unit in a memory data read/write processing apparatus according to an embodiment of the present invention;

FIG. 16 is a structural diagram of another apparatus for reading and writing memory data according to an embodiment of the present invention;

FIG. 1 is a structural diagram of a determining unit in a memory data read/write processing apparatus according to an embodiment of the present invention;

FIG. 18 is a structural diagram of another apparatus for reading and writing memory data according to an embodiment of the present invention;

FIG. 19 is a structural diagram of a determining unit in a memory data read/write processing apparatus according to an embodiment of the present invention;

FIG. 20 is a structural diagram of still another apparatus for reading and writing memory data according to an embodiment of the present invention; FIG. FIG. 2 is a structural diagram of a processing unit in another read/write processing device for memory data according to an embodiment of the present disclosure;

FIG. 22 is a structural diagram of still another apparatus for reading and writing memory data according to an embodiment of the present invention;

FIG. 2 is a structural diagram of another apparatus for reading and writing memory data according to an embodiment of the present invention; FIG.

24 is a hardware structural diagram of a device for reading and writing memory data according to an embodiment of the present invention;

FIG. 25 is a hardware structural diagram of another apparatus for reading and writing memory data according to an embodiment of the present invention.

detailed description

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

When the local memory and the remote memory perform data interaction, the memory data read and write operation with a small amount of data wastes a lot of overhead on the data request packet, resulting in inefficient transmission. The embodiment of the present invention provides a memory data read. The write processing method is to combine the requests with the same destination node address to improve the utilization of the link. The execution source source node of the method of the embodiment of the present invention may be a computer node (also referred to as a computing node), as shown in FIG. 1 . As shown, the method includes the following steps:

1 01. Acquire each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet.

Each of the to-be-sent request packets acquired by the source node includes a read request packet, and/or a write request packet. The target node identifiers of the to-be-sent request packets are parsed, and the destination node identifiers of the to-be-sent request packets are determined to be the same.

The above step 012 can be implemented in two ways.

Manner 1: First, obtain the destination node identifiers carried in each of the to-be-sent request packets, and then determine whether the destination node identifiers of the to-be-sent request packets are the same.

In the first method, before the source node performs the read/write processing of the memory data to the destination node, the source node sends a memory borrowing request to the destination node according to the prior art, and the destination node determines according to the state of the memory. Accepting or rejecting, after the destination node accepts the request of the source node, the destination node sends the allocated memory address and the address length to the source node, where the source node is the memory of the destination node The address allocates a local memory address, and establishes an address mapping table, where the address mapping table is a mapping list of the local memory address and the destination node identifier and the memory address of the destination node.

When the source node sends a read or write request packet to the destination node, the request packet includes the source node identifier, the destination node identifier, and a memory address of the destination node, where the destination node identifier And the memory address of the destination node is obtained by the source node by using the address mapping table after obtaining the request that carries the local memory address, as shown in FIG. 2 To read the read request packet, Figure 3 shows the write request packet to be sent. The destination node identifiers carried in the to-be-sent request packets shown in FIG. 2 and/or FIG. 3 are obtained, and the destination node identifiers of the respective to-be-sent request packets are determined to be the same.

Manner 2: obtaining a destination node memory address carried in each request packet to be sent; searching for the destination node identifier to which the destination node memory address belongs in the address attribution table according to the destination node memory address, and the address attribution table is for the purpose a correspondence table between the node memory address and the destination node identifier; determining whether the destination node identifiers of the respective to-be-sent request packets are the same.

In the second method, the PC IE is used as an example to illustrate that the read request packet to be sent is as shown in FIG. 4, and the write request packet to be sent is as shown in FIG. 5, wherein the read request packet and the to-be-sent are sent. The address carried in the write request packet is the destination node memory address, and the destination node identifier of the destination node memory address is searched in the address attribution table, where the address attribution table is a correspondence table between the node memory address and the destination node identifier. For example, in the address attribution table, when the destination node memory address is OHIOOO-OHlfff, the corresponding destination node identifier is 1, and the destination node memory address is When the value is 0H2000-0H2fff, the corresponding destination node ID is 2. When the destination node memory address is 0H3000-0H2fff, the corresponding destination node identifier is 3. After obtaining the memory address of the destination node carried in the to-be-sent request packet, the destination node identifier corresponding to the memory address of the destination node is obtained according to the address attribution table.

For example, when the memory address of the destination node carried in the read request packet to be sent is 0H1001, and the memory address of the destination node carried in the write request packet to be sent is OH1ff0, the read request packet to be sent and The to-be-sent write request packet is in the range of the destination node memory address corresponding to the destination node identifier 1 , and therefore the to-be-sent read request packet and the destination node identifier of the to-be-sent write request packet are the same.

103. The at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet, and the merge request packet includes a common packet header and at least two sub-request packets. The header includes a merge packet identifier, a sub-request packet number identifier, and a source node identifier, where the merge packet identifier is used to indicate that the request packet is a combined packet of the at least two sub-request packets, and the number of the sub-request packets The identifier is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information.

When the sub-request packet is packaged into a merge request packet, a common header and a sub-request packet are included in the merge request packet, and the common packet header includes a merge package identifier (Combine, C) for indicating that the request packet is Determining a combined packet of at least two sub-request packet combinations, and setting the merged packet identifier to a predetermined value, where the predetermined value indicates that the request packet is a combined packet of the at least two sub-request packet combinations, for example, C= "1 The time indicates that the request packet is a merge request packet, and includes a number of sub-request packets carried by the merge request packet, which is used when the package request is a merge request packet. And obtaining the number of the sub-request packets, and writing the number into the sub-request packet number identifier; further including a source node identifier (Requester ID), configured to indicate the source node sent by the merge request packet, It is the identifier of the local source node, and the common header of the merge request packet also contains the public information in each sub-request packet header. The sub-request packet in the merge request packet includes a sub-request packet header and sub-request information, and the sub-request packet header includes a type of the sub-request packet

(Type) and length, where the length is specifically Trans a c t i on Lay Pa cke t (TLP l eng th ). The type is used to distinguish whether the sub-request packet is a write request sub-package or a read request sub-package.

Wherein, the write request sub-packets with the same destination node identifier may be packaged into one merge write request packet, or the read request sub-packets with the same destination node identifier may be packaged into one merge read request packet, or to further improve link utilization. The write request sub-packet and the read request sub-package with the same destination node identifier are packaged into one merge read-write request packet, which is not limited herein.

In the merge request packet obtained by the method in the above step 102, the destination node identifier in the sub-request packet needs to be extracted, and the destination node identifier is placed in the common header of the merge request packet. As shown in FIG. 6, for the obtained merge write request packet, as shown in FIG. 7, for the obtained merge read request packet, as shown in FIG. 8, the obtained merge read/write request packet.

In the merge request packet obtained by the method 2 packaged in the above step 102, as shown in FIG. 9, the obtained merge write request packet is as shown in FIG. 10, and the obtained merge read request packet is obtained as shown in FIG. The merge read and write request packet. In addition, the header of each sub-request packet in the merge request packet obtained by the second method further includes a request code, which is a code of the source node to the sub-request packet, and is a unique identifier of each sub-request packet.

The merging request packet is sent to the destination node corresponding to the destination node identifier, where the merging request packet is used to instruct the destination node to perform read and/or write operations on the memory data of the destination node.

The method for reading and writing the memory data provided by the embodiment of the present invention, the source node acquires each to-be-sent request packet, and parses the destination node identifier of each to-be-sent request packet, and determines the destination node of each to-be-sent request packet. Whether the identifiers are the same, and then the at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet. And requesting the packet, and finally sending the merge request packet to the destination node corresponding to the destination node identifier, where the merge request packet is used to instruct the destination node to perform read and/or write operations on the memory data of the destination node. . The embodiment of the invention solves the problem that the memory data access operation with less data volume is less wasted on the data request packet in the prior art, resulting in low transmission efficiency and improving the utilization efficiency of the link.

The embodiment of the present invention further provides a method for reading and writing memory data. The target node of the method of the method of the embodiment of the present invention may be a remote computer node. As shown in FIG. 12, the method includes the following steps:

1 201. Receive a request packet.

1 202. Determine whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a merge request packet of at least two sub-request packet combinations, where the sub-request package includes: a read request sub-package , and / or write request sub-package.

When the destination node receives the request packet, it first determines whether the request packet includes the merged package identifier C. For example, when it is determined that C= "1", the request packet is a merge request packet.

1 203. When the request packet carries the merged package identifier, determine that the request packet is a merge request packet, where the merge request packet includes a common header and a subrequest package, where the common header includes a merge package identifier and a subrequest And the number of the sub-request packets is used to indicate the number of sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information;

1 204. Extract, according to the number of sub-request packets carried in the merge request packet, and each sub-request packet header, and extract a sub-request packet corresponding to the number of sub-request packets in the merge request packet.

When it is determined that the received request packet is a merge request packet, the common packet header of the merge request packet includes, in addition to the merge packet identifier, a sub-request packet number identifier and a source node identifier, where the sub-request packet number identifier is used for And indicating a number of sub-request packets carried by the merge request packet, where the source node identifier is used to indicate a source node that is sent by the merge request packet. Sub-request packet headers and sub-request information are included in each sub-request packet. Extracting, by using the number of sub-request packets carried in the merge request packet, and each sub-request packet header, a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet, where Sub-request packet number identification and type and length in each sub-request packet header, or using the sub-request packet number identifier and the format, type and length in each sub-request packet header, extracting the merge request packet from the The number of sub-request packets identifies the corresponding number of sub-request packets.

1 2 05. Process the corresponding sub-request packet according to each sub-request packet header and sub-request information.

Wherein, since the merge request packet may include a write request sub-packet and/or a read request sub-package, the write request sub-packet and the read request sub-package may be distinguished according to the format and/or type in the sub-request packet header, thereby The write request sub-packets and read request sub-packets of different requests are processed differently.

Referring to the prior art, for a sub-write request, the payload is written to the memory address of the destination node according to the memory address and payload of the destination node carried in the write request sub-packet; And reading, according to the memory address of the destination node carried in the read request sub-packet, content on the memory address of the destination node, and returning the content to the source node.

The method for reading and writing the memory data provided by the embodiment of the present invention, the destination node determines whether the received request packet carries the merged package identifier, and when the request packet carries the merged package identifier, determining that the request packet is a merge request a packet, and then, according to the number of sub-request packets carried in the merge request packet and each sub-request packet header, extracting a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet, and finally Corresponding sub-request packets are processed according to each of the sub-request packet headers and sub-request information. The embodiment of the present invention solves the problem that the memory data access operation with less data volume is less wasted on the data request packet in the prior art, resulting in low transmission efficiency and improved link utilization efficiency.

The embodiment of the present invention can be applied to the RDMA, and the to-be-sent request packets having the same destination node in the memory data of a computing node are packaged and combined, and sent together to the remote computing node corresponding to the destination node, specifically, the local Ca che will The request is sent to the local memory controller, packaged and merged, and then sent to the remote memory controller for unpacking, which will get The request is sent to the remote Cache. In this embodiment, a PC IE is taken as an example for description. FIG. 4 is a read request packet format in a standard PC IE, and FIG. 5 is a write request packet format, where a format and a type are used to distinguish whether the request packet is a read request packet or a write request packet. The transaction layer packet length is used to indicate the request packet length, the source node identifier is used to indicate the source node that sends the request packet, and the address in the read request packet indicates the address of the memory data to be read by the read request packet. The address in the write request packet indicates the payload to be carried in the write request packet, and the memory address to be written. The meanings of other fields may refer to the packet format of the PC IE in the prior art. The following application scenario is taken as an example to describe a method for reading and writing memory data in detail. As shown in FIG. 13 , the method includes the following steps:

1301, the source node acquires each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet;

1302. The source node parses the destination node identifiers of the to-be-sent request packets, and determines whether the destination node identifiers of the to-be-sent request packets are the same.

Obtaining a destination node memory address carried in each of the to-be-sent request packets; searching, according to the destination node memory address, the destination node identifier of the destination node memory address in the address attribution table, where the address attribution table is the destination node memory address Corresponding relationship table with the destination node identifier; determining whether the destination node identifiers of the respective to-be-sent request packets are the same.

The read request packet to be sent is as shown in FIG. 4, and the write request packet to be sent is as shown in FIG. 5, wherein the address to be sent in the read request packet and the to-be-sent write request packet is a destination node memory address, where the destination node identifier of the destination node is located in the address attribution table, where the address attribution table is a correspondence table between the destination node memory address and the destination node identifier, for example, in the address attribution table. When the destination node memory address is OH I OOO-OH1 fff, the corresponding destination node identifier is 1 and the destination node memory address is 0H2000-0H2f ff, the corresponding destination node identifier is 2, the purpose When the node memory address is 0H 3000-0H2f ff, the corresponding destination node ID is 3. After obtaining the memory address of the destination node carried in the to-be-sent request packet, the destination node identifier corresponding to the memory address of the destination node is obtained according to the address attribution table. For example, when the memory address of the destination node carried in the read request packet to be sent is

0H1001, when the memory address of the destination node carried in the to-be-sent write request packet is OH1ff0, the to-be-sent read request packet and the to-be-sent write request packet are both present in the destination node identifier 1 The target node has a memory address range, and therefore the read request packet to be sent and the destination node identifier of the to-be-sent write request packet are the same.

1303. The source node separately encapsulates the at least two to-be-sent request packets having the same destination node identifier as a sub-request packet into a merge request packet, where the merge request packet includes a common packet header and at least two sub-request packets. The common packet header includes a merge packet identifier, a sub-request packet number identifier, and a source node identifier, where the merge packet identifier is used to indicate that the request packet is a merge packet of the at least two sub-request packet combinations, and the sub-packet The number of request packets is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information.

Wherein, the write request sub-packets with the same destination node identifier may be packaged into one merge write request packet, or the read request sub-packets with the same destination node identifier may be packaged into one merge read request packet, or to further improve link utilization. The write request sub-packet and the read request sub-packet having the same destination node identifier are packaged into one merge read-write request packet, so the merge request packet includes a write request sub-packet and/or a read request sub-packet.

When the sub-request packet is packaged into a merge request packet, a common header and a sub-request packet are included in the merge request packet, and the common packet header includes a merge packet identifier (Combine, C) for indicating that the request packet is A merge request packet of a sub-request packet combination, for example, C= "1" indicates that the request packet is a merge request packet; and includes a combined request number (CR NUM ) for indicating the carry request packet The number of the sub-request packets further includes a source node identifier (Requester ID), which is used to indicate the source node that is sent by the merge request packet, that is, the identifier of the local source node, and the common header of the merge request packet further includes The public information in the header of each sub-request. The sub-request packet in the merge request packet includes a sub-request packet header and sub-request information, where the sub-request packet header includes a format (Fmt), a type (Type), and a length of the sub-request packet, wherein the length is specifically a transaction layer Packet length (Transaction Lay Packet, TLP length ). The format and type are used to distinguish whether the sub-request packet is a write request sub-package or a read request sub-package.

In addition, the header of each sub-request packet further includes a request encoding, which is a coding of the sub-request packet by the source node, and is a unique identifier of each sub-request packet, so that the source node receives the response returned by the destination node. In the message, the completion of the sub-request is determined according to the request code carried in the response message.

Specifically, as shown in FIG. 10, the two read request sub-packets are combined to obtain a combined read request packet, where the merged package identifier and the number of the sub-request packet identifiers are newly added additional information, where the number of the sub-request packets is identified. It is 4 bits, and it can be indicated that the number of read request sub-packets included in the combined read request packet is 16. The PCIE originally contains a request encoding (ie, tag) field, which is 8 bits in length. In the combined read request packet, the upper 4 bits of the field are shared, and the 0-3 bits are used to distinguish different read requests. The sub-packets can distinguish 16 read request sub-packets, and match the number represented by the sub-request packet number identification field. Wherein, the communication classification (TC), the pollution reservation bit (EP), the attribute (Attr), and the source node identifier (Requester ID) belong to a common header in the combined read request packet; format (Fmt), type (Type), whether ECRC (TD), TLP length, Request Coding, Last DW BE, and First Byte Alignment (1st DW BE) belong to the read request sub-packet header, the above fields The meaning can be referred to the read request packet format in the PCIE in the prior art.

As shown in FIG. 9, the two write request sub-packets are combined to obtain a merge write request packet, and the meaning of the field included in the merge write request packet is similar to the meaning of the field included in the merge read request packet shown in FIG. 10, except that The write request subpackage also contains the payload to be written to the destination node.

To further improve the link utilization, the write request sub-packet and the read request sub-packet of the same destination node identifier may be packaged into one merge request packet, as shown in FIG.

For the sub-requests in the above-mentioned merge request packet, the package merge is performed in the order in which the sub-requests are sent. 1304, the source node sends the merge request packet to a destination node corresponding to the destination node identifier, where the merge request packet is used to instruct the destination node to read and/or read memory data of the destination node. Write operation.

1 305. The destination node receives a request packet.

1 306. The destination node determines whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a combined request packet of at least two sub-request packet combinations, where the sub-request package includes: Read request sub-packages, and/or write request sub-packages.

1 307. When the request packet carries the merged package identifier, determine that the request packet is a merge request packet, where the merge request packet includes a common header and a subrequest package, where the common header includes a merge package identifier and a subrequest. And the number of the sub-request packets is used to indicate the number of sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information;

1 308. Extract, according to the number of sub-request packets carried in the merge request packet and each sub-request packet header, a sub-request packet corresponding to the number of sub-request packet identifiers in the merge request packet.

When it is determined that the request packet is a merge request packet, each sub-request packet is obtained according to the number of the sub-request packets, and the format, type, and length of each sub-request packet. In addition, since the merge request packet may include a write request sub-packet and/or a read request sub-package, the write request sub-package and the read request sub-package may be distinguished according to the format and type in the sub-request packet header, thereby being different. The requested write request sub-packet and the read request sub-package are processed differently.

1 309. Process the corresponding sub-request packet according to each sub-request packet header and sub-request information.

Writing the data to be written into the destination according to the write request identifier in the write request sub-packet header, the data to be written in the write request information of the write request sub-packet, and the memory address of the destination node to be written Node memory address; and/or, according to the read request header in the read request sub-packet header And the memory address of the destination node where the data to be read is located in the read request information of the read request sub-packet, and the data to be read is read in the memory address of the destination node.

Further, after the destination node processes the sub-request packet, it is convenient to send a response message to the source node, and write the source node identifier into a sub-request packet header of each sub-request packet, or The other information of the common header in the merge request packet is written into the sub-request packet header of each sub-request packet, so that each sub-request packet is converted into the format before the merge.

In addition, after processing the sub-request, the destination node may return a response message to the source node to inform the source node of the completion status of each sub-request. There are two ways to get feedback:

method one:

The destination node sends the response message of the processed sub-request packet separately: acquiring the source node identifier in the write request sub-package;

Sending a write request response message to the source node corresponding to the source node identifier, where the write request response message includes a write request identifier corresponding to the write request sub-package with a successful write request.

When the write request sub-packet is a non-push-type write request, the destination node needs to send feedback to the source node, or may request a write request sub-packet that completes the write request, and sends a response to the source node. Message.

The source node receives a write request response message sent by the destination node, where the write request response message includes a write request identifier corresponding to the write request sub-package with a successful write request.

Referring to the prior art, after receiving the write request response message, the source node compares with the write request identifier in the sent write request sub-packet according to the carried write request identifier, and obtains which write request sub-packets are available. The destination node is successfully written, which write request sub-packets are not written, and the request packet for which no write request is successful may be sent again.

and / or,

Obtaining, by the destination node, the source node identifier in the read request sub-package;

Sending a read request response message to the source node corresponding to the source node identifier, the read request The response message includes a request code corresponding to the read request sub-packet whose read request is successful, and an acquired payload, where the payload is data acquired by the sub-read request.

Receiving, by the source node, a read request response message sent by the destination node, where the read request response message includes a read request identifier corresponding to the read request sub-package with a successful read request, and the acquired payload, the payload The data obtained for the sub-read request.

The payload in the read request sub-packet is data obtained by the destination node according to the remote memory address in the read request sub-package, and the destination node will successfully obtain the read request corresponding to the sub-read request of the payload. Sending, by the source node, the source node, after receiving the payload, according to which addresses of the local end need to read data when the read request sub-packet is sent, according to the read request identifier The corresponding payload can be read back to the corresponding address.

Method 2:

After obtaining the merge response message packet, the source node obtains a corresponding response message according to the write request identifier and/or the read request identifier in the packet, and the processing manner after the acquisition is the same as that described in the first method, where No longer.

The method in the embodiment of the present invention may also be applied to data transmission between multiple ports in the system, and searching for the destination node identifier in the routing table according to the destination node identifier corresponding to each to-be-sent request packet. a routing port, the routing table includes a routing port corresponding to the destination node identifier, where the routing port is a port on the path of the to-be-sent request packet to the destination node, and determining a route of each to-be-sent request packet If the ports are the same, the to-be-sent request packets with the same routing port are merged to obtain a merge request packet, and the merge request packet is sent to the routing port. The routing port receives the merge request After the packet is requested, the unpacking operation is performed, the source node identifier in the common header of the merge request packet is extracted, written in the header of each sub-request packet, and then the subsequent forwarding operation is continued. The embodiment of the present invention provides a memory data read/write processing device 14. The memory data read/write processing device 14 may be present in a computing node, or may be a computing node, as shown in FIG. The acquisition unit 141, the determination unit 142, the processing unit 143, and the transmission unit 144 are included.

The obtaining unit 141 is configured to obtain each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet;

The determining unit 1 42 is configured to parse the destination node identifiers of the to-be-sent request packets, and determine whether the destination node identifiers of the to-be-sent request packets are the same;

The processing unit 14 3 is configured to package the at least two to-be-sent request packets having the same destination node identifier as a sub-request packet into a merge request packet, where the merge request packet includes a common header and at least two sub-requests. a packet, where the common packet header includes a merge packet identifier, a sub-request packet number identifier, and a source node identifier, where the merge packet identifier is used to indicate that the request packet is a merge packet of the at least two sub-request packet combinations, The number of sub-request packets is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information;

The sending unit 144 is configured to send the merge request packet to the destination node corresponding to the destination node identifier, where the merge request packet is used to instruct the destination node to read and/or write the memory data of the destination node. operating.

The device for reading and writing the memory data provided by the embodiment of the present invention acquires each to-be-sent request packet, and parses the destination node identifier of each to-be-sent request packet, and determines whether the destination node identifier of each to-be-sent request packet is The same, the at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet, and the merge request packet is finally sent to the destination node corresponding to the destination node identifier. The merge request packet is used to indicate that the destination node performs the memory data of the destination node. Read and / or write operations. The embodiment of the invention solves the problem that the remote data access operation with a small amount of data is wasted more on the data request packet in the prior art, resulting in low transmission efficiency and improving the utilization efficiency of the link.

Optionally, as shown in FIG. 15, the processing unit includes:

The first identifier module 151 is configured to set the merged packet identifier to a predetermined value, where the predetermined value indicates that the request packet is a merged package of the at least two sub-request packet combinations;

The second identifier module 152 is configured to obtain the number of sub-request packets, and write the number into the identifier of the sub-request packet.

Optionally, as shown in FIG. 16, the determining unit includes:

The first obtaining module 161 is configured to obtain a destination node identifier carried in each request packet to be sent;

The first determining module 162 is configured to determine whether the destination node identifiers of the to-be-sent request packets are the same;

The processing unit further includes:

The third identifier module 163 is configured to extract a destination node identifier in the sub-request packet, and place the destination node identifier in a common header of the merge request packet.

Optionally, as shown in FIG. 17, the determining unit includes:

The second obtaining module 171 is configured to obtain a memory address of the destination node carried in each request packet to be sent;

The first search module 172 is configured to search, according to the destination node memory address, the destination node identifier of the destination node memory address in the address attribution table, where the address attribution table is a correspondence between the destination node memory address and the destination node identifier. Relational tables;

The second determining module 173 is configured to determine whether the destination node identifiers of the to-be-sent request packets are the same.

Optionally, the sub-request packet included in the merge request packet obtained by the processing unit includes a write request sub-packet and/or a read request sub-packet, and the write request sub-packet header includes a write that identifies the write request sub-packet. Requesting an identifier, the write request information in the write request sub-package includes a to-be-written target The data of the node and the memory address of the destination node to be written, the read request sub-packet header includes a read request identifier identifying the read request sub-packet, and the read request information in the read request sub-package includes a to-be-read The memory address of the destination node where the data is located, as shown in FIG. 18, the device further includes:

The receiving unit 181 is configured to receive a write request response message sent by the destination node, where the write request response message includes a write request identifier corresponding to the write request sub-package with a successful write request; and/or, a read request response message sent by the destination node, where the read request response message includes a read request identifier corresponding to the read request sub-package with a successful read request, and a read payload, where the payload is the read request The data corresponding to the remote memory address in the subpackage.

Optionally, as shown in FIG. 19, the determining unit further includes:

The second search module 191 is configured to search for a routing port corresponding to the destination node identifier in the routing table according to the destination node identifier corresponding to the to-be-sent request packet, where the routing table includes a route corresponding to the destination node identifier. a port, where the routing port is a port on a path of the to-be-sent request packet to the destination node;

The third determining module 192 is configured to determine whether the routing ports of the to-be-sent request packets are the same.

Correspondingly, the embodiment of the present invention further provides a memory data read/write processing device 20, wherein the memory data read/write processing device 20 may exist in a computing node, or may be a computing node, as shown in FIG. The apparatus includes a receiving unit 201, a determining unit 202, a determining unit 203, an extracting unit 204, and a processing unit 205.

The receiving unit 201 is configured to receive a request packet.

The determining unit 202 is configured to determine whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a combined request packet of at least two sub-request packet combinations, where the sub-request package includes: Requesting a sub-package, and/or writing a request sub-package;

The determining unit 203 is configured to: when the request packet carries the merged package identifier, determine that the request packet is a merge request packet, where the merge request packet includes a common header and a sub-request packet, where The common packet header includes a merging packet identifier, a sub-request packet number identifier, and a source node identifier, where the sub-request packet number identifier is used to indicate the number of sub-request packets carried by the merging request packet, and the sub-request packet Contains subrequest packet header and subrequest information;

The extracting unit 2 04 is configured to extract, according to the number of sub-request packets carried in the merge request packet and each sub-request packet header, the number of the number of the sub-request packet identifiers in the merge request packet Request package

The processing unit 2 05 is configured to process the corresponding sub-request packet according to each of the sub-request packet header and sub-request information.

The apparatus for reading and writing the memory data provided by the embodiment of the present invention determines whether the received request packet carries the merged package identifier, and when the request packet carries the merged package identifier, determining that the request packet is a merge request packet, And then, according to the number of sub-request packets carried in the merge request packet and each sub-request packet header, extracting a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet, and finally according to the Each sub-request packet header and sub-request information is processed, and the corresponding sub-request packet is processed. The embodiment of the invention solves the problem that the remote data access operation with a small amount of data is wasteful on the data request packet in the prior art, which causes the transmission efficiency to be low, and improves the utilization efficiency of the link.

Optionally, as shown in FIG. 21, the processing unit includes:

The first processing module 2 1 1 is configured to: according to the write request identifier in the write request sub-packet header, the to-be-written data in the write request information of the write request sub-packet, and the memory address of the destination node to be written, Writing the write data to the memory address of the destination node;

The second processing module 2 1 2 is configured to: according to the read request identifier in the read request sub-packet header, and the memory address of the destination node where the data to be read in the read request information of the read request sub-package is located, The data to be read is read in the destination node memory address.

Optionally, as shown in FIG. 2, the device further includes:

The identifier writing unit 22 1 is configured to write the source node identifier into a sub-request packet header of each sub-request packet, so that the destination node sends a response message to the source node.

Optionally, the determining unit is further configured to determine, according to each response message to be sent, Whether the source node identifiers are the same;

The processing unit is further configured to package the to-be-sent response message with the same source node identifier as a sub-response message, and the merge response message packet includes a write request sub-response message and/or Or read the request sub-response message;

As shown in FIG. 23, the device further includes:

The sending unit 231 is configured to send the merge response message packet to a source node corresponding to the source node identifier.

Optionally, when the destination node is a routing port, the processing unit is further configured to extract a source node identifier in a common header of the merge request packet, and write the header of each sub-request packet.

For the operation of the above device 14 and device 20, refer to the processing of the above-mentioned read and write processing method of the memory data. The read/write processing device for the memory data shown in FIG. 14 to FIG. 19 can be implemented based on the hardware structure of a computer or other network device. As shown in FIG. 24, the hardware structure of the read/write processing device 24 for the in-memory data includes the memory 241. Transceiver 242, processor 243, and bus 244.

The processor 243, the memory 241, and the transceiver 242 are connected by a bus 244. The memory 241 may be a Read Only Memory (ROM), a static storage device, a dynamic storage device, or a Random Access Memory (RAM). The memory 241 can store instructions of the operating system and other applications as well as application data. The instructions stored in the memory 241 are executed by the processor 243 for execution.

The memory 241 in the embodiment of the present invention can be used to store instructions for implementing the functional units in the embodiments shown in FIGS. 14 to 19.

The transceiver 242 is used to implement communication between the read/write processing device 24 of the memory data and other devices.

In the embodiment of the present invention, the transceiver 242 is configured to send a merge request packet and receive a response message. The processor 243 can use a general-purpose central processing unit (Central Processing) Unit, CPU), processor, Application Specific Integrated Circuit (ASIC), or one or more integrated circuits for executing related programs.

In the embodiment of the present invention, the processor 243 is configured to execute the instructions in the memory 241, and specifically:

Obtaining each to-be-sent request packet, where the to-be-sent request packet includes: a read request packet, and/or a write request packet;

The at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet, and the merge request packet includes a common header and at least two sub-request packets, where the common header is included. The merging packet identifier, the sub-request packet number identifier, and the source node identifier, where the merging packet identifier is used to indicate that the request packet is a combined packet of the at least two sub-request packet combinations, and the sub-request packet number identifier is used by And indicating the number of sub-request packets carried by the merge request packet, where the sub-request packet includes a sub-request packet header and sub-request information; and sending the merge request packet to the destination node identifier through the transceiver 242. a node, the merge request packet is used to instruct the destination node to perform a read and/or write operation on the memory data of the destination node.

The device for reading and writing the memory data provided by the embodiment of the present invention acquires each to-be-sent request packet, and parses the destination node identifier of each to-be-sent request packet, and determines whether the destination node identifier of each to-be-sent request packet is The same, the at least two to-be-sent request packets having the same destination node identifier are respectively packaged as a sub-request packet, and the merge request packet is finally sent to the destination node corresponding to the destination node identifier. The merge request packet is used to instruct the destination node to perform a read and/or write operation on the memory data of the destination node. The embodiment of the invention solves the problem that the remote data access operation with a small amount of data is wasted more on the data request packet in the prior art, resulting in low transmission efficiency and improving the utilization efficiency of the link. The specific processing procedure in the processor 243 can refer to the method embodiment part and the apparatus embodiment part shown in FIG. 14 to FIG. 19, and details are not described herein again.

Bus 244 can include a path for communicating information between various components of the read and write processing device of memory data, such as processor 243, memory 241, and transceiver 242.

The read/write processing device for the memory data shown in FIG. 20 to FIG. 23 described above can be implemented based on the hardware configuration of the computer. As shown in FIG. 25, the hardware structure of the read/write processing device 25 for the in-memory data includes the memory 251 and the transceiver 252. , processor 253 and bus 254.

The processor 253, the memory 251, and the transceiver 252 are communicatively coupled via a bus 254.

The memory 251 may be a read only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 251 can store instructions of the operating system and other applications as well as application data. The instructions stored in the memory 251 are executed by the processor 253 for execution.

The memory 251 in the embodiment of the present invention can be used to store instructions for implementing the functional units in the embodiments shown in Figs. 20 to 23.

The transceiver 252 is used to implement communication between the read and write processing device 25 of the memory data and other devices.

In the embodiment of the present invention, the transceiver 252 is configured to receive a merge request packet and send a response message.

The processor 253 can use a general-purpose central processing unit (CPU), a processor, an application specific integrated circuit (ASIC), or one or more integrated circuits for executing related programs.

In the embodiment of the present invention, the processor 253 is configured to execute the instructions in the memory 251, and specifically:

Receiving a request packet through the transceiver 252;

Determining whether the received request packet carries a merged package identifier, where the merged package identifier is used to indicate that the request packet is a merge request packet of at least two sub-request packet combinations, where the sub-request package includes: a read request sub-package, and/ Or write a request sub-package; When the request packet carries the merged package identifier, the request packet is determined to be a merge request packet, where the merge request packet includes a common header and a sub-request packet, where the common header includes the merge packet identifier and the number of sub-request packets. The identifier of the sub-request packet is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information;

Corresponding sub-request packets are processed according to each of the sub-request packet headers and sub-request information. The apparatus for reading and writing the memory data provided by the embodiment of the present invention determines whether the received request packet carries the merged package identifier, and when the request packet carries the merged package identifier, determining that the request packet is a merge request packet, And then, according to the number of sub-request packets carried in the merge request packet and each sub-request packet header, extracting a sub-request packet corresponding to the number of the sub-request packet identifiers in the merge request packet, and finally according to the Each sub-request packet header and sub-request information is processed, and the corresponding sub-request packet is processed. The embodiment of the invention solves the problem that the remote data access operation with a small amount of data is wasteful on the data request packet in the prior art, which causes the transmission efficiency to be low, and improves the utilization efficiency of the link.

The specific processing procedure in the processor 253 can be referred to the method embodiment part and the apparatus embodiment part shown in FIG. 20 to FIG. 23, and details are not described herein again.

Bus 254 can include a path for communicating information between various components of the memory data read and write processing device, such as processor 253, memory 251, and transceiver 252.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the division of each functional module described above is exemplified. In practical applications, the above-mentioned function assignment can be completed by different functional modules as needed. The internal structure of the device is divided into different functional modules to perform all or part of the functions described above. For the specific working process of the system, the device and the unit described above, refer to the corresponding process in the foregoing method embodiment, and details are not described herein again. In the several embodiments provided by the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be another division manner, for example, multiple units or components may be used. Combined or can be integrated into another system, or some features can be ignored, or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.

The units described as separate components may or may not be physically separate. The components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed over multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may contribute to the prior art or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. , including a number of instructions to make a computer device

(may be a personal computer, a server, or a network device, etc.) or a processor (p roce s s or ) performs all or part of the steps of the method of the various embodiments of the present invention. The foregoing storage medium includes: a USB flash drive, a removable hard disk, a read only memory (ROM, Rea d-Only Memory), a random access memory (RAM, Random Acce ss Memory), a magnetic disk or an optical disk, and the like. The medium of the code.

The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not It is to be understood that those skilled in the art are susceptible to variations and substitutions within the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the appended claims.

Claims

claims

1. A method for reading and writing memory data, which is characterized by including:

Obtain each request packet to be sent, which includes: a read request packet, and/or a write request packet;

Analyze the destination node identification of each request packet to be sent, and determine whether the destination node identification of each request packet to be sent is the same;

The at least two request packets to be sent with the same destination node identification are respectively used as sub-request packets and packaged into a merge request packet. The merge request packet includes a common header and at least two sub-request packets. In the common header Contains a merged package identifier, a sub-request package number identifier and a source node identifier. The merged package identifier is used to indicate that the request packet is a merged package that is a combination of the at least two sub-request packages. The sub-request package number identifier is To indicate the number of sub-request packets carried by the merge request packet, the sub-request packet includes a sub-request packet header and sub-request information;

The merge request packet is sent to the destination node corresponding to the destination node identification, and the merge request packet is used to instruct the destination node to read and/or write the memory data of the destination node.

2. The memory data reading and writing processing method according to claim 1, characterized in that the packaging into a merge request package includes:

Set the merged packet identifier to a predetermined value, and the predetermined value indicates that the request packet is a merged packet that is a combination of the at least two sub-request packets;

Obtain the number of sub-request packets, and write the number into the sub-request packet number identifier.

3. The memory data reading and writing processing method according to claim 1 or 2, characterized in that: parsing the destination node identification of each request packet to be sent, and determining the destination node of each request packet to be sent. Whether the logos are the same includes:

Obtain the destination node identification carried in each request packet to be sent;

Determine whether the destination node identifiers of each request packet to be sent are the same;

The package as a merge request package includes:

Extract the destination node identifier in the sub-request packet, and place the destination node identifier in the in the public header of the merge request packet.

4. The method for reading and writing memory data according to claim 1 or 2, characterized in that: parsing the destination node identification of each request packet to be sent, and determining the destination node of each request packet to be sent. Whether the logos are the same includes:

Obtain the memory address of the destination node carried in each request packet to be sent;

According to the memory address of the destination node, search the destination node identifier to which the memory address of the destination node belongs in the address ownership table. The address ownership table is a correspondence table between the memory address of the destination node and the destination node identifier;

Determine whether the destination node identifiers of each request packet to be sent are the same.

5. The memory data reading and writing processing method according to claim 3 or 4, characterized in that: the sub-request packet includes a write request sub-packet and/or a read request sub-packet, and the write request sub-packet header includes a header identifying the Describe the write request identification of the write request sub-packet. The write request information in the write request sub-package includes the data to be written to the destination node and the memory address of the destination node to be written. The read request sub-packet header includes Identify the read request identifier of the read request sub-packet, and the read request information in the read request sub-packet includes the memory address of the destination node where the data to be read is located. The method further includes: receiving the message sent by the destination node. The write request response message contains the write request identifier corresponding to the write request sub-package for which the write request is successful; and/or,

Receive a read request response message sent by the destination node. The read request response message includes a read request identifier corresponding to the read request subpackage for which the read request is successful, and a read payload, where the payload is the Read the data corresponding to the remote memory address in the request subpackage.

6. The method for reading and writing memory data according to any one of claims 1 to 5, characterized in that: parsing the destination node identification of each request packet to be sent, and judging the request packet to be sent. Whether the destination node identifiers are the same include:

According to the destination node identifier corresponding to each request packet to be sent, search the routing port corresponding to the destination node identifier in the routing table. The routing table contains the routing port corresponding to the destination node identifier, and the routing port is the routing port. The port on the path on which the request packet to be sent reaches the destination node; Determine whether the routing ports of each request packet to be sent are the same.

7. A method for reading and writing memory data, which is characterized by including:

Receive request packet;

Determine whether the received request packet carries a merge packet identifier. The merge packet identifier is used to indicate that the request packet is a merge request packet composed of at least two sub-request packets. The sub-request packet includes: a read request sub-packet, and/ Or write request sub-package;

When the request packet carries a merge packet identifier, it is determined that the request packet is a merge request packet. The merge request packet includes a common header and a sub-request packet. The public header includes a merge packet identifier and the number of sub-request packets. Identifier and source node identifier, the sub-request packet number identifier is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information;

According to the sub-request packet number identifier carried in the merge request packet and the header of each sub-request packet, extract the number of sub-request packets in the merge request packet corresponding to the sub-request packet number identifier; Sub-request packet header and sub-request information, and process the corresponding sub-request packet.

8. The memory data reading and writing processing method according to claim 7, characterized in that, processing the corresponding sub-request packet according to each sub-request packet header and sub-request information includes: According to the write request sub-packet header The write request identifier, the data to be written in the write request information of the write request sub-package and the memory address of the destination node to be written, the data to be written is written into the memory address of the destination node; and/or ,

According to the read request identifier in the header of the read request sub-packet and the memory address of the destination node where the data to be read in the read request information of the read request sub-packet is located, the read request is read in the memory address of the destination node. Read data.

9. The memory data reading and writing processing method according to claim 8, characterized in that, after extracting the number of sub-request packets corresponding to the sub-request packet number identifier in the merge request packet, include:

The source node identifier is written into the sub-request packet header of each sub-request packet, so that the destination node sends a response message to the source node.

10. The method for reading and writing memory data according to claim 9, wherein the destination node sending a response message to the source node includes:

Determine whether the source node identifiers corresponding to each response message to be sent are the same;

The response message to be sent with the same source node identification is used as a sub-response message and packaged into a merged response message package. The merged response message package includes a write request sub-response message and/or a read request sub-response message;

Send the merge response message package to the source node corresponding to the source node identification.

1 1. The memory data reading and writing processing method according to claim 7, characterized in that, when the destination node is a routing port, the number of sub-request packets in the extraction of the merge request packet is After identifying the corresponding number of sub-request packets, it also includes:

Extract the source node identifier from the common header of the merge request package and write it into the header of each sub-request package.

1 2. A device for reading and writing memory data, which is characterized by including:

The acquisition unit is used to acquire each request packet to be sent, which includes: a read request packet, and/or a write request packet;

A judgment unit configured to parse the destination node identifiers of each request packet to be sent, and determine whether the destination node identifiers of each request packet to be sent are the same;

A processing unit configured to package the at least two request packets to be sent with the same destination node identification as sub-request packets into a merge request packet, where the merge request packet includes a common header and at least two sub-request packets, The public header includes a merged packet identifier, a sub-request packet number identifier and a source node identifier. The merged packet identifier is used to indicate that the request packet is a merged packet that is a combination of the at least two sub-request packets. The sub-request The packet number identifier is used to indicate the number of sub-request packets carried by the merge request packet, and the sub-request packet includes a sub-request packet header and sub-request information;

A sending unit, configured to send the merge request packet to the destination node corresponding to the destination node identification, and the merge request packet is used to instruct the destination node to read and/or write the memory data of the destination node. .

13. The device for reading and writing memory data according to claim 12, characterized in that the processing unit includes:

A first identification module, configured to set the merged packet ID to a predetermined value, the predetermined value indicating that the request packet is a merged packet that is a combination of the at least two sub-request packets;

The second identification module is used to obtain the number of sub-request packets, and write the number into the sub-request packet number identification.

14. The memory data reading and writing processing device according to claim 12 or 13, characterized in that,

The judgment unit includes:

The first acquisition module is used to obtain the destination node identification carried in each request packet to be sent; the first judgment module is used to determine whether the destination node identification of each request packet to be sent is the same;

The processing unit also includes:

The third identification module is used to extract the destination node identifier in the sub-request packet, and place the destination node identifier in the common header of the merge request packet.

15. The memory data reading and writing processing device according to claim 12 or 13, characterized in that the judgment unit includes:

The second acquisition module is used to obtain the memory address of the destination node carried in each request packet to be sent;

The first search module is used to search the destination node identifier to which the destination node memory address belongs in the address ownership table according to the destination node memory address. The address ownership table is the corresponding relationship between the destination node memory address and the destination node identifier. surface;

The second judgment module is used to judge whether the destination node identifiers of each request packet to be sent are the same.

16. The device for reading and writing memory data according to claim 14 or 15, characterized in that, the sub-request packet included in the merge request packet obtained by the processing unit includes a write request sub-packet and/or Or a read request sub-packet, and the write request sub-packet header includes a write request identifying the write request sub-packet. Find the identification. The write request information in the write request sub-package includes the data to be written to the destination node and the memory address of the destination node to be written. The read request sub-packet header includes identification of the read request sub-packet. The read request identifier, the read request information in the read request subpackage includes the memory address of the destination node where the data to be read is located, and the device further includes:

A receiving unit, configured to receive a write request response message sent by the destination node, where the write request response message contains a write request identifier corresponding to the write request subpackage for which the write request is successful; and/or, configured to receive the The read request response message sent by the destination node, the read request response message includes the read request identifier corresponding to the read request sub-packet for which the read request is successful, and the read payload, where the payload is the read request sub-packet. The data corresponding to the remote memory address in the package.

17. The memory data reading and writing processing device according to any one of claims 12-16, characterized in that the judgment unit further includes:

The second search module is configured to search for the routing port corresponding to the destination node identifier in the routing table according to the destination node identifier corresponding to each request packet to be sent. The routing table includes the routing port corresponding to the destination node identifier. , the routing port is a port on the path of the request packet to be sent to the destination node;

The third judgment module is used to judge whether the routing ports of each request packet to be sent are the same.

18. A device for reading and writing memory data, which is characterized by including:

Receiving unit, used to receive request packets;

A judgment unit, used to judge whether the received request packet carries a merge packet identifier, where the merge packet identifier is used to indicate that the request packet is a merge request packet composed of at least two sub-request packets, and the sub-request packet includes: Read request subpackage, and/or write request subpackage;

Determining unit, configured to determine that the request packet is a merge request packet when the request packet carries a merge packet identifier, the merge request packet includes a common header and a sub-request packet, and the public header includes the merge packet identifier, A sub-request packet number identifier and a source node identifier. The sub-request packet number identifier is used to indicate the number of sub-request packets carried by the merge request packet. The sub-request packet includes a sub-request packet header and sub-request information;

An extraction unit, configured to extract information based on the number of sub-request packets carried in the merge request packet and the number of each sub-request packet. Sub-request packet header, extract the number of sub-request packets corresponding to the sub-request packet number identifier in the merge request packet;

A processing unit, configured to process the corresponding sub-request packet according to the header and sub-request information of each sub-request packet.

19. The device for reading and writing memory data according to claim 18, characterized in that the processing unit includes:

The first processing module is configured to write the to-be-written data according to the write request identifier in the write request sub-packet header, the data to be written in the write request information of the write request sub-package, and the memory address of the destination node to be written. The incoming data is written into the memory address of the destination node;

The second processing module is configured to, based on the read request identification in the read request sub-packet header and the memory address of the destination node where the data to be read in the read request information of the read request sub-packet is located, in the memory of the destination node The data to be read is read from the address.

20. The device for reading and writing memory data according to claim 19, characterized in that the device further includes:

An identity writing unit is configured to write the source node identity into the sub-request packet header of each sub-request packet, so that the destination node sends a response message to the source node.

21. The device for reading and writing memory data according to claim 20, characterized in that the judgment unit is also used to judge whether the source node identifiers corresponding to each response message to be sent are the same;

The processing unit is also configured to package the response message to be sent with the same source node identification as a sub-response message into a merged response message package, where the merged response message package includes a write request sub-response message and/ Or read request sub-response message, ■

The device also includes:

A sending unit, configured to send the merge response message packet to the source node corresponding to the source node identification.

22. The device for reading and writing memory data according to claim 18, characterized in that when the destination node is a routing port, the processing unit is also used to extract the merge request The source node identifier in the public header of the package is written into the header of each sub-request package.