CN109828843A - Method, system and the electronic equipment that data are transmitted between a kind of calculate node - Google Patents
Method, system and the electronic equipment that data are transmitted between a kind of calculate node Download PDFInfo
- Publication number
- CN109828843A CN109828843A CN201910091960.6A CN201910091960A CN109828843A CN 109828843 A CN109828843 A CN 109828843A CN 201910091960 A CN201910091960 A CN 201910091960A CN 109828843 A CN109828843 A CN 109828843A
- Authority
- CN
- China
- Prior art keywords
- data
- transmitted
- dma engine
- computing nodes
- target computing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000005540 biological transmission Effects 0.000 claims abstract description 73
- 238000012546 transfer Methods 0.000 claims abstract description 26
- 239000000284 extract Substances 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Landscapes
- Computer And Data Communications (AREA)
Abstract
This application discloses the methods that data between a kind of calculate node are transmitted, comprising: receives the data transfer task of input；Transmission information is determined according to data transfer task；It calls the local PCIe DMA engine to read data to be transmitted from memory, and data to be transmitted is transmitted to long-range DMA engine；Call long-range DMA engine that data to be transmitted is transmitted to target computing nodes.Technical solution provided herein, by increasing the local PCIe DMA engine component and long-range DMA engine component in calculate node, so that data to be transmitted is by DMA transfer direct in memory into the local PCIe DMA engine, data transmit the load to CPU and the demand to caching between reducing calculate node, it is next data cached not need additional opening space, reduces the demand to caching.The application additionally provides system, equipment and the computer readable storage medium that data between a kind of calculate node are transmitted simultaneously, has above-mentioned beneficial effect.
Description
Technical field
This application involves field of data transmission, in particular to method, the system, electronics of data transmission between a kind of calculate node
Equipment and computer readable storage medium.
Background technique
Core equipment of the server as internet provides a user various network services.With the development of internet, no
By being that user requests scale and complexity, the data volume also related to increase at geometric index, at this stage to the place of server
Reason ability proposes superelevation requirement, a large number of users request is not only responded in a very short period of time, often involved in single request
Data volume is also huge.Under this demand, server cluster, data center etc. by multiple server centereds get up together into
The processing system of row processing and operation is come into being, and is like only one server in client.
Cluster or data center can use multiple servers and carry out parallel computation to obtain very high calculating speed, often
A server is referred to as a calculate node, and is equipped with a certain amount of memory.When same service is completed in each calculate node division of labor, need
Each calculate node is in communication with each other collaboration, and the data in node are also required to shared transmission each other.
Referring to FIG. 1, the schematic diagram of Fig. 1 data transport frame between a kind of calculate node in the prior art.Such as Fig. 1 institute
Show, by traditional network interface card interconnection between prior art calculate node, multiple nodes form a local area network.Each node quilt
It is distributed into different IP address, the application program for running on each node transfers data to destination node by ICP/IP protocol,
Or the order of request data is sent to another node.
Prior art realizes server cluster or data center, is primarily present following problems:
1. CPU needs continual participation transmission process during the entire process of data transmitting, lead to accounting for for cpu resource
With transmitting task computation and data cannot carry out parallel well, fight for the generation of CPU operational capability.
2. operating system nucleus, during layer-by-layer packaged data, data successively replicate transmitting, opening for memory is increased
Pin.
Therefore, how to reduce data load of the transmission to CPU and the demand to caching between calculate node is this field skill
The current technical issues that need to address of art personnel.
Summary of the invention
The purpose of the application is to provide method, system, electronic equipment and the computer that data between a kind of calculate node are transmitted
Readable storage medium storing program for executing, for reducing data load of the transmission to CPU and the demand to caching between calculate node.
In order to solve the above technical problems, the application provides the method that data between a kind of calculate node are transmitted, this method comprises:
Receive the data transfer task of input；
Transmission information is determined according to the data transfer task；Wherein, the transport packet includes the object of data to be transmitted
Manage address and target computing nodes；
It calls the local PCIe DMA engine to read the data to be transmitted from memory, and the data to be transmitted is transmitted
To long-range DMA engine；
Call the long-range DMA engine that the data to be transmitted is transmitted to the target computing nodes.
Optionally, the local PCIe DMA engine is called to read data to be transmitted from memory, comprising:
The starting PCIE-DMA transmission operation of the local PCIe DMA engine is called, according to the physics of the data to be transmitted
The data to be transmitted is read from the memory according to preset granularity in address.
Optionally, it calls the long-range DMA engine that the data to be transmitted is transmitted to the target computing nodes, wraps
It includes:
It calls the long-range DMA engine to configure network interface card according to the target computing nodes, and the data to be transmitted is led to
It crosses the network interface card and is transmitted to exchange network；
After the network interface card of the target computing nodes receives the data to be transmitted by the exchange network, it is transferred to institute
State the long-range DMA engine of target computing nodes；
The long-range DMA engine of the target computing nodes extracts destination address information, and by the destination address information
The local the PCIe DMA engine for passing to the target computing nodes is transmitted to the data to be transmitted；
The local the PCIe DMA engine of the target computing nodes initiates dma operation according to the destination address information, by institute
State CPU and/or memory that data to be transmitted passes to the target computing nodes.
Optionally, call the long-range DMA engine by the data to be transmitted be transmitted to the target computing nodes it
Afterwards, further includes:
Present data transmission number is recorded, and generates data transmission log.
The application also provides the system that data between a kind of calculate node are transmitted, which includes:
Receiving module, data transfer task for receiving input；
Determining module, for determining transmission information according to the data transfer task；Wherein, the transport packet include to
Transmit the physical address and target computing nodes of data；
First calling module, for calling the local PCIe DMA engine to read the data to be transmitted from memory, and by institute
It states data to be transmitted and is transmitted to long-range DMA engine；
Second calling module, by calling the long-range DMA engine that the data to be transmitted is transmitted to based on the target
Operator node.
Optionally, first calling module includes:
First calls submodule, for calling the starting PCIE-DMA transmission operation of the local PCIe DMA engine, according to institute
The physical address for stating data to be transmitted reads the data to be transmitted according to preset granularity from the memory.
Optionally, second calling module includes:
Second calls submodule, configures network interface card according to the target computing nodes for the long-range DMA engine, and by institute
It states data to be transmitted and exchange network is transmitted to by the network interface card；
First transmission submodule, for when the target computing nodes network interface card by the exchange network receive described in
After transmitting data, it is transferred to the long-range DMA engine of the target computing nodes；
Second transmission submodule, the long-range DMA engine for the target computing nodes extract destination address information, and
The destination address information and the data to be transmitted are transmitted to and passes to the local the PCIe DMA of the target computing nodes and draws
It holds up；
Third transmits submodule, and the local the PCIe DMA engine for the target computing nodes is according to the destination address
Information initiates dma operation, and the data to be transmitted is passed to the CPU and/or memory of the target computing nodes.
Optionally, further includes:
Logging modle for recording present data transmission number, and generates data transmission log.
The application also provides a kind of electronic equipment, which includes:
Memory, for storing computer program；
Processor, data are transmitted between the calculate node as described in any of the above-described is realized when for executing the computer program
Method the step of.
The application also provides a kind of computer readable storage medium, and calculating is stored on the computer readable storage medium
Machine program realizes the side of data transmission between the calculate node as described in any of the above-described when the computer program is executed by processor
The step of method.
The method that data are transmitted between calculate node provided herein, comprising: receive the data transfer task of input；According to
Data transfer task determines transmission information；Wherein, transport packet includes the physical address and target computing nodes of data to be transmitted；
It calls the local PCIe DMA engine to read data to be transmitted from memory, and data to be transmitted is transmitted to long-range DMA engine；It adjusts
Data to be transmitted is transmitted to target computing nodes with long-range DMA engine.
Technical solution provided herein, by increasing the local PCIe DMA engine component and remote in calculate node
Journey DMA engine component, so that first calling the local PCIe DMA engine to read from memory to be passed when receiving data transfer task
Transmission of data, and data to be transmitted is transmitted to long-range DMA engine；Then long-range DMA engine is called to be transmitted to data to be transmitted
Target computing nodes；Data to be transmitted, into the local PCIe DMA engine, is not required to by DMA transfer direct in memory in transmission process
It wants CPU to participate in, cpu load can be effectively reduced；Meanwhile data to be transmitted need not be copied mutually in memory, according to transmission network
The loading condition of network is directly fetched in memory according to being transmitted, and is not needed additional opening space and is come data cached, reduces pair
The demand of caching.The application additionally provides system, electronic equipment and the computer that data between a kind of calculate node are transmitted simultaneously can
Storage medium is read, there is above-mentioned beneficial effect, details are not described herein.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
The schematic diagram of Fig. 1 data transport frame between a kind of calculate node in the prior art；
Fig. 2 is the schematic diagram of data transmission stream journey in the prior art；
The flow chart of Fig. 3 method that data are transmitted between a kind of calculate node provided by the embodiment of the present application；
The schematic diagram of Fig. 4 data transport frame between a kind of calculate node provided by the embodiment of the present application；
A kind of practical manifestation mode of S304 in Fig. 5 method that data are transmitted between a kind of calculate node provided by Fig. 3
Flow chart；
The structure chart of Fig. 6 system that data are transmitted between a kind of calculate node provided by the embodiment of the present application；
The structure chart of Fig. 7 system that data are transmitted between another kind calculate node provided by the embodiment of the present application；
Fig. 8 is the structure chart of a kind of electronic equipment provided by the embodiment of the present application.
Specific embodiment
The core of the application is to provide method, system, electronic equipment and the computer that data between a kind of calculate node are transmitted
Readable storage medium storing program for executing, for reducing data load of the transmission to CPU and the demand to caching between calculate node.
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Referring to FIG. 3, the process of Fig. 3 method that data are transmitted between a kind of calculate node provided by the embodiment of the present application
Figure.
It specifically comprises the following steps:
S301: the data transfer task of input is received；
Referring to FIG. 2, Fig. 2 is the schematic diagram of data transmission stream journey in the prior art.As shown in Fig. 2, application program is initiated
After data-transfer requests, operating system nucleus will be according to the layer-by-layer packaged data of network ICP/IP protocol, and copy between the layers
Data after packing are finally passed to network interface card and carry out network transmission by data.Meanwhile when receiving data, operating system nucleus
Layer-by-layer unpacked data is needed, and completes other business required by ICP/IP protocol (data inspection, re-transmission etc.).
In the whole process, CPU needs continual participation transmission process, leads to the occupancy of cpu resource, makes task computation
It cannot carry out parallel well with data transmission, the generation of CPU operational capability is fought for.And operating system nucleus is packaged successively
During data, data successively replicate transmitting, increase the expense of memory.Based on this, this application provides a kind of calculating sections
The method that data are transmitted between point, for solving the above problems.
S302: transmission information is determined according to data transfer task；
Transport packet mentioned herein includes the physical address and target computing nodes of data to be transmitted；
Optionally, which can also include the information such as length, the destination address of data to be transmitted.
S303: it calls the local PCIe DMA engine to read data to be transmitted from memory, and data to be transmitted is transmitted to far
Journey DMA engine；
As shown in figure 4, the signal of Fig. 4 data transport frame between a kind of calculate node provided by the embodiment of the present application
Figure must participate in transmission process by CPU based on the transmission of internodal data in the prior art, cause cpu resource occupied, the application
Innovative increases the local PCIe DMA engine component and long-range DMA engine component in calculate node, passes through each calculate node
Between the local PCIe DMA engine component and the long-range complete paired data of DMA engine component transmission；
DMA (Direct Memory Access, direct memory access) mentioned herein allows the hardware of friction speed to fill
It sets to link up, a large amount of interrupt loads without depending on CPU, DMA transfer is to copy to data from an address space
Another address space.When CPU initialize this transmission movement, transmission movement be in itself carried out by dma controller with it is complete
At DMA transfer mode directly controls transmission without CPU, also retains scene and restoring scene like that without Interrupt Processing Model
Process, by hardware be RAM and I/O equipment opens up the access for directly transmitting data, greatly improves the efficiency of CPU；
Optionally, mentioned herein, call the local PCIe DMA engine to read data to be transmitted from memory, and will be to be passed
Transmission of data is transmitted to long-range DMA engine, is specifically as follows:
Call the local PCIe DMA engine starting PCIE-DMA transmission operation, according to the physical address of data to be transmitted according to
Preset granularity reads data to be transmitted from memory.
S304: call long-range DMA engine that data to be transmitted is transmitted to target computing nodes.
Long-range DMA engine is used to the data to be transmitted that the local PCIe DMA engine is sent being transmitted to target computing nodes, complete
At the transmission process of data between calculate node.
Optionally, after calling long-range DMA engine that data to be transmitted is transmitted to target computing nodes, further includes:
Present data transmission number is recorded, and generates data transmission log.
Based on the above-mentioned technical proposal, the method that data are transmitted between a kind of calculate node provided herein, by counting
The local PCIe DMA engine component and long-range DMA engine component are increased on operator node, so that when receiving data transfer task,
First the local PCIe DMA engine is called to read data to be transmitted from memory, and data to be transmitted is transmitted to long-range DMA engine；
Then call long-range DMA engine that data to be transmitted is transmitted to target computing nodes；Data to be transmitted is passed by DMA direct in memory
It is defeated into the local PCIe DMA engine, do not needed in transmission process CPU participation, cpu load can be effectively reduced；Meanwhile it is to be transmitted
Data need not be copied mutually in memory, according to the loading condition of transmission network, directly be fetched in memory according to being transmitted, no
It needs additional opening space data cached, reduces the demand to caching.
It is directed to the step S304 of an embodiment, wherein described call long-range DMA engine to pass data to be transmitted
Target computing nodes are transported to, is specifically as follows step as shown in Figure 5, is illustrated below with reference to Fig. 5.
Referring to FIG. 5, a kind of reality of Fig. 5 S304 in the method for data transmission between a kind of calculate node provided by Fig. 3
The flow chart of border manifestation mode.
Itself specifically includes the following steps:
S501: it calls long-range DMA engine to configure network interface card according to target computing nodes, and data to be transmitted is passed by network interface card
Transport to exchange network；
S502: after the network interface card of target computing nodes receives data to be transmitted by exchange network, it is transferred to target calculating
The long-range DMA engine of node；
S503: the long-range DMA engine of target computing nodes extracts destination address information, and by destination address information and to
Transmission data are transmitted to the local the PCIe DMA engine for passing to target computing nodes；
S504: the local the PCIe DMA engine of target computing nodes initiates dma operation according to destination address information, will be to be passed
Transmission of data passes to the CPU and/or memory of target computing nodes.
Based on the above-mentioned technical proposal, the process that the embodiment of the present application receives data to be transmitted to target computing nodes carries out
Description by calling long-range DMA engine to configure network interface card according to target computing nodes, and data to be transmitted is transmitted to by network interface card
Exchange network；After the network interface card of target computing nodes receives data to be transmitted by exchange network, it is transferred to target computing nodes
Long-range DMA engine；The long-range DMA engine of target computing nodes extracts destination address information, and by destination address information and
Data to be transmitted is transmitted to the local the PCIe DMA engine for passing to target computing nodes；The local the PCIe DMA of target computing nodes
Engine initiates dma operation according to destination address information, and data to be transmitted is passed to the CPU and/or memory of target computing nodes,
Complete the process that entire target computing nodes receive data to be transmitted.
Referring to FIG. 6, the structure of Fig. 6 system that data are transmitted between a kind of calculate node provided by the embodiment of the present application
Figure.
The system may include:
Receiving module 100, data transfer task for receiving input；
Determining module 200, for determining transmission information according to data transfer task；Wherein, transport packet includes to be transmitted
The physical address and target computing nodes of data；
First calling module 300, for calling the local PCIe DMA engine to read data to be transmitted from memory, and will be to
Transmission data are transmitted to long-range DMA engine；
Second calling module 400, for calling long-range DMA engine that data to be transmitted is transmitted to target computing nodes.
Referring to FIG. 7, the knot of Fig. 7 system that data are transmitted between another kind calculate node provided by the embodiment of the present application
Composition.
First calling module 300 can specifically include:
First calls submodule, for calling the starting PCIE-DMA transmission operation of the local PCIe DMA engine, according to be transmitted
The physical address of data reads data to be transmitted according to preset granularity from memory.
Second calling module 400 can specifically include:
Second calls submodule, configures network interface card according to target computing nodes for long-range DMA engine, and by data to be transmitted
Exchange network is transmitted to by network interface card；
First transmission submodule, for after the network interface card of target computing nodes receives data to be transmitted by exchange network,
It is transferred to the long-range DMA engine of target computing nodes；
Second transmission submodule, the long-range DMA engine for target computing nodes extract destination address information, and by mesh
Address information and data to be transmitted be transmitted to the local the PCIe DMA engine for passing to target computing nodes；
Third transmits submodule, and the local the PCIe DMA engine for target computing nodes is initiated according to destination address information
Data to be transmitted is passed to the CPU and/or memory of target computing nodes by dma operation.
The system can also include:
Logging modle for recording present data transmission number, and generates data transmission log.
Each component part in system above can be applied in embodiment below:
Receiving module receives the data transfer task of input；Determining module determines transmission information according to data transfer task；
First calls submodule to call the starting PCIE-DMA transmission operation of the local PCIe DMA engine, physically according to data to be transmitted
Data to be transmitted is read from memory according to preset granularity in location；
Second calls the long-range DMA engine of submodule to configure network interface card according to target computing nodes, and data to be transmitted is passed through
Network interface card is transmitted to exchange network；First transmission submodule receives number to be transmitted by exchange network when the network interface card of target computing nodes
According to rear, it is transferred to the long-range DMA engine of target computing nodes；The long-range DMA engine of second transmission submodule target computing nodes
Destination address information is extracted, and destination address information and data to be transmitted are transmitted to the PCIe for passing to target computing nodes
Local DMA engine；The local the PCIe DMA engine that third transmits submodule target computing nodes is initiated according to destination address information
Data to be transmitted is passed to the CPU and/or memory of target computing nodes by dma operation；
Logging modle records present data transmission number, and generates data transmission log.
Referring to FIG. 8, Fig. 8 is the structure chart of a kind of electronic equipment provided by the embodiment of the present application.
The electronic equipment 800 can generate bigger difference because configuration or performance are different, may include one or one
The above processor (central processing units, CPU) 822 (for example, one or more processors) and storage
Device 832, one or more storage application programs 842 or data 844 storage medium 830 (such as one or more
Mass memory unit).Wherein, memory 832 and storage medium 830 can be of short duration storage or persistent storage.It is stored in storage
The program of medium 830 may include one or more modules (diagram does not mark), and each module may include in device
Series of instructions operation.Further, central processing unit 822 can be set to communicate with storage medium 830, set in electronics
The series of instructions operation in storage medium 830 is executed on standby 800.
Electronic equipment 800 can also include one or more power supplys 828, one or more wired or wireless nets
Network interface 850, one or more input/output interfaces 858, and/or, one or more operating systems 841, such as
Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
The step in method that data are transmitted between calculate node described in above-mentioned Fig. 3 to Fig. 5 is based on by electronic equipment should
Structure shown in Fig. 8 is realized.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and module, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed device, device and method, it can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the division of module,
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple module or components can be with
In conjunction with or be desirably integrated into another system, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING of device or module or
Communication connection can be electrical property, mechanical or other forms.
Module may or may not be physically separated as illustrated by the separation member, show as module
Component may or may not be physical module, it can and it is in one place, or may be distributed over multiple networks
In module.Some or all of the modules therein can be selected to achieve the purpose of the solution of this embodiment according to the actual needs.
It, can also be in addition, can integrate in a processing module in each functional module in each embodiment of the application
It is that modules physically exist alone, can also be integrated in two or more modules in a module.Above-mentioned integrated mould
Block both can take the form of hardware realization, can also be realized in the form of software function module.
If integrated module is realized and when sold or used as an independent product in the form of software function module, can
To be stored in a computer readable storage medium.Based on this understanding, the technical solution of the application substantially or
Say that all or part of the part that contributes to existing technology or the technical solution can embody in the form of software products
Out, which is stored in a storage medium, including some instructions are used so that a computer equipment
The whole of (can be personal computer, funcall device or the network equipment etc.) execution each embodiment method of the application
Or part steps.And storage medium above-mentioned include: USB flash disk, mobile hard disk, read-only memory (Read-Only Memory,
ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. are various can store program
The medium of code.
Method, system, equipment and the computer transmitted above between data a kind of calculate node provided herein can
Storage medium is read to be described in detail.Specific case used herein explains the principle and embodiment of the application
It states, the description of the example is only used to help understand the method for the present application and its core ideas.It should be pointed out that for this skill
For the those of ordinary skill in art field, under the premise of not departing from the application principle, several change can also be carried out to the application
Into and modification, these improvement and modification also fall into the protection scope of the claim of this application.
It should also be noted that, in the present specification, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or equipment for including element.
Claims (10)
1. a kind of method that data are transmitted between calculate node characterized by comprising
Receive the data transfer task of input；
Transmission information is determined according to the data transfer task；Wherein, the transport packet includes data to be transmitted physically
Location and target computing nodes；
It calls the local PCIe DMA engine to read the data to be transmitted from memory, and the data to be transmitted is transmitted to far
Journey DMA engine；
Call the long-range DMA engine that the data to be transmitted is transmitted to the target computing nodes.
2. the method according to claim 1, wherein it is to be passed to call the local PCIe DMA engine to read from memory
Transmission of data, comprising:
The starting PCIE-DMA transmission operation of the local PCIe DMA engine is called, according to the physical address of the data to be transmitted
The data to be transmitted is read from the memory according to preset granularity.
3. the method according to claim 1, wherein calling the long-range DMA engine by the data to be transmitted
It is transmitted to the target computing nodes, comprising:
It calls the long-range DMA engine to configure network interface card according to the target computing nodes, and the data to be transmitted is passed through into institute
It states network interface card and is transmitted to exchange network；
After the network interface card of the target computing nodes receives the data to be transmitted by the exchange network, it is transferred to the mesh
Mark the long-range DMA engine of calculate node；
The long-range DMA engine of the target computing nodes extracts destination address information, and by the destination address information and institute
It states data to be transmitted and is transmitted to the local the PCIe DMA engine for passing to the target computing nodes；
The local the PCIe DMA engine of the target computing nodes initiates dma operation according to the destination address information, will it is described to
Transmission data pass to the CPU and/or memory of the target computing nodes.
4. the method according to claim 1, wherein call the long-range DMA engine by the number to be transmitted
According to being transmitted to after the target computing nodes, further includes:
Present data transmission number is recorded, and generates data transmission log.
5. the system that data are transmitted between a kind of calculate node characterized by comprising
Receiving module, data transfer task for receiving input；
Determining module, for determining transmission information according to the data transfer task；Wherein, the transport packet includes to be transmitted
The physical address and target computing nodes of data；
First calling module, for calling the local PCIe DMA engine to read the data to be transmitted from memory, and will it is described to
Transmission data are transmitted to long-range DMA engine；
Second calling module calculates section for calling the long-range DMA engine that the data to be transmitted is transmitted to the target
Point.
6. system according to claim 5, which is characterized in that first calling module includes:
First call submodule, for call the local PCIe DMA engine starting PCIE-DMA transmission operation, according to it is described to
The physical address of transmission data reads the data to be transmitted according to preset granularity from the memory.
7. system according to claim 5, which is characterized in that second calling module includes:
Second calls submodule, configures network interface card according to the target computing nodes for the long-range DMA engine, and will it is described to
It transmits data and exchange network is transmitted to by the network interface card；
First transmission submodule, is received for the network interface card when the target computing nodes by the exchange network described to be transmitted
After data, it is transferred to the long-range DMA engine of the target computing nodes；
Second transmission submodule, the long-range DMA engine for the target computing nodes extract destination address information, and by institute
It states destination address information and the data to be transmitted is transmitted to the local the PCIe DMA engine for passing to the target computing nodes；
Third transmits submodule, and the local the PCIe DMA engine for the target computing nodes is according to the destination address information
Dma operation is initiated, the data to be transmitted is passed to the CPU and/or memory of the target computing nodes.
8. system according to claim 5, which is characterized in that further include:
Logging modle for recording present data transmission number, and generates data transmission log.
9. a kind of electronic equipment characterized by comprising
Memory, for storing computer program；
Processor realizes data between the calculate node as described in any one of Claims 1-4 when for executing the computer program
The step of method of transmission.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program, data pass between the calculate node as described in any one of Claims 1-4 is realized when the computer program is executed by processor
The step of defeated method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910091960.6A CN109828843A (en) | 2019-01-30 | 2019-01-30 | Method, system and the electronic equipment that data are transmitted between a kind of calculate node |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910091960.6A CN109828843A (en) | 2019-01-30 | 2019-01-30 | Method, system and the electronic equipment that data are transmitted between a kind of calculate node |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109828843A true CN109828843A (en) | 2019-05-31 |
Family
ID=66863021
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910091960.6A Withdrawn CN109828843A (en) | 2019-01-30 | 2019-01-30 | Method, system and the electronic equipment that data are transmitted between a kind of calculate node |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109828843A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112446474A (en) * | 2019-08-31 | 2021-03-05 | 安徽寒武纪信息科技有限公司 | Chip, multichip system, electronic equipment and data transmission method |
| CN112506828A (en) * | 2020-12-18 | 2021-03-16 | 展讯半导体(成都)有限公司 | Transmission configuration method and device for direct memory access |
| CN113098955A (en) * | 2021-03-30 | 2021-07-09 | 山东英信计算机技术有限公司 | Data transmission method, device, equipment and computer readable storage medium |
| WO2022156376A1 (en) * | 2021-01-19 | 2022-07-28 | 苏州浪潮智能科技有限公司 | Method, system and device for prefetching target address, and medium |
| CN116401189A (en) * | 2023-06-07 | 2023-07-07 | 湖北芯擎科技有限公司 | Multi-chip communication method, equipment and storage medium based on PCIE controller |
| WO2024093112A1 (en) * | 2022-10-31 | 2024-05-10 | 广东浪潮智慧计算技术有限公司 | Computing engine communication method and apparatus, electronic device, and storage medium |
-
2019
- 2019-01-30 CN CN201910091960.6A patent/CN109828843A/en not_active Withdrawn
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112446474A (en) * | 2019-08-31 | 2021-03-05 | 安徽寒武纪信息科技有限公司 | Chip, multichip system, electronic equipment and data transmission method |
| CN112446474B (en) * | 2019-08-31 | 2022-11-22 | 安徽寒武纪信息科技有限公司 | Chip, multichip system, electronic equipment and data transmission method |
| CN112506828A (en) * | 2020-12-18 | 2021-03-16 | 展讯半导体(成都)有限公司 | Transmission configuration method and device for direct memory access |
| CN112506828B (en) * | 2020-12-18 | 2024-05-17 | 展讯半导体(成都)有限公司 | Transmission configuration method and device for direct memory access |
| WO2022156376A1 (en) * | 2021-01-19 | 2022-07-28 | 苏州浪潮智能科技有限公司 | Method, system and device for prefetching target address, and medium |
| CN113098955A (en) * | 2021-03-30 | 2021-07-09 | 山东英信计算机技术有限公司 | Data transmission method, device, equipment and computer readable storage medium |
| WO2024093112A1 (en) * | 2022-10-31 | 2024-05-10 | 广东浪潮智慧计算技术有限公司 | Computing engine communication method and apparatus, electronic device, and storage medium |
| CN116401189A (en) * | 2023-06-07 | 2023-07-07 | 湖北芯擎科技有限公司 | Multi-chip communication method, equipment and storage medium based on PCIE controller |
| CN116401189B (en) * | 2023-06-07 | 2023-10-10 | 湖北芯擎科技有限公司 | Multi-chip communication method, equipment and storage medium based on PCIE controller |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109828843A (en) | Method, system and the electronic equipment that data are transmitted between a kind of calculate node | |
| CN106549878B (en) | Service distribution method and device | |
| CN106375142B (en) | The test method and device of application program | |
| CN110300115A (en) | A kind of data transmission method and relevant apparatus based on multichannel | |
| CN102984189B (en) | Wireless network and implementation method and terminal thereof | |
| US20150363340A1 (en) | Providing multiple synchronous serial console sessions using data buffering | |
| US11750418B2 (en) | Cross network bridging | |
| CN107104824A (en) | A kind of network topology determines method and apparatus | |
| CN106484322A (en) | A kind of method of carry file system, device and equipment | |
| CN106980582A (en) | Data processing method and device | |
| CN107819891A (en) | Data processing method, device, computer equipment and storage medium | |
| CN105472291A (en) | Digital video recorder with multiprocessor cluster and realization method of digital video recorder | |
| CN106453625A (en) | Information synchronization method and high-availability cluster system | |
| CN106921578A (en) | The generation method and device of a kind of forwarding-table item | |
| CN110061892A (en) | Method, system, storage medium and the electronic equipment of real-time display Network status | |
| CN112468589A (en) | Data distribution method and device, computer equipment and storage medium | |
| CN109417507A (en) | The message of section retards accesses | |
| Bergstrom et al. | The distributed open network emulator: Using relativistic time for distributed scalable simulation | |
| CN111786887A (en) | Data forwarding method, apparatus, computing device, and medium executed by control device | |
| CN102845042B (en) | The aggregation of bandwidth system and method for the multiple movable physical interface of a kind of application layer | |
| CN111431921A (en) | Configuration synchronization method | |
| CN105991633B (en) | Session log processing method and processing device | |
| CN108011801A (en) | Method, unit and the system of data transfer | |
| CN110417573A (en) | A kind of method and system of data transmission | |
| US11941445B2 (en) | RLC channel management for low memory 5G devices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190531 |