CN115827285A - Cross-platform communication method, system, device, equipment and medium - Google Patents
Cross-platform communication method, system, device, equipment and medium Download PDFInfo
- Publication number
- CN115827285A CN115827285A CN202310154595.5A CN202310154595A CN115827285A CN 115827285 A CN115827285 A CN 115827285A CN 202310154595 A CN202310154595 A CN 202310154595A CN 115827285 A CN115827285 A CN 115827285A
- Authority
- CN
- China
- Prior art keywords
- layer
- cross
- operation request
- platform communication
- upper layer
- 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.)
- Granted
Links
- 238000004891 communication Methods 0.000 title claims abstract description 261
- 238000000034 method Methods 0.000 title claims abstract description 173
- 230000001133 acceleration Effects 0.000 claims abstract description 96
- 230000005540 biological transmission Effects 0.000 claims abstract description 90
- 230000007246 mechanism Effects 0.000 claims abstract description 40
- 238000004806 packaging method and process Methods 0.000 claims abstract description 14
- 230000008569 process Effects 0.000 claims description 44
- 230000006870 function Effects 0.000 claims description 41
- 238000004590 computer program Methods 0.000 claims description 10
- 230000004044 response Effects 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 8
- 238000012546 transfer Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 230000003993 interaction Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007274 generation of a signal involved in cell-cell signaling Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Computer And Data Communications (AREA)
Abstract
The application relates to the technical field of communication, and discloses a cross-platform communication method, a system, a device, equipment and a medium, which comprises the following steps: a first communication layer of a control server provides a cross-platform communication interface for a first upper layer of the server; the cross-platform communication interface is physically paired by a first communication layer packaging a PCIE driving interface of the transmission layer to obtain the PCIE driving interface; controlling a first upper layer to call a cross-platform communication interface so as to send an operation request to an acceleration card end based on a direct memory access mechanism of a physical transmission layer; and controlling a second communication layer of the acceleration card end to acquire the operation request from the direct memory access controller of the physical transmission layer and send the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer. According to the method and the device, data transparent transmission is realized for upper developers, any details of cross-platform transmission do not need to be concerned, meanwhile, data transmission time can be greatly saved, and communication efficiency is improved.
Description
Technical Field
The present invention relates to the field of communications technologies, and in particular, to a cross-platform communication method, system, apparatus, device, and medium.
Background
The artificial intelligent accelerator card is a board card specially used for accelerating a deep learning algorithm, and the accelerator card is connected with a server through a PCIE slot to accelerate a neural network. Generally, the accelerator card and the server usually adopt Linux as an operating system, the Linux system completes specific logic functions through threads or processes, communication between each process and each thread can be realized through a signal mode, a message queue mode, a shared memory mode and the like, and the key point is how to realize multi-user cross-platform communication under a Linux system.
At present, there is no open framework or scheme for artificial intelligence acceleration board card cross-platform communication, and there are two related schemes: (1) The data transmission method comprises the steps that the data are divided according to the internal modules of the board card, each module development engineer is respectively responsible for cross-platform communication, cross-platform data transmission is carried out through a data format appointed by the module development engineer, and data transmission is generally completed through an interface provided by PCIE driving and an interruption or traversal memory mode. In this way, developers need to know the PCIE technology to a certain extent, the technical requirements of the developers in the middle layer on the bottom layer driver are increased, and when multiple users or multiple modules are in a data acquisition state at the same time, excessive processes or threads are caused to access a certain fixed memory area at the same time to determine whether there is data required by the developers, so that system resources are not in reasonable competition, and performance is reduced; (2) Data acquisition and distribution of the board card end are managed in a unified mode by adding a forwarding process and a management process to the board card end, after data are monitored, the data are copied and filled into a message queue data structure, the artificial intelligent accelerator card can frequently transmit a large amount of data, such as model data, input data and the like, the artificial intelligent accelerator card is performance-sensitive equipment, and extra data carrying reduces board card efficiency.
Therefore, the above technical problems need to be solved by those skilled in the art.
Disclosure of Invention
In view of this, an object of the present invention is to provide a cross-platform communication method, system, apparatus, device and storage medium, which implement transparent data transfer for upper developers, do not need to care about any details of cross-platform transfer, and can greatly save data transfer time and improve communication efficiency. The specific scheme is as follows:
a first aspect of the present application provides a cross-platform communication method, including:
a first communication layer of a control server end provides a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transmission layer;
and controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer.
Alternatively to this, the first and second parts may, the first upper layer comprises a first application layer and a first runtime layer of the server end;
the first communication layer of the control server provides a cross-platform communication interface to a first upper layer of the server, and the cross-platform communication interface comprises:
and controlling the first communication layer to send the cross-platform communication interface to the first runtime layer, and forwarding the cross-platform communication interface to the first application layer through the first runtime layer.
Optionally, the controlling the first upper layer to call the cross-platform communication interface to send the operation request to the acceleration card end based on the direct memory access mechanism of the physical transport layer includes:
controlling the first application layer to call the cross-platform communication interface so as to send the operation request to the acceleration card terminal based on a direct memory access mechanism of the physical transmission layer.
Optionally, the controlling the first upper layer to call the cross-platform communication interface to send the operation request to the acceleration card end based on the direct memory access mechanism of the physical transport layer includes:
controlling the first upper layer to call the cross-platform communication interface so as to perform serialization operation on the operation request, and sending the operation request after the serialization operation to the acceleration card end based on a direct memory access mechanism of the physical transmission layer;
after the second communication layer controlling the accelerator card terminal obtains the operation request from the direct memory access controller of the physical transport layer, further comprising:
and controlling the second communication layer to perform deserialization operation on the operation request obtained after the serialization operation, and sending the operation request subjected to the deserialization operation to the second upper layer.
Optionally, the process of the serialization operation and the deserialization operation includes:
and carrying out serialization and deserialization operations on the operation request by using a flatbuffers tool.
Optionally, the second upper layer includes a second application layer and a second runtime layer of the acceleration card end;
the sending the operation request to a second upper layer of the accelerator card end includes:
and sending the operation request to the second runtime layer, and forwarding the operation request to the second application layer through the second runtime layer.
Optionally, before the controlling the second communication layer of the accelerator card end obtains the operation request from the direct memory access controller of the physical transport layer, the method further includes:
controlling PCIE drive of the physical transmission layer to generate an interrupt signal;
sending the interrupt signal to the second communication layer, and forwarding the interrupt signal to the second application layer through the second communication layer, so that the second application layer controls the second communication layer to acquire the operation request from a direct memory access controller of the physical transport layer according to the interrupt signal.
Optionally, the responding to the operation request by using the second upper layer includes:
controlling the second application layer to analyze the operation request to obtain a function parameter containing a target sub-process identifier;
determining a target sub-process having a binding relation with the target sub-process identifier; the target sub-process and the target business process running on the first upper layer and sending the operation request have a corresponding relation;
and responding to the operation request by utilizing the target subprocess.
Optionally, the target sub-process identifier is a target sub-process handle.
Optionally, the controlling the second application layer to parse the operation request includes:
controlling the second application layer to analyze the operation request to obtain an operation function type;
the responding to the operation request by using the target sub-process comprises:
and controlling the second application layer to call a target function corresponding to the type of the operation function from the second runtime layer, and executing the target function in the second runtime layer.
Optionally, the cross-platform communication method further includes:
and controlling the server to send an initialization command to the accelerator card end so that the accelerator card end creates the target sub-process handle and the target sub-process after receiving the initialization command and returns the target sub-process handle to the server end.
Optionally, the controlling the first upper layer to call the cross-platform communication interface to send the operation request to the acceleration card end based on the direct memory access mechanism of the physical transport layer includes:
and controlling the first upper layer to call the cross-platform communication interface so as to send the operation request carrying the target sub-process handle to the acceleration card terminal based on a direct memory access mechanism of the physical transmission layer.
Optionally, the creating, by the accelerator card end, the target sub-process handle after receiving the initialization command includes:
and the acceleration card terminal distributes the target subprocess handle in an idle state after receiving the initialization command.
Optionally, after the accelerator card terminal receives the initialization command and allocates one target sub-process handle in an idle state, the method further includes:
the acceleration card terminal binds the target sub-process handle with the target sub-process, and obtaining the binding relationship between the target sub-process handle and the target sub-process.
A second aspect of the present application provides a cross-platform communication system, including a server end and an accelerator card end, wherein the server end and the accelerator card end are arranged in a hierarchical manner, and the accelerator card end includes:
the first communication layer is arranged at the server end and used for providing a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
the first upper layer is arranged on the server side and used for calling the cross-platform communication interface so as to send an operation request to an acceleration card side based on a direct memory access mechanism of the physical transmission layer;
the second communication layer is arranged at the acceleration card end and used for acquiring the operation request from the direct memory access controller of the physical transmission layer and sending the operation request to a second upper layer of the acceleration card end;
the second upper layer is arranged at the acceleration card end and used for responding to the operation request.
Optionally, the first upper layer includes a first application layer and a first runtime layer of the server, where:
the first communication layer is further configured to send the cross-platform communication interface to the first runtime layer;
the first runtime layer is to forward the cross-platform communication interface to the first application layer.
Optionally, the second upper layer includes a second application layer and a second runtime layer of the accelerator card end, where:
the second communication layer is further configured to send the operation request to the second runtime layer;
the first runtime layer is configured to forward the operation request to the second application layer.
Optionally, the acceleration card end is further provided with a hardware module layer, configured to provide hardware and drive support for the acceleration card end.
A third aspect of the present application provides a cross-platform communication device comprising:
the interface providing module is used for controlling a first communication layer of the server end to provide a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
a request sending module, configured to control the first upper layer to call the cross-platform communication interface so as to send an operation request to an accelerator card end based on a direct memory access mechanism of the physical transport layer;
and the response module is used for controlling the second communication layer of the acceleration card end to acquire the operation request from the direct memory access controller of the physical transmission layer and sending the operation request to the second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer.
A fourth aspect of the present application provides an electronic device comprising a processor and a memory; wherein the memory is for storing a computer program that is loaded and executed by the processor to implement the aforementioned cross-platform communication method.
A fifth aspect of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are loaded and executed by a processor, the foregoing cross-platform communication method is implemented.
In the application, a first communication layer of a server end is controlled to provide a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer; then controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card end based on a direct memory access mechanism of the physical transmission layer; and finally, controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer. Therefore, the server end and the accelerator card end are designed in a layered mode, when cross-platform communication is conducted, the upper layer only needs to call the cross-platform communication interface provided by the communication layer, data transparent transmission is achieved for an upper layer developer, and any details of the cross-platform transmission do not need to be concerned. Meanwhile, the direct memory access mechanism based on the physical transmission layer avoids extra data copying operation, greatly saves data transmission time and improves communication efficiency.
Drawings
To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, reference will now be made in brief to the embodiments or drawings that are required for use in the description of the prior art, and it will be apparent that, the drawings in the following description are only embodiments of the invention and other drawings may be derived from the provided drawings by those skilled in the art without inventive effort.
FIG. 1 is a flowchart of a cross-platform communication method provided herein;
fig. 2 is a flowchart of a specific method for responding to the accelerator card end according to the present application;
FIG. 3 is a flowchart of a specific cross-platform communication method provided herein;
FIG. 4 is an embodiment provided by the present application a cross-platform communication system framework diagram of (1);
FIG. 5 is a specific cross-platform communication interaction diagram provided herein;
FIG. 6 is a schematic structural diagram of a cross-platform communication device according to the present application;
fig. 7 is a block diagram of a cross-platform communication electronic device according to the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The scheme that the existing artificial intelligence accelerates the cross-platform communication of the board card needs developers to have certain knowledge about PCIE technology on the one hand, the technical requirement commands of middle layer developers to bottom layer driving are increased, on the other hand, the artificial intelligence accelerator card has frequent mass data transmission, and extra data carrying reduces board card efficiency. In order to overcome the defect, the application provides a cross-platform communication scheme, the server end and the acceleration card end are designed in a layered mode, during cross-platform communication, an upper layer only needs to call a cross-platform communication interface provided by a communication layer, data transparent transmission is achieved for an upper layer developer, and any details of cross-platform transmission do not need to be concerned. At the same time, the user can select the desired position, the physical transport layer based direct memory access mechanism avoids additional data copy operations, greatly saving data transmission time and improving communication efficiency.
Fig. 1 is a flowchart of a cross-platform communication method according to an embodiment of the present disclosure. Referring to fig. 1, the cross-platform communication method includes:
s11: a first communication layer of a control server end provides a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer.
In this embodiment, a first communication layer of a server is controlled to provide a cross-platform communication interface to a first upper layer of the server, where the cross-platform communication interface is obtained by encapsulating a PCIE drive interface of a physical transmission layer by the first communication layer. The cross-platform communication interface is a simpler initialization function interface, a data receiving function interface, a data sending function interface and the like.
In this embodiment, the server is provided with the first communication layer and the first upper layer, where the first upper layer includes a first application layer and a first runtime layer of the server. When the cross-platform communication interface is provided, the first communication layer is mainly controlled to send the cross-platform communication interface to the first runtime layer, and the cross-platform communication interface is forwarded to the first application layer through the first runtime layer.
S12. The method of preparation the method comprises the following steps: and controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transmission layer.
In this embodiment, the first upper layer is controlled to call the cross-platform communication interface so as to send the operation request to the accelerator card terminal based on the direct memory access mechanism of the physical transport layer. The Direct Memory Access mechanism of the physical transmission layer is a Direct Memory Access (DMA) transmission mechanism of the PCIE, and the DMA can be used for quickly copying data of different hardware devices without depending on a CPU (central processing unit) and a DMA controller, so that the execution efficiency of the CPU is greatly improved. Specifically, the first application layer is controlled to call the cross-platform communication interface so as to send the operation request to the acceleration card terminal based on the direct memory access mechanism of the physical transport layer.
S13: and controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer.
In this embodiment, the second communication layer of the accelerator card end is controlled to obtain the operation request from the dma controller of the physical transport layer, and send the operation request to the second upper layer of the accelerator card end, so as to respond to the operation request by using the second upper layer. Similarly, the second upper layer includes a second application layer and a second runtime layer of the accelerator card end, and when the operation request is transferred, the operation request is mainly sent to the second runtime layer, and the operation request is forwarded to the second application layer through the second runtime layer. The specific process of the step is as follows (fig. 2):
s131: and controlling the PCIE drive of the physical transmission layer to generate an interrupt signal.
S132: sending the interrupt signal to the second communication layer, and forwarding the interrupt signal to the second application layer through the second communication layer, so that the second application layer controls the second communication layer to acquire the operation request from a direct memory access controller of the physical transport layer according to the interrupt signal.
In this embodiment, after receiving the operation request sent by the server layer, the physical transport layer controls the PCIE driver of the physical transport layer to generate an interrupt signal, then sends the interrupt signal to the second communication layer, and forwards the interrupt signal to the second application layer through the second communication layer, so that the second application layer controls the second communication layer to obtain the operation request from the dma controller of the physical transport layer according to the interrupt signal. At this time, the operation request is sent to the second runtime layer, and the operation request is forwarded to the second application layer through the second runtime layer.
S133: controlling the second application layer to analyze the operation request to obtain a function parameter and an operation function type which comprise a target sub-process identifier; and the target sub-process identifier is a target sub-process handle.
S134: determining a target sub-process having a binding relation with the target sub-process identifier; and the target sub-process and the target business process which is operated by the first upper layer and sends the operation request have a corresponding relation.
S135: and controlling the second application layer to call a target function corresponding to the type of the operation function from the second runtime layer, and executing the target function in the second runtime layer.
In this embodiment, the second application layer is controlled to parse the operation request to obtain a function parameter and an operation function type that include the identifier of the target sub-process. And the target sub-process identifier is a target sub-process handle, and the operation request comprises a function parameter and an operation function type. And then determining the target sub-process having the binding relation with the target sub-process identifier. And the target sub-process and the target business process running on the first upper layer for sending the operation request have a corresponding relation. And finally, controlling the second application layer to call a target function corresponding to the type of the operation function from the second runtime layer, and executing the target function in the second runtime layer.
Before cross-platform communication, the target sub-process handle and the target sub-process are created through accelerator card initialization operation, and the target sub-process handle is carried during cross-platform communication. Namely, the first upper layer is controlled to call the cross-platform communication interface so as to send the operation request carrying the target sub-process handle to the acceleration card terminal based on the direct memory access mechanism of the physical transmission layer. That is to say, the handle is used by subsequent access requests of the business process to the board card end, so that the consistency of the operation of the acceleration card end is realized in a handle mode, for example, the acceleration card end process configures model information first and then specifies input data, and different processes handle the process independently, which results in the invalidation of the operation of other processes to the current process. Specifically, the server is controlled to send an initialization command to the accelerator card end, so that the accelerator card end creates the target sub-process handle and the target sub-process after receiving the initialization command, and returns the target sub-process handle to the server end. And when the target sub-process handle is created, the acceleration card terminal receives the initialization command and then allocates the target sub-process handle in an idle state. And simultaneously, the acceleration card end binds the target sub-process handle with the target sub-process to obtain the binding relationship between the target sub-process handle and the target sub-process.
As can be seen, in the embodiment of the present application, a first communication layer of a server end is controlled to provide a cross-platform communication interface to a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer; then controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card end based on a direct memory access mechanism of the physical transmission layer; and finally, controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer. According to the embodiment of the application, the server end and the acceleration card end are designed in a layered mode, when cross-platform communication is conducted, the upper layer only needs to call the cross-platform communication interface provided by the communication layer, data transparent transmission is achieved for upper developers, and any details of cross-platform transmission do not need to be concerned. Meanwhile, the direct memory access mechanism based on the physical transmission layer avoids extra data copying operation, greatly saves data transmission time and improves communication efficiency.
Fig. 3 is a flowchart of a specific cross-platform communication method according to an embodiment of the present disclosure. Referring to fig. 3, the cross-platform communication method includes:
s21: a first communication layer of a control server end provides a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer.
In this embodiment, as to the specific process of the step S21, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
S22: and controlling the first upper layer to call the cross-platform communication interface so as to perform serialization operation on the operation request, and sending the operation request after the serialization operation to an acceleration card end based on a direct memory access mechanism of the physical transmission layer.
The accelerator card needs to be compatible with more Linux system platforms, such as the common 32bit/64bit Linux system, which requires simple data transfer, and the parsing fails due to the difference of bit width. For example, the request data structure contains a long type, the type occupies 32 bits in a 32-bit Linux system, and the type occupies 64 bits in a 64-bit Linux system, which causes a difference in data analysis between transmission and reception, and further causes data errors, and the serialization operation realizes the analysis consistency of cross-platform data by uniformly defining the bit width length of transmission data, for example, the long type data is stipulated to be 64bit, and a 32bit system needs to receive the data by using a 64-bit data type, so that the problem caused by platform difference is solved by using a method of stipulating the data length.
Therefore, the present embodiment overcomes the problem of inconsistent data bit width caused by the difference of the Linux system through the serialization operation and the deserialization operation. Specifically, the first upper layer is controlled to call the cross-platform communication interface to perform serialization operation on an operation request, and the operation request after the serialization operation is sent to an acceleration card end based on a direct memory access mechanism of the physical transmission layer.
S23: and controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer.
In this embodiment, as to the specific process of the step S23, reference may be made to corresponding contents disclosed in the foregoing embodiments, and details are not repeated here.
S24: and controlling the second communication layer to perform deserialization operation on the operation request obtained after the serialization operation, and sending the operation request after the deserialization operation to a second upper layer of the acceleration card end, wherein the second upper layer of the acceleration card end is a layer of the operation request obtained after the serialization operation.
In this embodiment, the second communication layer is controlled to perform deserialization operation on the operation request after the serialization operation is obtained, and the operation request after the deserialization operation is sent to a second upper layer of the acceleration card end, where the second upper layer of the acceleration card end is located. Specifically, the operations request may be serialized and deserialized by using a flitbuffers tool, and the tool for the serialization and deserialization operations is not limited in this embodiment.
As can be seen, in the embodiment of the present application, a first communication layer of a server end is controlled to provide a cross-platform communication interface to a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer; then controlling the first upper layer to call the cross-platform communication interface so as to perform serialization operation on the operation request, and sending the operation request after the serialization operation to an acceleration card end based on a direct memory access mechanism of the physical transmission layer; then controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer; and finally, controlling the second communication layer to perform deserialization operation on the operation request after the serialization operation is obtained, and sending the operation request after the deserialization operation to a second upper layer of the acceleration card end, wherein the second upper layer of the acceleration card end is a second upper layer of the acceleration card end. The embodiment of the application carries out serialization and deserialization operations on the transmitted data so as to be compatible with different systems and solve the problem of data analysis errors caused by system differences.
Fig. 4 is a specific cross-platform communication system framework diagram provided in an embodiment of the present application. Referring to fig. 4, the cross-platform communication system includes a server end and an accelerator card end, wherein the server end and the accelerator card end are arranged in a hierarchical manner, and:
the first communication layer L1 is arranged at the server end and used for providing a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by encapsulating a PCIE driving interface of a physical transmission layer by the first communication layer L1; the first upper layer L2 is arranged at the server end and used for calling the cross-platform communication interface so as to send an operation request to the acceleration card end based on a direct memory access mechanism of the physical transmission layer L3; the second communication layer L4 is arranged at the accelerator card end, and is configured to obtain the operation request from the direct memory access controller of the physical transport layer L3, and send the operation request to the second upper layer L5 of the accelerator card end; the second upper layer L5 disposed at the accelerator card end, for responding to the operation request.
The first upper layer L2 includes a first application layer L21 and a first runtime layer L22 of the server side, where: the first communication layer L1 is further configured to send the cross-platform communication interface to the first runtime layer L22; the first runtime layer L22 is configured to forward the cross-platform communication interface to the first application layer L21. The second upper layer L5 includes a second application layer L51 and a second runtime layer L52 at the accelerator card end, where: the second communication layer L4 is further configured to send the operation request to the second runtime layer L52; the second runtime layer L52 is configured to forward the operation request to the second application layer L51. And the acceleration card end is also provided with a hardware module layer L6 for providing hardware and drive support for the acceleration card end.
The frame diagram of fig. 4 is divided into a server end portion on the left and an accelerator card end portion on the right. The first communication layer L1 and the first upper layer L2 (the first application layer L21 and the first runtime layer L22) are server-side software components. The L1 layer provides a cross-platform communication interface for the L22 layer upwards, and the transparent transmission between the L22 layer and the board card is realized mainly in the layer: the PCIE driving interface is uniformly controlled and packaged into a simpler function interface for initialization, data receiving and data sending to an L22 layer, transparent transmission of the L22 layer and an L52 layer is realized, and cross-platform transmission details are not required; and carrying out serialization and deserialization operations on the transmitted data so as to be compatible with different systems and solve the problem of data analysis errors caused by platform differences. The L21 layer is a service program of a user, the function realized by the L22 layer is to provide an interface used by the L21 layer upwards so as to use the board card resource to carry out the service program, and call the interface of the L1 layer downwards to realize cross-platform transmission of a service layer request, and the layer and the L4 layer complete the processing of the L21 layer software logic together.
The physical transmission layer L3 is a PCIE physical interface device at the server and the accelerator card end and a corresponding driver thereof. And realizing cross-platform transfer of data. When the board card end or the server end receives data, the PCIE drive generates an interrupt to inform the upper layer so that the upper layer can carry out subsequent operation, and the PCIE also supports a DMA data transmission mode to finish the high-efficiency transmission of the data.
The second communication layer L4 and the second upper layer L5 (the second application layer L51 and the second runtime layer L52) are software parts of an acceleration card end, and the hardware module layer L6 is each module and its driver inside the acceleration card end. The L4 layer has the same function as the L1 layer, the L51 layer is a service process of the acceleration card end, the process senses that data is received by the acceleration card end by receiving an interrupt signal sent by the L3 layer, the service process deserializes the data and then obtains handle information, and the handle marks which sub-process the data is handed to for execution, namely, the subsequent request data is obtained from a data receiving memory area and the request is executed. The L52 layer is a process in which the accelerator card end receives access data sent by the server end and processes the data in response, if the return data exists, the return data is placed in the accelerator card end sending memory area, the PCIE driver will generate an interrupt at the server end at the same time, so that the server end receives the data in the related process, and the above operation is a cross-platform transmission process.
The following describes in detail a processing scheme for implementing data transfer and service request by using the cross-platform communication framework of this embodiment, taking a scenario of using an accelerator card to train or infer a neural network as an example:
firstly, a service process of a service end needs to initialize an accelerator card, namely, an access accelerator card end executes initialization operation, because the operation does not acquire a handle from the accelerator card end yet, the operation which needs to be performed by the service process is known through a function, the service process of the accelerator card end detects that a subprocess is created after the initialization operation of a board card, an unused handle number is allocated to mark the subprocess, and the handle is returned to the service process, and the service process performs data interaction based on the handle later. And during each interaction, the server provides the handle and the corresponding function name and parameter, the acceleration card end identifies the handle and hands the following request data to the corresponding process for processing, the acceleration card end returns the data to the sending memory after processing, and the server end obtains the data. The specific interaction process is shown in fig. 5.
In the above flow, the initialization board operation at the server side in the L1 layer of fig. 4 serializes the operation, and the serialization is generally implemented by using a flitbuffers technology, where the serializing content includes a function type and a function parameter, for example, an initialization operation, where the function is initialization devinit and the parameter is handle. And after the serialization is finished, the data is driven by the PCIE to be finished and sent, the PCIE places the data in a shared memory based on a DMA data transmission mode so as to be directly accessed and acquired by the accelerator card end, and simultaneously triggers an interrupt at the accelerator card end so as to inform that the data is waiting to be acquired. And after receiving the interrupt, the accelerated card end service process acquires data from a receiving memory, namely a shared memory, and judges that the function is the initialization of the board card after deserialization, so that the service process creates a sub-process, allocates an idle handle to be bound with the sub-process, returns the handle to the server end, and returns the data to be the same as the serialized sending operation of the server end. After the operations are completed, the service process of the server and the processing operation of the acceleration card end can be bound one by one, and the service process of the acceleration card end is assigned according to the request handle to complete the full-flow data transmission from the service process to the board card processing process. Business operations after the business process, such as transferring model data, configuring model parameters, issuing inference execution commands and the like, are performed based on the above flows.
Referring to fig. 6, an embodiment of the present application further discloses a cross-platform communication apparatus, which includes:
the interface providing module 11 is configured to control a first communication layer of a server to provide a cross-platform communication interface to a first upper layer of the server; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
a request sending module 12, configured to control the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card end based on a direct memory access mechanism of the physical transport layer;
the response module 13 is configured to control the second communication layer of the acceleration card end to obtain the operation request from the dma controller of the physical transport layer, and send the operation request to the second upper layer of the acceleration card end, so as to use the second upper layer to respond to the operation request.
As can be seen, the method comprises the steps that a first communication layer of a server end is controlled to provide a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer; then controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card end based on a direct memory access mechanism of the physical transmission layer; finally, the second communication layer of the acceleration card end is controlled to obtain the operation request from the direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer. According to the embodiment of the application, the server end and the acceleration card end are designed in a layered mode, when cross-platform communication is conducted, the upper layer only needs to call the cross-platform communication interface provided by the communication layer, data transparent transmission is achieved for upper developers, and any details of cross-platform transmission do not need to be concerned. Meanwhile, the direct memory access mechanism based on the physical transmission layer avoids extra data copying operation, greatly saves data transmission time and improves communication efficiency.
In some specific embodiments, the first upper layer includes a first application layer and a first runtime layer of the server, and the interface providing module 11 specifically includes:
a first sending unit for sending the data to the first sending unit, the cross-platform communication interface is used for transmitting the cross-platform communication interface to the first runtime layer;
a second sending unit, configured to forward the cross-platform communication interface to the first application layer via the first runtime layer.
In some embodiments, the request sending module 12 is specifically configured to control the first application layer to call the cross-platform communication interface so as to send the operation request to the accelerator card terminal based on a direct memory access mechanism of the physical transport layer.
In some specific embodiments, the request sending module 12 is further specifically configured to be a serialization operation module, configured to control the first upper layer to call the cross-platform communication interface to perform serialization operation on the operation request, and send the operation request after the serialization operation to the acceleration card terminal based on a direct memory access mechanism of the physical transport layer.
In some specific embodiments, the second upper layer includes a second application layer and a second runtime layer of the accelerator card end, and the response module 13 is specifically configured to send the operation request to the second runtime layer, and forward the operation request to the second application layer through the second runtime layer.
In some embodiments, the cross-platform communication device further comprises:
the deserializing operation module is used for controlling the second communication layer to perform deserializing operation on the operation request obtained after the serializing operation and sending the operation request subjected to the deserializing operation to the second upper layer;
the interrupt signal generation module is used for controlling the PCIE drive of the physical transmission layer to generate an interrupt signal;
an interrupt signal sending module, configured to send the interrupt signal to the second communication layer, and forward the interrupt signal to the second application layer via the second communication layer, so that the second application layer controls the second communication layer to obtain the operation request from a direct memory access controller of the physical transport layer according to the interrupt signal;
and the initialization module is used for controlling the server to send an initialization command to the accelerator card end so that the accelerator card end creates the target sub-process handle and the target sub-process after receiving the initialization command and returns the target sub-process handle to the server end.
In some specific embodiments, the response module 13 specifically further includes:
the analysis unit is used for controlling the second application layer to analyze the operation request to obtain a function parameter containing a target subprocess identifier;
the determining unit is used for determining the target sub-process having the binding relation with the target sub-process identifier; the target sub-process and the target business process running on the first upper layer and sending the operation request have a corresponding relation;
and the response unit is used for responding to the operation request by utilizing the target subprocess.
In some embodiments, the target sub-process in the cross-platform communication device is identified as a target sub-process handle.
In some specific embodiments, the parsing unit is further configured to control the second application layer to parse the operation request to obtain the operation function type.
In some specific embodiments, the response unit in the cross-platform communication device specifically includes:
a function calling subunit, configured to control the second application layer to call, from the second runtime layer, a target function corresponding to the operation function type;
an execution subunit, configured to execute the objective function at the second runtime level.
In some specific embodiments, the request sending module 12 is further specifically configured to control the first upper layer to call the cross-platform communication interface so as to send the operation request carrying the target sub-process handle to the acceleration card terminal based on a direct memory access mechanism of the physical transport layer.
In some specific embodiments, the initialization module specifically includes:
a handle creating unit, configured to allocate, by the accelerator card end, a handle of the target sub-process in an idle state after receiving the initialization command;
and the binding unit is used for binding the target sub-process handle with the target sub-process by the accelerator card end to obtain the binding relationship between the target sub-process handle and the target sub-process.
Further, the embodiment of the application also provides electronic equipment. FIG. 7 is a block diagram illustrating an electronic device 20 according to an exemplary embodiment, and the contents of the diagram should not be construed as limiting the scope of use of the present application in any way.
Fig. 7 is a schematic structural diagram of an electronic device 20 according to an embodiment of the present disclosure. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input output interface 25, and a communication bus 26. Wherein the memory 22 is used for storing a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the cross-platform communication method disclosed in any of the foregoing embodiments.
In this embodiment, the power supply 23 is configured to provide a working voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and an external device, and a communication protocol followed by the communication interface is any communication protocol applicable to the technical solution of the present application, and is not specifically limited herein; the input/output interface 25 is configured to obtain external input data or output data to the outside, and a specific interface type thereof may be selected according to specific application requirements, which is not specifically limited herein.
In addition, the storage 22 is used as a carrier for resource storage, and may be a read-only memory, a random access memory, a magnetic disk or an optical disk, etc., and the resources stored thereon may include an operating system 221, a computer program 222, data 223, etc., and the storage may be a transient storage or a permanent storage.
The operating system 221 is used for managing and controlling each hardware device and the computer program 222 on the electronic device 20, so as to realize the operation and processing of the mass data 223 in the memory 22 by the processor 21, and may be Windows Server, netware, unix, linux, and the like. The computer programs 222 may further include computer programs that can be used to perform other specific tasks in addition to the computer programs that can be used to perform the cross-platform communication method performed by the electronic device 20 disclosed in any of the foregoing embodiments. Data 223 may include operational requests collected by electronic device 20.
Further, an embodiment of the present application further discloses a storage medium, in which a computer program is stored, and when the computer program is loaded and executed by a processor, the steps of the cross-platform communication method disclosed in any of the foregoing embodiments are implemented.
The embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same or similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a" \8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises the element.
The cross-platform communication method, system, device, equipment and medium provided by the invention are described in detail, and the principle and the implementation mode of the invention are explained by applying specific examples, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (21)
1. A cross-platform communication method, comprising:
a first communication layer of a control server end provides a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
controlling the first upper layer to call the cross-platform communication interface so as to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transmission layer;
and controlling a second communication layer of the acceleration card end to acquire the operation request from a direct memory access controller of the physical transmission layer, and sending the operation request to a second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer.
2. The cross-platform communication method according to claim 1, wherein the first upper layer comprises a first application layer and a first runtime layer of the server side;
the first communication layer of the control server provides a cross-platform communication interface to a first upper layer of the server, and the cross-platform communication interface comprises:
and controlling the first communication layer to send the cross-platform communication interface to the first runtime layer, and forwarding the cross-platform communication interface to the first application layer through the first runtime layer.
3. The cross-platform communication method according to claim 2, wherein the controlling the first upper layer to call the cross-platform communication interface to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transport layer comprises:
controlling the first application layer to call the cross-platform communication interface so as to send the operation request to the acceleration card terminal based on a direct memory access mechanism of the physical transmission layer.
4. The cross-platform communication method according to claim 1, wherein the controlling the first upper layer to call the cross-platform communication interface to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transport layer comprises:
controlling the first upper layer to call the cross-platform communication interface so as to perform serialization operation on the operation request, and sending the operation request after the serialization operation to the acceleration card end based on a direct memory access mechanism of the physical transmission layer;
after the controlling the second communication layer of the acceleration card end obtains the operation request from the direct memory access controller of the physical transport layer, the method further includes:
and controlling the second communication layer to perform deserialization operation on the operation request obtained after the serialization operation, and sending the operation request subjected to the deserialization operation to the second upper layer.
5. The cross-platform communication method of claim 4, wherein the process of serializing and deserializing comprises:
and carrying out serialization and deserialization operations on the operation request by using a flatbuffers tool.
6. The cross-platform communication method according to claim 1, wherein the second upper layer comprises a second application layer and a second runtime layer of the acceleration card end;
the sending the operation request to a second upper layer of the accelerator card end includes:
and sending the operation request to the second runtime layer, and forwarding the operation request to the second application layer through the second runtime layer.
7. The cross-platform communication method according to claim 6, wherein before the controlling of the second communication layer of the acceleration card end obtains the operation request from the direct memory access controller of the physical transport layer, the method further comprises:
controlling PCIE drive of the physical transmission layer to generate an interrupt signal;
sending the interrupt signal to the second communication layer, and forwarding the interrupt signal to the second application layer through the second communication layer, so that the second application layer controls the second communication layer to acquire the operation request from a direct memory access controller of the physical transport layer according to the interrupt signal.
8. The cross-platform communication method according to claim 6, wherein responding to the operation request by the second upper layer comprises:
controlling the second application layer to analyze the operation request to obtain a function parameter containing a target sub-process identifier;
determining a target sub-process having a binding relation with the target sub-process identifier; the target sub-process and the target business process running on the first upper layer and sending the operation request have a corresponding relation;
and responding to the operation request by utilizing the target subprocess.
9. The cross-platform communication method of claim 8, wherein the target sub-process is identified as a target sub-process handle.
10. The cross-platform communication method according to claim 9, wherein the controlling the second application layer to parse the operation request comprises:
controlling the second application layer to analyze the operation request to obtain an operation function type;
the responding to the operation request by using the target sub-process comprises:
and controlling the second application layer to call a target function corresponding to the type of the operation function from the second runtime layer, and executing the target function in the second runtime layer.
11. The cross-platform communication method of claim 9, further comprising:
and controlling the server to send an initialization command to the accelerator card end so that the accelerator card end creates the target sub-process handle and the target sub-process after receiving the initialization command and returns the target sub-process handle to the server end.
12. The cross-platform communication method according to claim 11, wherein the controlling the first upper layer to call the cross-platform communication interface to send an operation request to an acceleration card terminal based on a direct memory access mechanism of the physical transport layer comprises:
and controlling the first upper layer to call the cross-platform communication interface so as to send the operation request carrying the target sub-process handle to the acceleration card terminal based on a direct memory access mechanism of the physical transmission layer.
13. The cross-platform communication method according to claim 11, wherein the creating the target sub-process handle after the accelerator card terminal receives the initialization command comprises:
and the acceleration card terminal distributes the target subprocess handle in an idle state after receiving the initialization command.
14. The method according to claim 13, wherein after the accelerator card terminal receives the initialization command and allocates one of the target sub-process handles in an idle state, the method further comprises:
and the acceleration card terminal binds the target sub-process handle with the target sub-process to obtain the binding relationship between the target sub-process handle and the target sub-process.
15. A cross-platform communication system is characterized by comprising a server end and an acceleration card end which are arranged in a layered mode, wherein:
the first communication layer is arranged at the server end and used for providing a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
the first upper layer is arranged at the server end and used for calling the cross-platform communication interface so as to send an operation request to the acceleration card end based on a direct memory access mechanism of the physical transmission layer;
the second communication layer is arranged at the acceleration card end and used for acquiring the operation request from the direct memory access controller of the physical transmission layer and sending the operation request to a second upper layer of the acceleration card end;
and the second upper layer is arranged at the acceleration card end and used for responding to the operation request.
16. The cross-platform communication system according to claim 15, wherein the first upper layer comprises a first application layer and a first runtime layer of the server side, wherein:
the first communication layer is further configured to send the cross-platform communication interface to the first runtime layer;
the first runtime layer is to forward the cross-platform communication interface to the first application layer.
17. The cross-platform communication system according to claim 15, wherein the second upper layer comprises a second application layer and a second runtime layer of the accelerator card terminal, wherein:
the second communication layer is further configured to send the operation request to the second runtime layer;
the second runtime layer is configured to forward the operation request to the second application layer.
18. The cross-platform communication system according to claim 15, wherein the acceleration card end is further provided with a hardware module layer for providing hardware and driving support for the acceleration card end.
19. A cross-platform communication device, comprising:
the interface providing module is used for controlling a first communication layer of the server end to provide a cross-platform communication interface for a first upper layer of the server end; the cross-platform communication interface is obtained by packaging a PCIE driving interface of a physical transmission layer by the first communication layer;
a request sending module, configured to control the first upper layer to call the cross-platform communication interface so as to send an operation request to an accelerator card end based on a direct memory access mechanism of the physical transport layer;
and the response module is used for controlling the second communication layer of the acceleration card end to acquire the operation request from the direct memory access controller of the physical transmission layer and sending the operation request to the second upper layer of the acceleration card end so as to respond to the operation request by utilizing the second upper layer.
20. An electronic device, comprising a processor and a memory; wherein the memory is for storing a computer program that is loaded and executed by the processor to implement the cross-platform communication method of any of claims 1 to 14.
21. A computer-readable storage medium storing computer-executable instructions which, when loaded and executed by a processor, implement the cross-platform communication method of any one of claims 1 to 14.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310154595.5A CN115827285B (en) | 2023-02-23 | 2023-02-23 | Cross-platform communication method, system, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310154595.5A CN115827285B (en) | 2023-02-23 | 2023-02-23 | Cross-platform communication method, system, device, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115827285A true CN115827285A (en) | 2023-03-21 |
| CN115827285B CN115827285B (en) | 2023-05-23 |
Family
ID=85522136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310154595.5A Active CN115827285B (en) | 2023-02-23 | 2023-02-23 | Cross-platform communication method, system, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115827285B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648211A (en) * | 2024-01-29 | 2024-03-05 | 山东云海国创云计算装备产业创新中心有限公司 | Runtime unified interface, server and calling method of artificial intelligent framework |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115374043A (en) * | 2022-09-22 | 2022-11-22 | 成都乐创自动化技术股份有限公司 | Communication protocol data interaction method based on PCIE |
-
2023
- 2023-02-23 CN CN202310154595.5A patent/CN115827285B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115374043A (en) * | 2022-09-22 | 2022-11-22 | 成都乐创自动化技术股份有限公司 | Communication protocol data interaction method based on PCIE |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648211A (en) * | 2024-01-29 | 2024-03-05 | 山东云海国创云计算装备产业创新中心有限公司 | Runtime unified interface, server and calling method of artificial intelligent framework |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115827285B (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110620712B (en) | Method for realizing real-time EtherCAT master station of Window platform | |
| CN111542064B (en) | Container arrangement management system and arrangement method for wireless access network | |
| CN112416840B (en) | Remote mapping method, device, equipment and storage medium for computing resources | |
| US7231638B2 (en) | Memory sharing in a distributed data processing system using modified address space to create extended address space for copying data | |
| EP0312739B1 (en) | Apparatus and method for interconnecting an application of a transparent services access facility to a remote source | |
| KR100253930B1 (en) | Dynamic eaecution unit management for high performance user level network protocol server system | |
| CN102103526A (en) | Method and system for performing inter-process communication between server and client by service management | |
| CN110658794B (en) | Manufacturing execution system | |
| CN103747107B (en) | A kind of compatible cloud operating platform and its implementation | |
| CN111176801B (en) | Multi-process management method, device, equipment and storage medium | |
| CN103116520A (en) | Remote procedure call remote position control (RPC) method based on transmission control protocol (TCP)/user datagram protocol (UDP) | |
| CN112068970B (en) | RPC implementation method and medium based on heterogeneous multi-core processor | |
| CN115827285B (en) | Cross-platform communication method, system, device, equipment and medium | |
| JP2022522609A (en) | How to execute user code of programming platform and platform, node, device, medium | |
| CN114691286A (en) | Server system, virtual machine creation method and device | |
| US6052729A (en) | Event-reaction communication protocol in an object oriented processor array | |
| CN112887121B (en) | Method, equipment and system for managing network service | |
| CN105487403A (en) | Establishment of motion control system based on CAN and simulation modeling method | |
| JPH11312079A (en) | Method for controlling portable information processor | |
| CN116800616B (en) | Management method and related device of virtualized network equipment | |
| CN110430110B (en) | On-site bus gateway and protocol conversion method thereof | |
| CN113342456A (en) | Connection method, device, equipment and storage medium | |
| CN116450109A (en) | Modbus communication library adaptation method under MATLAB and CCS combined development mode | |
| KR100331519B1 (en) | Computerized method and system for implementing distributed applications | |
| CN116643854A (en) | Service arrangement method and device, flow controller and task processor |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |