CN114844888B - Power computer and processing method thereof - Google Patents
Power computer and processing method thereof Download PDFInfo
- Publication number
- CN114844888B CN114844888B CN202210347088.9A CN202210347088A CN114844888B CN 114844888 B CN114844888 B CN 114844888B CN 202210347088 A CN202210347088 A CN 202210347088A CN 114844888 B CN114844888 B CN 114844888B
- Authority
- CN
- China
- Prior art keywords
- power
- computing
- cloud
- computing power
- computer
- 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.)
- Active
Links
- 238000003672 processing method Methods 0.000 title abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 18
- 230000008569 process Effects 0.000 claims abstract description 9
- 238000006243 chemical reaction Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000012544 monitoring process Methods 0.000 claims description 13
- 230000008901 benefit Effects 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 claims description 4
- 239000002699 waste material Substances 0.000 abstract description 6
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000005304 joining Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/06—Generation of reports
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- 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
Abstract
The invention discloses a power computer and a processing method thereof. The power computing machine comprises a power computing processing device, a network interface and a cloud aggregator, wherein the power computing processing device is connected with the network interface and the cloud aggregator, and the network interface and the cloud aggregator are connected with at least two power computing clouds through a network; the computing power processing device is communicated with the computing power cloud through a network interface, when the computing power processing device is in an idle state, if the computing power demand of any computing power cloud is detected, a computing power resource pool corresponding to the computing power cloud is added, the cloud aggregator receives computing tasks distributed by the computing power cloud and converts the computing tasks into a computer computing instruction, and the computing power processing device processes the computer computing instruction. By adopting the method and the device, one power machine can dynamically migrate among different power clouds, and resource waste is avoided.
Description
Technical Field
The invention relates to the technical field of cloud computing, in particular to a power computer and a processing method thereof.
Background
With the rapid development of high-performance computing and artificial intelligence technology, the demands of the fields of automatic driving, image recognition, gene detection, blockchain and the like on computing power are continuously increased. On the basis, the industry provides a novel resource integration scheme of an 'computational power network', and the computational power information, the storage information, the algorithm information and the like of the service nodes are distributed through the network, and the information of network states (such as paths and time delays) and the like are combined, so that an optimal resource distribution and network connection scheme is provided for customer demands, and the optimal use of the whole network resources is realized.
The computing power network utilizes novel network technologies such as cloud network fusion technology, SDN (software defined network)/NFV (Network Functions Virtualization network function virtualization) and the like to deeply fuse edge computing nodes, cloud computing nodes and various network resources including wide area networks together, reduces the management and control complexity of the edge computing nodes, and cooperates with computing and storage resources of the cloud computing nodes and network resources of the wide area networks through a centralized control or distributed scheduling method to form a new generation information infrastructure, so that an overall computing power service comprising computing, storage and connection is provided for clients, and flexible and schedulable service according to service characteristics is provided.
At present, because different computing networks have different management operations, required performances, provided services and services, computing nodes are limited to one computing network, such as one computing machine for one computing cloud, so that the computing machine cannot move to the computing cloud with other resources in shortage even when the computing machine is idle, and resource waste is caused.
Disclosure of Invention
The invention aims to solve the technical problems that: in the prior art, a power computer can not flexibly and dynamically migrate to other power clouds, so that resource waste is caused.
In order to solve the technical problems, the invention provides a power machine and a processing method thereof.
A power computer, comprising: the computing power system comprises a computing power processing device, a network interface and a cloud aggregator, wherein the computing power processing device is connected with the network interface and the cloud aggregator, and the network interface and the cloud aggregator are connected with at least two computing power clouds through a network;
the computing power processing device is communicated with the computing power cloud through the network interface, if the computing power demand of any computing power cloud is detected when the computing power processing device is in an idle state, a computing power resource pool corresponding to the computing power cloud is added, the cloud aggregator receives the computing tasks distributed by the computing power cloud and converts the computing tasks into a computer computing instruction, and the computing power processing device processes the computer computing instruction.
In one embodiment, the computing power processing device comprises a processor and a computing power board, wherein the processor is connected with the computing power board and the network interface, and the computing power board is connected with the cloud aggregator;
the processor communicates with the computing power clouds through the network interface, and if the computing power demand of any computing power cloud is detected when the processor is in an idle state, the processor adds a computing power resource pool corresponding to the computing power clouds;
and the power computing board receives and processes the local computing instruction sent by the cloud aggregator.
In one embodiment, the computing board comprises a controller and a plurality of computing chips, the controller connecting the processor, the cloud aggregator, and the plurality of computing chips;
and the controller receives the local computing instruction sent by the cloud aggregator and sends the local computing instruction to the power computing chip for processing.
In one embodiment, the cloud aggregator includes an access interface and a conversion module, the access interface is connected to the conversion module and to the computing power cloud via a network, and the conversion module is connected to the computing power processing device.
In one embodiment, the cloud aggregator further comprises an account management module, and the account management module is connected with the conversion module.
In one embodiment, the cloud aggregator further comprises a monitoring module, and the monitoring module connects the conversion module and the computing power processing device.
In one embodiment, the cloud aggregator further comprises a policy module connecting the monitoring module, the account management module, and the conversion module.
In one embodiment, the power computer further comprises at least one of a memory and an input/output port, and the memory and the input/output port are connected with the power computing device.
A method of processing the power computer, comprising:
detecting whether the connected computing power cloud has computing power requirements when the computing power cloud is in an idle state;
if yes, adding an computing power resource pool of the computing power cloud with computing power requirements;
acquiring a local computing instruction obtained by conversion after a cloud aggregator receives a computing task distributed by the added computing power cloud;
and processing the local computing instruction.
In one embodiment, the detecting whether the connected computing power cloud has a computing power demand includes:
judging whether an add power machine request of any one power cloud is received or not;
if yes, determining that the computing power cloud sending the adding computing power machine request has a computing power requirement;
if not, sending an inquiry message to the connected power computer, wherein the inquiry message is used for inquiring whether the power computer is added;
when receiving a positive reply message returned by any one of the computers according to the inquiry message, determining that the computer returning the positive reply message has a calculation power requirement;
and when receiving negative reply messages returned by all the computers according to the inquiry messages, determining that no calculation force is required.
In one embodiment, before detecting whether the connected computing power cloud has a computing power demand when in the idle state, the method further includes:
and interacting the device characteristics and state information of both parties with the connected computing power cloud.
One or more embodiments of the above-described solution may have the following advantages or benefits compared to the prior art:
and when in an idle state, the computing power cloud with computing power requirements is added, the computing power cloud is connected through the cloud aggregator, and the computing tasks sent by the computing power cloud are converted into instructions which can be understood by the computing power machine, so that the computing tasks of different computing power clouds can be processed by the computing power machine, one computing power machine can dynamically migrate among different computing power clouds, namely, one computing power machine can be suitable for a plurality of computing power clouds, and resource waste is avoided.
Drawings
The scope of the present disclosure may be better understood by reading the following detailed description of exemplary embodiments in conjunction with the accompanying drawings. The drawings included herein are:
FIG. 1 is a block diagram of a power machine in one embodiment;
FIG. 2 is a block diagram of a power machine in another embodiment;
FIG. 3 is a block diagram of the architecture of a force plate in one embodiment;
FIG. 4 is a block diagram of the architecture of a cloud aggregator in one embodiment;
FIG. 5 is a schematic diagram of a connection relationship between a computing machine and a computing cloud in one embodiment;
FIG. 6 is a flow chart of a method of processing a power machine in one embodiment;
FIG. 7 is a flow chart of a processing method of the power machine according to another embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the following detailed description of the implementation method of the present invention will be given with reference to the accompanying drawings and examples, by which the technical means are applied to solve the technical problems, and the implementation process for achieving the technical effects can be fully understood and implemented accordingly.
In one embodiment, a power machine is provided, as shown in FIG. 1, comprising: the computing power processing device 110, the cloud aggregator 120 and the network interface 130, the computing power processing device 110 is connected with the network interface 130 and the cloud aggregator 120, and the network interface 130 and the cloud aggregator 120 are connected with at least two computing power clouds through a network. The computing power cloud is a cloud platform, and different computing power clouds can be used for carrying out different types of computing tasks or carrying out the same type of computing tasks. In fig. 1, computing force cloud a and computing force cloud B are two different computing force clouds.
The computing power processing device 110 communicates with the computing power cloud through the network interface 130, and if the computing power demand of any computing power cloud is detected while in the idle state, the computing power processing device joins the computing power resource pool corresponding to the computing power cloud. Wherein the computing power demand is a demand for adding a computing power machine; and if the computing power demand of any computing power cloud is detected, adding the computing power resource pool of the computing power cloud with the computing power demand. For example, if the computing power cloud A is detected to have the computing power requirement, adding the computing power cloud A into a computing power resource pool of the computing power cloud A. The idle state is a state in which no operation task is executed.
The cloud aggregator 120 receives computing tasks assigned by the computing power cloud and converts the computing tasks into local computing instructions that the computing power processing device 110 processes. Specifically, the cloud aggregator 120 receives the computing tasks allocated by the added computing power cloud, converts the computing tasks into instructions that can be processed by the computing power machine to obtain local computing instructions, and sends the local computing instructions to the computing power processing device 110, and the computing power processing device 110 processes the local computing instructions.
According to the power computer, the power computing cloud with power computing requirements is added in the idle state, the power computing cloud is connected through the cloud aggregator 120, the computing tasks sent by the power computing cloud are converted into the instructions which can be understood by the power computing machine, and the computing tasks of different power computing clouds can be processed by the power computing machine, so that one power computing machine can dynamically migrate among different power computing clouds, namely one power computing machine can be suitable for a plurality of power computing clouds, and resource waste is avoided.
In one embodiment, the computing force processing device 110 detects whether the connected computing force cloud has a computing force demand, including: judging whether an add power machine request of any one power cloud is received or not; if yes, determining that the computing power cloud for sending the request of adding the computing power machine has a computing power requirement; if not, sending an inquiry message to each connected computer, wherein the inquiry message is used for inquiring whether the computers are added; when any one of the computers receives the positive reply message returned according to the inquiry message, determining that the computer returning the positive reply message has a power demand; upon receipt of a negative reply message returned by all computers in accordance with the inquiry message, it is determined that there is no demand for computing power.
Wherein the affirmative reply message is a message indicating that there is an added computer demand; the negative reply message is a message indicating that the computer demand is not added. By actively querying each computing power cloud whether to add a computing power machine when no request to add a computing power machine is received, it is possible to more accurately determine whether there is a computing power demand.
Specifically, if a power computer request of a plurality of power clouds is received or a positive reply message of the plurality of power clouds is received, one of the power clouds is selected for joining according to a preset joining rule. For example, a computing power cloud may be selected randomly to join its computing power resource pool, or a computing power cloud that first received a computing power machine request/affirmative reply message may be selected to join.
In one embodiment, the computing power processing apparatus 110, while in the idle state, also interacts with the device characteristics and status information of both connected computing power clouds before detecting whether the connected computing power clouds have computing power requirements. And realizing synchronous information through the characteristics and state information of the interactive equipment. For example, taking the example of connecting the computing power cloud a and the computing power cloud B in fig. 1, the computing power processing device 110 interacts with the computing power cloud a and also interacts with the computing power cloud B.
In one embodiment, as shown in fig. 2, the computing power processing device 110 includes a processor 111 and a computing power board 112, where the processor 111 is connected to the computing power board 112 and the network interface 130, and the computing power board 112 is connected to the cloud aggregator 120. The processor 111 communicates with the computing power clouds via the network interface 130 and, while in an idle state, joins a pool of computing power resources for the corresponding computing power cloud if a computing power demand for any one of the computing power clouds is detected. The computing pad 112 receives and processes the local computing instructions sent by the cloud aggregator 120.
The computing board 112 is a computing component of a power machine. By adopting the processor 111 and the computing board 112 to perform processing and computation, respectively, the processing and computation efficiency is high.
In one embodiment, as shown in FIG. 3, the computing board 112 includes a controller and a plurality of computing chips, the controller connecting the processor 111, the cloud aggregator 120, and the plurality of computing chips; the controller receives the local computing instruction sent by the cloud aggregator 120, and sends the local computing instruction to the computing chip for processing.
Wherein the controller comprises an instruction execution unit; the power computing chip comprises an instruction cache unit, a data cache unit and a special operation unit. The instruction cache unit stores instructions obtained from the controller, the special operation unit extracts instruction execution from the instruction cache unit, and the data cache unit stores calculation data and calculation results. The special purpose arithmetic unit performs computationally intensive tasks including, but not limited to, matrix calculations (or multiply-accumulate, activate, vector operations, etc. involved in matrix calculations), encryption algorithms, hash algorithms, zero knowledge proof algorithms, etc.
The controller may receive code issued by the processor 111 (central processing unit) including instructions executed by a special purpose arithmetic unit of the computing power chip. The instruction execution unit sends the instruction to the instruction cache unit of the power computing chip, and the special operation unit runs the instruction.
In one embodiment, cloud aggregator 120 includes an access interface and a conversion module, where the access interface is coupled to the conversion module and the conversion module is coupled to computing power processing device 110 via a network. In particular, the conversion module may be a controller in the connection computing pad 112.
The access interface is used for setting a communication protocol between the power computer and the power cloud, and transmitting operation related data comprising instructions, original data and calculation results; management related data such as resource requests, login information, device status, computing power cloud status, payment information, etc. are transmitted.
The conversion module is used for realizing consistency and transparency of the computing power cloud to the computing power machine and converting and translating the interaction data between the access interfaces; such as converting different instruction formats of different computing power clouds into instruction formats which can be understood by the same computing power machine, or converting equipment states sent by the computing power machine into different data formats which meet the requirements of different computing power clouds.
In one embodiment, referring to fig. 4, cloud aggregator 120 further includes an account management module coupled to the conversion module. The account management module is used for account/login and payment management of the power computer in different power clouds.
In one embodiment, referring to fig. 4, cloud aggregator 120 further includes a monitoring module that connects the conversion module and computing power processing device 110. In particular, the monitoring module may be a controller in the connection computing pad 112.
The monitoring module is used for collecting the running state of the power machine, including but not limited to CPU/memory/stored equipment parameters, computing/storage resource utilization rate, equipment reservation plan, task type suitable for the power machine, special operation unit function, service price and the like; the computing power cloud may request computing power information from the computing power machine when computing resources need to be added, or report to the computing power cloud when a change in timing or status occurs by the present module. The monitoring module acquires a resource request of the computing power cloud, wherein the resource request comprises a time period of the computing power cloud requiring resources, a price willing to be provided, a service type of the computing power cloud, a required operation function type, network connection cost and the like.
In one embodiment, referring to fig. 4, cloud aggregator 120 further includes a policy module that connects the monitoring module, the account management module, and the conversion module.
The strategy module is used for analyzing by adopting an artificial intelligence/expert system/strategy database according to the information provided by the monitoring module, making a business strategy with the maximum benefit, and selecting the computing power cloud to be added, the adding time length and the like.
In one embodiment, the power computer may further include at least one of a memory and an input/output port, where the memory and the input/output port are connected to the power processing device 110. The computing force processing device 110 can send the data to be stored to the memory for storage, and can be connected with other devices through the input and output ports, so that the service performance is improved. Specifically, with continued reference to fig. 2, the memory 140 and the input/output 150 may be coupled to the processor 111.
To better illustrate the application of a computing machine, consider the example of a computing machine migrating between two computing clouds, as shown in FIG. 5. The computing power cloud A/B can be used for high-performance computing, deep neural network training, block chain consensus demonstration and the like. The computing force clouds A and B can be used for the same type of computing task as the same type of deep neural network training, or different types of computing tasks as the computing force cloud A is used for the deep neural network training and the computing force cloud B is used for the blockchain consensus proof. When two computing power clouds perform different types of computing tasks, computing components in a computing machine need to adapt to two or more applications to achieve a balance of performance and applicability. The deep neural network training computing component can perform special optimization for typical and high-computation-intensive computing tasks (such as multiply-accumulate, activation operation, vector operation and the like involved in matrix computation or matrix computation) in the neural network; the blockchain consensus proving power component optimizes various algorithms commonly used in blockchains, such as encryption algorithms, hash algorithms, zero knowledge proving algorithms, and the like.
When the computing power cloud A needs to increase computing power resources, adding a computing power machine into a computing power resource pool of the computing power cloud, distributing tasks to the computing power machine for computing, and paying consideration to an owner of the computing power machine according to the service condition of the computing power machine; after the task is finished, the computing power machine can request the computing power cloud to release equipment, and the computing power cloud removes the equipment from the computing power resource pool; the computing machine may request to join other computing cloud B or wait for a request for other computing cloud B. The computing machine has a unique identification associated with the physical device, such as a Globally Unique Identifier (GUID), or other locally unique or statistically likely unique identifier, such as a network card MAC (media access control) address, motherboard ID, CPU ID, BIOS number, etc.
In one embodiment, a processing method of the power computer is provided, which can be applied to a power computing processing device, as shown in fig. 6, and the method includes:
s110: while in the idle state, it is detected whether the connected computing power cloud has a computing power demand.
If there is an calculation power demand from the calculation power cloud, step S130 is executed; otherwise, step S110 is repeatedly performed.
S130: and adding a computing power resource pool of the computing power cloud with the computing power requirement.
S150: and acquiring a local computing instruction obtained by conversion after the cloud aggregator receives the computing tasks distributed by the added computing power cloud.
S170: processing the native computing instructions.
According to the processing method of the power computer, the power computing cloud with power computing requirements is added in the idle state, and the computing tasks sent by the power computing cloud are converted into the instructions which can be understood by the power computer through the cloud aggregator, so that the computing tasks of different power computing clouds can be processed by the power computer, one power computing machine can dynamically migrate among different power computing clouds, namely, one power computing machine can be suitable for a plurality of power computing clouds, and resource waste is avoided.
In one embodiment, the step of detecting whether the connected computing power cloud has a computing power demand comprises: judging whether an add power machine request of any one power cloud is received or not; if yes, determining that the computing power cloud for sending the request of adding the computing power machine has a computing power requirement; if not, sending an inquiry message to the connected power computer, wherein the inquiry message is used for inquiring whether the power computer is added; when any one of the computers receives the positive reply message returned according to the inquiry message, determining that the computer returning the positive reply message has a power demand; upon receipt of a negative reply message returned by all computers in accordance with the inquiry message, it is determined that there is no demand for computing power.
In one embodiment, before step S110, the processing method of the power machine further includes: and interacting the device characteristics and state information of both parties with the connected computing power cloud.
Describing one detailed embodiment, as shown in fig. 7, the computing machine interacts device characteristics and status information of two parties with at least two connected computing force clouds to synchronize information; when the computing machine detects that no computing task exists, namely in an idle state, judging whether a computing machine adding request of a computing cloud is received or not; if yes, requesting to join the computing power cloud, making a strategy, joining the computing power cloud, and then transmitting data/instructions for operation. If the request of adding the computing power machine of the computing power cloud is not received, querying the computing power cloud, inquiring whether the computing power cloud has the requirement of adding the computing power machine, if so, requesting to join the corresponding computing power cloud, and if not, re-synchronizing the information.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
Although the embodiments of the present invention are disclosed above, the embodiments are only used for the convenience of understanding the present invention, and are not intended to limit the present invention. Any person skilled in the art can make any modification and variation in form and detail without departing from the spirit and scope of the present disclosure, but the scope of the present disclosure is still subject to the scope of the present disclosure as defined by the appended claims.
Claims (8)
1. A power computer, comprising: the computing power system comprises a computing power processing device, a network interface and a cloud aggregator, wherein the computing power processing device is connected with the network interface and the cloud aggregator, and the network interface and the cloud aggregator are connected with at least two computing power clouds through a network;
the computing power processing device is communicated with the computing power clouds through the network interface, detects whether the connected computing power clouds have computing power demands or not when the computing power processing device is in an idle state, if the computing power demands of any one computing power cloud are detected, the computing power resource pools corresponding to the computing power clouds are added, the cloud aggregator receives computing tasks distributed by the computing power clouds and converts the computing tasks into computer computing instructions, and the computing power processing device processes the computer computing instructions; wherein the detecting whether the connected computing power cloud has a computing power demand comprises: judging whether an add power machine request of any one power cloud is received or not; if yes, determining that the computing power cloud sending the adding computing power machine request has a computing power requirement; if not, sending an inquiry message to each connected power computer, wherein the inquiry message is used for inquiring whether the power computers are added; when receiving a positive reply message returned by any one of the computers according to the inquiry message, determining that the computer returning the positive reply message has a calculation power requirement; when negative reply messages returned by all the power computers according to the inquiry messages are received, determining that no power calculation requirement exists;
the cloud aggregator comprises an access interface, a conversion module, a monitoring module and a strategy module; the access interface is connected with the conversion module and is connected with the computing power cloud through a network; the conversion module is connected with the computing force processing device; the monitoring module is connected with the conversion module and the computing power processing device and is used for collecting the running state information of the computing power machine and acquiring a computing power cloud resource request; the strategy module is connected with the monitoring module and the conversion module and is used for providing the running state information of the power machine and the resource request information of the power cloud according to the monitoring module, analyzing by adopting an artificial intelligence, expert system or strategy database, formulating a service strategy with the maximum benefit, and selecting the power cloud to be added and the adding time.
2. The power computer of claim 1, wherein the power processing device comprises a processor and a power board, the processor connecting the power board and the network interface, the power board connecting the cloud aggregator;
the processor communicates with the computing power clouds through the network interface, and if the computing power demand of any computing power cloud is detected when the processor is in an idle state, the processor adds a computing power resource pool corresponding to the computing power clouds;
and the power computing board receives and processes the local computing instruction sent by the cloud aggregator.
3. The power computer of claim 2, wherein the power board comprises a controller and a plurality of power chips, the controller connecting the processor, the cloud aggregator, and the plurality of power chips;
and the controller receives the local computing instruction sent by the cloud aggregator and sends the local computing instruction to the power computing chip for processing.
4. The computing machine of claim 1, wherein the cloud aggregator further comprises an account management module, the account management module coupled to the conversion module.
5. The computer of claim 4, wherein the policy module is further coupled to the account management module.
6. The power computer of claim 1, further comprising at least one of a memory and an input-output port, the memory and the input-output port being coupled to the power processing device.
7. A method of processing a power machine according to any one of claims 1 to 6, comprising:
detecting whether the connected computing power cloud has computing power requirements when the computing power cloud is in an idle state; wherein the detecting whether the connected computing power cloud has a computing power demand comprises: judging whether an add power machine request of any one power cloud is received or not; if yes, determining that the computing power cloud sending the adding computing power machine request has a computing power requirement; if not, sending an inquiry message to the connected power computer, wherein the inquiry message is used for inquiring whether the power computer is added; when receiving a positive reply message returned by any one of the computers according to the inquiry message, determining that the computer returning the positive reply message has a calculation power requirement; when negative reply messages returned by all the power computers according to the inquiry messages are received, determining that no power calculation requirement exists;
if yes, adding an computing power resource pool of the computing power cloud with computing power requirements; wherein, the computing power resource pool of the computing power cloud with computing power demand comprises: according to the operation state information of the power computers collected by the cloud aggregator and the acquired power cloud resource request information, analyzing by adopting an artificial intelligence, expert system or strategy database, formulating a service strategy with maximum benefit, and selecting the power cloud to be added and the adding time length so as to add a power resource pool of the power cloud with power demand;
acquiring a local computing instruction obtained by conversion after a cloud aggregator receives a computing task distributed by the added computing power cloud;
and processing the local computing instruction.
8. The method of claim 7, wherein the detecting, while in the idle state, whether the connected computing power cloud has a computing power demand is preceded by:
and interacting the device characteristics and state information of both parties with the connected computing power cloud.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210347088.9A CN114844888B (en) | 2022-04-01 | 2022-04-01 | Power computer and processing method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210347088.9A CN114844888B (en) | 2022-04-01 | 2022-04-01 | Power computer and processing method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114844888A CN114844888A (en) | 2022-08-02 |
| CN114844888B true CN114844888B (en) | 2024-01-19 |
Family
ID=82562999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210347088.9A Active CN114844888B (en) | 2022-04-01 | 2022-04-01 | Power computer and processing method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114844888B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109406025A (en) * | 2018-10-16 | 2019-03-01 | 武汉斗鱼网络科技有限公司 | Method for testing pressure, device and electronic equipment |
| CN110221913A (en) * | 2019-04-26 | 2019-09-10 | 深圳市致宸信息科技有限公司 | The cloud of monitoring server calculates method, terminal, equipment and the storage medium of power |
| CN112269981A (en) * | 2020-11-17 | 2021-01-26 | 深圳杰微芯片科技有限公司 | Block-based linkage computing power equipment data construction method, server and storage medium |
| CN112817751A (en) * | 2021-01-21 | 2021-05-18 | 浙江大华技术股份有限公司 | Computing resource management method and device, storage medium and electronic device |
| CN113037858A (en) * | 2021-03-23 | 2021-06-25 | 海南师范大学 | System and method for realizing sharing computing power between mobile phone and terminal |
| WO2022022584A1 (en) * | 2020-07-29 | 2022-02-03 | 展讯半导体(南京)有限公司 | Hashrate sharing method and related device |
| WO2022027224A1 (en) * | 2020-08-04 | 2022-02-10 | 北京大学深圳研究生院 | In-network computing power or resource service-oriented communication method |
| CN114070854A (en) * | 2021-11-26 | 2022-02-18 | 中国联合网络通信集团有限公司 | Computing power perception and routing method, system, equipment and medium in computing power network |
-
2022
- 2022-04-01 CN CN202210347088.9A patent/CN114844888B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109406025A (en) * | 2018-10-16 | 2019-03-01 | 武汉斗鱼网络科技有限公司 | Method for testing pressure, device and electronic equipment |
| CN110221913A (en) * | 2019-04-26 | 2019-09-10 | 深圳市致宸信息科技有限公司 | The cloud of monitoring server calculates method, terminal, equipment and the storage medium of power |
| WO2022022584A1 (en) * | 2020-07-29 | 2022-02-03 | 展讯半导体(南京)有限公司 | Hashrate sharing method and related device |
| WO2022027224A1 (en) * | 2020-08-04 | 2022-02-10 | 北京大学深圳研究生院 | In-network computing power or resource service-oriented communication method |
| CN112269981A (en) * | 2020-11-17 | 2021-01-26 | 深圳杰微芯片科技有限公司 | Block-based linkage computing power equipment data construction method, server and storage medium |
| CN112817751A (en) * | 2021-01-21 | 2021-05-18 | 浙江大华技术股份有限公司 | Computing resource management method and device, storage medium and electronic device |
| CN113037858A (en) * | 2021-03-23 | 2021-06-25 | 海南师范大学 | System and method for realizing sharing computing power between mobile phone and terminal |
| CN114070854A (en) * | 2021-11-26 | 2022-02-18 | 中国联合网络通信集团有限公司 | Computing power perception and routing method, system, equipment and medium in computing power network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114844888A (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021190482A1 (en) | Computing power processing network system and computing power processing method | |
| CN107087019B (en) | Task scheduling method and device based on end cloud cooperative computing architecture | |
| CN112799789A (en) | Node cluster management method, device, equipment and storage medium | |
| CN107133109B (en) | Method and device for communication between modules and computing equipment | |
| CN108063813B (en) | Method and system for parallelizing password service network in cluster environment | |
| US20230136612A1 (en) | Optimizing concurrent execution using networked processing units | |
| CN113032166A (en) | Inter-core communication method, processor, inter-core communication system, and computer-readable storage medium | |
| CN110545302A (en) | calculation migration method, equipment and storage medium | |
| CN115328645A (en) | Computing task scheduling method, computing task scheduling device and electronic equipment | |
| CN110737425B (en) | Method and device for establishing application program of charging platform system | |
| US20090132582A1 (en) | Processor-server hybrid system for processing data | |
| CN114844888B (en) | Power computer and processing method thereof | |
| CN111422078B (en) | Electric vehicle charging data allocation monitoring method based on block chain | |
| CN116402318A (en) | Multi-stage computing power resource distribution method and device for power distribution network and network architecture | |
| CN116192849A (en) | Heterogeneous accelerator card calculation method, device, equipment and medium | |
| KR101081932B1 (en) | Apparatus and method of load dispersion in multi-agent system | |
| CN107480302A (en) | A kind of loose coupling data integration synchronization realizing method based on enterprise-level application scene | |
| Chang et al. | An M2M computing model for improving the performance among devices | |
| Xiang et al. | Robust and cost-effective resource allocation for complex iot applications in edge-cloud collaboration | |
| CN115022403B (en) | Information processing system for multi-channel service agent | |
| CN114936098B (en) | Data transfer method, device, back-end equipment and storage medium | |
| Bansal et al. | An Android-based Application for Computation Offloading in Mobile Cloud Computing | |
| CN117076057B (en) | AI service request scheduling method, device, equipment and medium | |
| CN116954927B (en) | Distributed heterogeneous data acquisition method, storage medium and electronic equipment | |
| KR102642396B1 (en) | Batch scheduling device for deep learning inference model using limited gpu resources |
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 |