CN113609068A - Cloud service architecture based on hybrid heterogeneous processor - Google Patents
Cloud service architecture based on hybrid heterogeneous processor Download PDFInfo
- Publication number
- CN113609068A CN113609068A CN202110914305.3A CN202110914305A CN113609068A CN 113609068 A CN113609068 A CN 113609068A CN 202110914305 A CN202110914305 A CN 202110914305A CN 113609068 A CN113609068 A CN 113609068A
- Authority
- CN
- China
- Prior art keywords
- processing
- cloud service
- heterogeneous processor
- layer
- computing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/76—Architectures of general purpose stored program computers
- G06F15/78—Architectures of general purpose stored program computers comprising a single central processing unit
- G06F15/7807—System on chip, i.e. computer system on a single chip; System in package, i.e. computer system on one or more chips in a single package
- G06F15/7825—Globally asynchronous, locally synchronous, e.g. network on chip
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45595—Network integration; Enabling network access in virtual machine instances
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Multi Processors (AREA)
Abstract
The invention discloses a cloud service architecture based on a hybrid heterogeneous processor, which specifically comprises cloud service software and a hybrid heterogeneous processor hardware processing unit, wherein the cloud service software runs on the hybrid heterogeneous processor hardware processing unit; the hardware processing unit of the hybrid heterogeneous processor comprises a plurality of computing resource nodes; all the computing resource nodes are interconnected and intercommunicated through a high-speed network to form a computing network, and information among all the computing resource nodes is transmitted by adopting an interactive mechanism based on information; a computing network formed by each computing resource node adopts a full-network topology architecture based on cloud computing to realize real-time calling of the computing resource nodes; the hybrid heterogeneous processor hardware unit comprises a CPU processing logic subunit and an FPGA processing logic subunit. The cloud service architecture based on the hybrid heterogeneous processor meets the system requirements on the distributed computing function nodes in the cloud computing, and can be seamlessly incorporated into the cloud computing environment under the cooperation of internal function software.
Description
Technical Field
The invention relates to the field of cloud service architectures, in particular to a cloud service architecture based on a hybrid heterogeneous processor.
Background
In recent years, to meet the increasing demand for computing in the field, a heterogeneous computing system incorporating various acceleration components has been developed. The CPU + GPU heterogeneous system is the most common, and compared with the CPU used for traditional calculation, the CPU + GPU carrying structure is suitable for parallel calculation of a large amount of data, achieves breakthrough achievement in the fields of floating point calculation and parallel calculation and has huge promotion potential.
In a traditional computer, both a CPU and a GPU as a coprocessor belong to a von neumann structure, i.e., instruction decoding execution and memory sharing. However, in this structure, since the control logic of the instruction stream is complex, there are not many independent instruction streams, and it is too costly to develop a dedicated chip for each new problem, which brings a great challenge to the improvement of the computational performance of the hardware.
The FPGA has abundant computing resources, has the advantages of low power consumption, small time delay, large throughput and the like, and has higher computing efficiency than a CPU (central processing unit) and even a GPU (graphics processing unit), because the FPGA is an architecture without instructions and sharing a memory, the function of each logic unit is determined when reprogramming without instructions; secondly, compared with almost only data parallelism of the GPU, the FPGA has both pipeline parallelism and data parallelism, different pipelines can be set up according to the steps of processing data packets, and when the computation units of the GPU perform computation according to unified pace, the FPGA dispersedly finishes processing the computation process.
Therefore, in a hybrid heterogeneous system composed of general-purpose processor computing and special-purpose coprocessor computing, as an FPGA with more efficient data processing capability and infrastructure, first-generation heterogeneous computing with 'CPU + GPU' is continued, and second-generation heterogeneous computing with 'CPU + FPGA' is formed. The logic is that a CPU + FPGA heterogeneous processing unit is deployed in each user server and a basic front-end processing system, and data can be processed in advance more quickly and efficiently before reaching the user servers.
Disclosure of Invention
The invention discloses a cloud service architecture based on a hybrid heterogeneous processor, aiming at the problem that an image processing system is difficult to meet the real-time processing requirement when processing a large amount of data, and specifically comprising cloud service software and a hybrid heterogeneous processor hardware processing unit, wherein the cloud service software runs on the hybrid heterogeneous processor hardware processing unit; the hardware processing unit of the hybrid heterogeneous processor comprises a plurality of computing resource nodes;
all the computing resource nodes are interconnected and intercommunicated through a high-speed network to form a computing network, and information among all the computing resource nodes is transmitted by adopting an interactive mechanism based on information;
a computing network formed by each computing resource node adopts a full-network topology architecture based on cloud computing to realize real-time calling of the computing resource nodes; the cloud service software realizes each processing function by adopting micro-service, thereby completing real-time loading starting and light deployment of each processing function.
The computing network formed by all the computing resource nodes dynamically configures all the computing resource nodes according to the image processing task requirement and the load operation condition;
the hardware unit of the hybrid heterogeneous processor comprises a CPU processing logic subunit and an FPGA processing logic subunit, and the two subunits use cloud service software as a carrier to uniformly define and describe hardware processing resources of the two subunits, so that the cloud service software can uniformly call the hardware processing resources of the two subunits.
The CPU processing logic subunit is internally integrated with general CPU hardware processing resources and is used for finishing scheduling, circulation, branching and logic judgment instruction processing operations in an operating system, system software and a general application program;
the FPGA processing logic subunit is internally integrated with a special FPGA hardware processing resource and is used for finishing the data processing operation of high-speed bitwise operation, repeated operation steps and bit operation; the FPGA processing logic subunit dynamically develops data processing logic with a specific function corresponding to the characteristics according to the characteristics of the remote sensing satellite image data, and can dynamically load FPGA hardware processing resources.
The cloud service software comprises a Web front end, a webpage server, a business application layer, a bridging layer and a processing layer; the Web front end is connected with a Web server, the Web server is connected with a service application layer, the service application layer is connected with a bridging layer, and the bridging layer is connected with a processing layer. Between the business application layer and the processing layer, the bridging layer is used as a message transmission channel, parameters of each module in the processing layer are sent to the business application layer through the bridging layer, and the processing layer is used for sending starting and stopping commands, collecting state information and the like; the layers send specific messages to each other to handle specific transactions. The bridge layer not only provides message service for the business application layer and the processing layer, but also serves as a bridge to transmit messages in the business application layer and among all modules of the business application layer. The Web front end is used for realizing the interaction between the user and the platform; the web server is used for providing information browsing service for the platform for the user.
The computing network formed by each computing resource Node is realized by adopting a Kubernets cluster, the cloud service function of a cloud service architecture based on the hybrid heterogeneous processor is realized by deploying a management tool on the Kubernets cluster, each link in the image processing process is mapped into a corresponding software function module which comprises a data receiving module, a decompression function module, a format analysis function module, a data sequencing function module and a front-end and back-end interface function module, meanwhile, a hardware unit of the hybrid heterogeneous processor is abstracted into a class of software function modules, the cloud service software comprises all the software function modules, all the software function modules are packaged into each Docker container, and each Docker container is grouped and configured into Node nodes in the Kubernets cluster by adopting the Kubernets cluster deployment management tool. Each Pod of the Kubernetes cluster is constructed into a micro-service, and the Pod is the minimum unit created and deployed in the Kubernetes cluster; each micro service comprises heterogeneous processing software, a plurality of micro services are connected in series to form a service processing chain, and data are processed step by step; the Kubernetes cluster dynamically establishes a service processing chain; and the management tool deployed on the Kubernetes cluster dynamically manages and calls the micro-service formed by the Pod.
The invention has the beneficial effects that:
the cloud service architecture based on the hybrid heterogeneous processor completely meets the system requirements on the distributed computing function nodes in the cloud computing, and can be completely and seamlessly brought into the cloud computing environment under the coordination of internal function software.
The traditional method has high hardware development threshold, high cost and long time, and can consume a great deal of energy, and the cloud service mode adopted by the invention can provide services with high added value and can focus on business integration and solution. The method saves a plurality of links such as hardware development board research and development and test while enjoying performance advantages.
Drawings
Fig. 1 is a schematic diagram of a hybrid heterogeneous processor cloud service architecture according to the present invention.
FIG. 2 is a schematic diagram of a message-driven mechanism according to the present invention.
Detailed Description
For a better understanding of the present disclosure, an example is given here.
The invention discloses a cloud service architecture based on a hybrid heterogeneous processor, wherein FIG. 1 is a schematic diagram of the cloud service architecture of the hybrid heterogeneous processor, FIG. 2 is a schematic diagram of a message driving mechanism of the invention, and the cloud service architecture specifically comprises cloud service software and a hardware processing unit of the hybrid heterogeneous processor, wherein the cloud service software runs on the hardware processing unit of the hybrid heterogeneous processor; the hardware processing unit of the hybrid heterogeneous processor comprises a plurality of computing resource nodes;
the method comprises the following steps that all computing resource nodes are interconnected and intercommunicated through a high-speed network to form a computing network, information transmission among all computing resource nodes is realized, and a message-based interaction mechanism is adopted, so that resource link establishment and release delay is optimized to millisecond level when information transmission is carried out;
a computing network formed by each computing resource node adopts a full-network topology architecture based on cloud computing to realize real-time calling of the computing resource nodes; the cloud service software realizes each processing function by adopting micro-service, thereby completing real-time loading starting and light deployment of each processing function, and supporting rapid development of a function module and rapid loading of a service function.
The computing network formed by all the computing resource nodes dynamically configures all the computing resource nodes according to the image processing task requirement and the load operation condition;
the hardware unit of the hybrid heterogeneous processor comprises a CPU processing logic subunit and an FPGA processing logic subunit, and the two subunits uniformly define and describe the hardware processing resources of the two subunits by using cloud service software as a carrier, so that the cloud service software can uniformly call the hardware processing resources of the two subunits, and the two units are presented as an independent heterogeneous processing software unit.
The CPU processing logic subunit is internally integrated with general CPU hardware processing resources and is used for finishing scheduling, circulation, branching and logic judgment instruction processing operations in an operating system, system software and a general application program;
the FPGA processing logic subunit is internally integrated with a special FPGA hardware processing resource and is used for finishing the data processing operation of high-speed bitwise operation, repeated operation steps and bit operation; the FPGA can flexibly customize pipeline parallelism and data parallelism, has the characteristics of high execution efficiency, large data throughput and the like, and has far higher performance than a CPU when processing specific operation.
The code stream format of the data mentioned in the architectural design has the characteristics of stream type, blocking, segmenting and layer-by-layer packaging, and is very suitable for chain type step-by-step processing, but the requirement on the speed of each stage is very high because of large data volume and high real-time requirement, and the heterogeneous processing unit in the invention can completely meet the requirement on the aspect of single-stage data processing performance.
The FPGA processing logic subunit dynamically develops data processing logic with specific functions corresponding to the characteristics according to the characteristics of the remote sensing satellite image data, can dynamically and logically load FPGA hardware processing resources, and has strong expandability and maintainability. The characteristics of the remote sensing satellite image data refer to the characteristics of large data volume and high processing real-time requirement of the remote sensing satellite image.
Referring to fig. 2, an interaction mechanism based on messages is adopted among the computing resource nodes, and resource link establishment and release delay is optimized to millisecond level; the cloud service software comprises a Web front end, a webpage server (WebServer layer), a service application layer, a bridging layer and a processing layer; the Web front end is connected with a Web server, the Web server is connected with a service application layer, the service application layer is connected with a bridging layer, and the bridging layer is connected with a processing layer. Between the business application layer and the processing layer, the bridging layer is used as a message transmission channel, parameters of each module in the processing layer are sent to the business application layer through the bridging layer, the processing layer is used for sending starting and stopping commands and collecting state information, and the two layers are not directly interconnected; the layers send specific messages to each other to handle specific transactions. The bridge layer not only provides message service for the business application layer and the processing layer, but also serves as a bridge to transmit messages in the business application layer and among all modules of the business application layer. The Web front end is used for realizing the interaction between the user and the platform; the web server is used for providing information browsing service for the platform for the user.
The computing network formed by the computing resource nodes is realized by a Kubernets cluster, and the cloud service function of the cloud service architecture based on the hybrid heterogeneous processor is realized by deploying a management tool on the Kubernets cluster, it maps each link in the image processing process into a corresponding software functional module, which comprises a data receiving module, a decompression functional module, a format analysis functional module, a data sorting functional module and a front-back interface functional module, meanwhile, the hardware unit of the hybrid heterogeneous processor is abstracted into a class of software functional modules, the cloud service software comprises all the software functional modules, all the software function modules are packaged into each Docker container, and each Docker container is grouped and configured into Node nodes in a Kubernetes cluster by adopting a Kubernetes cluster deployment management tool, so that the centralized management and the uniform deployment of the upgraded system are realized. Each Pod of the Kubernetes cluster is constructed into a micro-service, and the Pod is the minimum unit created and deployed in the Kubernetes cluster; each micro service comprises heterogeneous processing software, a plurality of micro services are connected in series to form a service processing chain, and data are processed step by step; the Kubernetes cluster dynamically establishes a service processing chain, and certain link equipment cannot work due to the failure of a single component node; and the management tool deployed on the Kubernetes cluster dynamically manages and calls the micro-service formed by the Pod.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (5)
1. A cloud service architecture based on a hybrid heterogeneous processor is characterized by specifically comprising cloud service software and a hybrid heterogeneous processor hardware processing unit, wherein the cloud service software runs on the hybrid heterogeneous processor hardware processing unit; the hardware processing unit of the hybrid heterogeneous processor comprises a plurality of computing resource nodes;
all the computing resource nodes are interconnected and intercommunicated through a high-speed network to form a computing network, and information among all the computing resource nodes is transmitted by adopting an interactive mechanism based on information;
a computing network formed by each computing resource node adopts a full-network topology architecture based on cloud computing to realize real-time calling of the computing resource nodes; the cloud service software realizes each processing function by adopting micro-service, so that the real-time loading starting and the light deployment of each processing function are completed;
the computing network formed by all the computing resource nodes dynamically configures all the computing resource nodes according to the image processing task requirement and the load operation condition;
the hardware unit of the hybrid heterogeneous processor comprises a CPU processing logic subunit and an FPGA processing logic subunit, and the two subunits use cloud service software as a carrier to uniformly define and describe hardware processing resources of the two subunits, so that the cloud service software can uniformly call the hardware processing resources of the two subunits.
2. The hybrid heterogeneous processor-based cloud service architecture of claim 1, wherein the CPU processing logic sub-unit is internally integrated with general CPU hardware processing resources for performing scheduling, looping, branching, and logic judgment instruction class processing operations in an operating system, system software, and general application programs.
3. The hybrid heterogeneous processor-based cloud service architecture of claim 1, wherein the FPGA processing logic subunit has dedicated FPGA hardware processing resources integrated therein for performing high-speed bitwise operations, operation step repetition, and bit operation-like data class processing operations; the FPGA processing logic subunit dynamically develops data processing logic with a specific function corresponding to the characteristics according to the characteristics of the remote sensing satellite image data, and can dynamically load FPGA hardware processing resources.
4. The hybrid heterogeneous processor-based cloud service architecture of claim 1,
the cloud service software comprises a Web front end, a webpage server, a business application layer, a bridging layer and a processing layer; the Web front end is connected with a Web server, the Web server is connected with a service application layer, the service application layer is connected with a bridging layer, and the bridging layer is connected with a processing layer; between the business application layer and the processing layer, the bridging layer is used as a message transmission channel, parameters of each module in the processing layer are sent to the business application layer through the bridging layer, and the processing layer is used for sending start and stop commands and collecting state information; sending specific messages to each other among layers to process specific transactions; the bridge layer not only provides message service for the business application layer and the processing layer, but also serves as a bridge to transmit messages in the business application layer and among all modules of the business application layer; the Web front end is used for realizing the interaction between the user and the platform; the web server is used for providing information browsing service for the platform for the user.
5. The hybrid heterogeneous processor-based cloud service architecture of claim 1,
the computing network formed by each computing resource Node is realized by adopting a Kubernets cluster, and the cloud service function of a cloud service architecture based on the hybrid heterogeneous processor is realized by deploying a management tool on the Kubernets cluster, each link in the image processing process is mapped into a corresponding software function module which comprises a data receiving module, a decompression function module, a format analysis function module, a data sequencing function module and a front-end and back-end interface function module, meanwhile, a hardware unit of the hybrid heterogeneous processor is abstracted into a class of software function modules, the cloud service software comprises all the software function modules, all the software function modules are packaged into each Docker container, and each Docker container is grouped and configured into Node nodes in the Kubernets cluster by adopting the Kubernets cluster deployment management tool; each Pod of the Kubernetes cluster is constructed into a micro-service, and the Pod is the minimum unit created and deployed in the Kubernetes cluster; each micro service comprises heterogeneous processing software, a plurality of micro services are connected in series to form a service processing chain, and data are processed step by step; the Kubernetes cluster dynamically establishes a service processing chain; and the management tool deployed on the Kubernetes cluster dynamically manages and calls the micro-service formed by the Pod.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110914305.3A CN113609068A (en) | 2021-08-10 | 2021-08-10 | Cloud service architecture based on hybrid heterogeneous processor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110914305.3A CN113609068A (en) | 2021-08-10 | 2021-08-10 | Cloud service architecture based on hybrid heterogeneous processor |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113609068A true CN113609068A (en) | 2021-11-05 |
Family
ID=78308012
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110914305.3A Pending CN113609068A (en) | 2021-08-10 | 2021-08-10 | Cloud service architecture based on hybrid heterogeneous processor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113609068A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114281521A (en) * | 2021-11-21 | 2022-04-05 | 苏州浪潮智能科技有限公司 | Method, system, device and medium for optimizing communication efficiency of deep learning heterogeneous resources |
CN114500222A (en) * | 2021-12-30 | 2022-05-13 | 中智行(上海)交通科技有限公司 | Deployment mode of vehicle-road cooperation field network and computing architecture |
CN115766865A (en) * | 2023-01-09 | 2023-03-07 | 利尔达科技集团股份有限公司 | Application service method and system for 5G intelligent terminal heterogeneous platform |
CN116361120A (en) * | 2023-05-31 | 2023-06-30 | 山东浪潮科学研究院有限公司 | Method, device, equipment and medium for managing and scheduling heterogeneous resources of database |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108829515A (en) * | 2018-05-29 | 2018-11-16 | 中国科学院计算技术研究所 | A kind of cloud platform computing system and its application method |
US20200228388A1 (en) * | 2020-03-27 | 2020-07-16 | Intel Corporation | Programmable integrated circuit configured as a remote trust anchor to support multitenancy |
CN111488051A (en) * | 2020-03-06 | 2020-08-04 | 复旦大学 | Cloud deep neural network optimization method based on CPU and FPGA cooperative computing |
CN111736988A (en) * | 2020-05-29 | 2020-10-02 | 浪潮电子信息产业股份有限公司 | Heterogeneous acceleration method, equipment and device and computer readable storage medium |
CN112000421A (en) * | 2020-07-15 | 2020-11-27 | 北京计算机技术及应用研究所 | Management scheduling technology based on super-fusion architecture |
CN112653728A (en) * | 2020-12-07 | 2021-04-13 | 同济大学 | Greenhouse environment control Internet of things system based on edge cloud cooperation |
-
2021
- 2021-08-10 CN CN202110914305.3A patent/CN113609068A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108829515A (en) * | 2018-05-29 | 2018-11-16 | 中国科学院计算技术研究所 | A kind of cloud platform computing system and its application method |
CN111488051A (en) * | 2020-03-06 | 2020-08-04 | 复旦大学 | Cloud deep neural network optimization method based on CPU and FPGA cooperative computing |
US20200228388A1 (en) * | 2020-03-27 | 2020-07-16 | Intel Corporation | Programmable integrated circuit configured as a remote trust anchor to support multitenancy |
CN111736988A (en) * | 2020-05-29 | 2020-10-02 | 浪潮电子信息产业股份有限公司 | Heterogeneous acceleration method, equipment and device and computer readable storage medium |
CN112000421A (en) * | 2020-07-15 | 2020-11-27 | 北京计算机技术及应用研究所 | Management scheduling technology based on super-fusion architecture |
CN112653728A (en) * | 2020-12-07 | 2021-04-13 | 同济大学 | Greenhouse environment control Internet of things system based on edge cloud cooperation |
Non-Patent Citations (2)
Title |
---|
伊恩·文斯·麦克洛克林: "《计算机***》", 30 June 2020, 机械工业出版社 * |
吕晓永: "《互联网金融》", 30 June 2018 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114281521A (en) * | 2021-11-21 | 2022-04-05 | 苏州浪潮智能科技有限公司 | Method, system, device and medium for optimizing communication efficiency of deep learning heterogeneous resources |
CN114281521B (en) * | 2021-11-21 | 2024-01-09 | 苏州浪潮智能科技有限公司 | Method, system, equipment and medium for optimizing deep learning heterogeneous resource communication efficiency |
CN114500222A (en) * | 2021-12-30 | 2022-05-13 | 中智行(上海)交通科技有限公司 | Deployment mode of vehicle-road cooperation field network and computing architecture |
CN114500222B (en) * | 2021-12-30 | 2024-06-07 | 中国电信股份有限公司 | Deployment mode of network and computing architecture in vehicle-road cooperation field and application thereof |
CN115766865A (en) * | 2023-01-09 | 2023-03-07 | 利尔达科技集团股份有限公司 | Application service method and system for 5G intelligent terminal heterogeneous platform |
CN116361120A (en) * | 2023-05-31 | 2023-06-30 | 山东浪潮科学研究院有限公司 | Method, device, equipment and medium for managing and scheduling heterogeneous resources of database |
CN116361120B (en) * | 2023-05-31 | 2023-08-15 | 山东浪潮科学研究院有限公司 | Method, device, equipment and medium for managing and scheduling heterogeneous resources of database |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113609068A (en) | Cloud service architecture based on hybrid heterogeneous processor | |
CN110119311B (en) | Distributed stream computing system acceleration method based on FPGA | |
US20060026169A1 (en) | Communication method with reduced response time in a distributed data processing system | |
CN103488775A (en) | Computing system and computing method for big data processing | |
CN111182239B (en) | AI video processing method and device | |
CN110515889B (en) | Embedded FPGA cluster intelligent computing platform hardware framework | |
Doyu et al. | A tinymlaas ecosystem for machine learning in iot: Overview and research challenges | |
CN110083454B (en) | Hybrid cloud service arrangement method with quantum computer | |
CN112631986B (en) | Large-scale DSP parallel computing device | |
CN111880911A (en) | Task load scheduling method, device and equipment and readable storage medium | |
Yin et al. | Scalable mapreduce framework on fpga accelerated commodity hardware | |
Wu et al. | Accelerating DNNs from local to virtualized FPGA in the Cloud: A survey of trends | |
CN104360962B (en) | Be matched with multistage nested data transmission method and the system of high-performance computer structure | |
Bosilca et al. | On scalability for mpi runtime systems | |
CN111459648B (en) | Heterogeneous multi-core platform resource optimization method and device for application program | |
CN111274018A (en) | Distributed training method based on DL framework | |
CN110502337B (en) | Optimization system for shuffling stage in Hadoop MapReduce | |
Xiao et al. | FPGA-based edge computing: Task modeling for cloud-edge collaboration | |
Sugiarto et al. | Optimized task graph mapping on a many-core neuromorphic supercomputer | |
Ueno et al. | VCSN: Virtual circuit-switching network for flexible and simple-to-operate communication in HPC FPGA cluster | |
CN114640681A (en) | Data processing method and system | |
CN114020618A (en) | High-availability test method and system based on FPGA and DPDK | |
CN110769037A (en) | Resource allocation method for embedded edge computing platform | |
Lee et al. | On-demand remote software code execution unit using on-chip flash memory cloudification for IoT environment acceleration | |
CN110380991A (en) | A kind of IOCP mechanism and the Internet of Things Network Communication acceleration system based on eFPGA and IOCP |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211105 |
|
RJ01 | Rejection of invention patent application after publication |