CN113839878B - Network-on-chip approximate communication system for data intensive application - Google Patents
Network-on-chip approximate communication system for data intensive application Download PDFInfo
- Publication number
- CN113839878B CN113839878B CN202111126409.4A CN202111126409A CN113839878B CN 113839878 B CN113839878 B CN 113839878B CN 202111126409 A CN202111126409 A CN 202111126409A CN 113839878 B CN113839878 B CN 113839878B
- Authority
- CN
- China
- Prior art keywords
- network
- data
- congestion
- chip
- router
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/02—Topology update or discovery
- H04L45/028—Dynamic adaptation of the update intervals, e.g. event-triggered updates
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/32—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/38—Flow control; Congestion control by adapting coding or compression rate
-
- 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
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a network-on-chip approximate communication system for data intensive application, and belongs to the technical field of network-on-chip. Aiming at the problems that the network on a chip is easy to be congested and delay is generated in the data-intensive application in the prior art, the invention adopts a data transmission mode of approximate communication; a global controller is arranged in the network on chip; a supervision unit of network congestion status is arranged in the router; the router and the processing core are connected with a port, and a data compression and decompression unit and a data screening unit are arranged in the port. The invention can effectively improve the communication congestion condition of the network on chip and reduce the communication delay of data intensive application by supervising the network congestion condition and a dynamic data approximation scheme, and has good use value and wide application prospect.
Description
Technical Field
The present invention relates to the field of Network on chip (NoC) technology, and more particularly, to a Network on chip proximity communication system for data intensive applications.
Background
With the continuous decrease of the size of semiconductor processes, the industry starts to enter the "dark silicon" era, and the performance of single processing cores gradually tends to be limited, so that multi-core has become a hot spot of chip industry research and a main development direction in the future. Compared with a bus structure, the multi-core communication architecture connected through the Network on Chip has the advantages of high bandwidth, good expandability, low delay and the like, and the Network on Chip (NoC) has become a normal form structure of multi-core interconnection, but at the same time, the multi-core communication architecture also means that more time is spent on data communication.
Currently, some large multi-core chips generally have hundreds of cores, and the number of processing cores of future chips is predicted to be more; as the number of processing cores increases, the communication time costs of several representative mega-level parallel applications are rapidly increasing, that is, the communication problem is rapidly becoming a major bottleneck for applications with extreme parallelism.
However, some data-intensive applications have fault-tolerant properties, such as image processing and scientific computing, which provide designers with new ideas in attempting to solve network-on-chip communication congestion problems. These applications can tolerate moderate errors while producing user acceptable results. Such as an image processing application, the differences are not apparent to the end user even if the output is not one hundred percent correct. Thus, conventional NoC designs that transmit all data with absolute accuracy are not the only choice for these applications.
The Chinese patent application, application number CN202010765578.1, 12 months and 18 days of publication 2020 discloses a method for optimizing energy consumption and performance of a many-core system based on collaborative approximation calculation, which can combine different abstract layers to adopt various approximation technologies on the premise of controlling an output result of an application program to meet a certain error range, comprises reducing the calculation workload of the application program at an application level, selectively deleting data at a network layer to reduce network congestion, and applying approximation calculation to different abstract layers of the many-core system through optimization regulation and control of a global controller and resource allocation of a local controller. According to the method, importance of discarded data is measured based on a quality model, collaborative management of communication and calculation is considered, a multi-objective optimization problem is formulated, network congestion is minimized, running time of an application program is limited, result quality is limited, a novel method for accelerating the running time of the application program, reducing energy consumption and improving chip-level energy efficiency is provided for a many-core system, all approximatable data in a network are cut and replied according to the application quality, network congestion is not considered when the scheme is used for data compression, unnecessary precision loss is caused when the network is smooth, and better effects cannot be brought under the condition of network congestion.
Disclosure of Invention
1. Technical problem to be solved
Aiming at the problems that the network-on-chip is easy to congestion and delay is generated in the data-intensive application in the prior art, the invention provides a network-on-chip approximate communication system for the data-intensive application, which can effectively improve the congestion condition of the network-on-chip, dynamically reduce the communication delay of the data-intensive application, and adapt to the topological structure of a two-dimensional network and a three-dimensional network and a plurality of dynamic routing algorithms.
2. Technical proposal
The invention discloses a scheme capable of improving communication bandwidth of a network-on-chip and reducing delay and power consumption generated by data-intensive application in communication of the network-on-chip.
The aim of the invention is achieved by the following technical scheme.
A network-on-chip approximate communication system facing data intensive application, the system network comprises a plurality of approximate communication architectures, each approximate communication architecture comprises a processing core, a router and a network interface, and the network interface is respectively connected with the processing core and the router;
a main control node is arranged in the processing core, and a global controller is arranged in the main control node; the router is provided with a network congestion condition supervision unit which is used for transmitting congestion information of the network to the global controller in real time; the network interface is provided with a data screening unit and a data compression and decompression unit, and system data is approximately processed by the data screening unit and the data compression and decompression unit in the network port and then transmitted to the network through the router.
Preferably, the global controller determines congestion status of network communication according to latency accumulated by each router in the regional network, wherein the regional network comprises all router paths possibly passed by data in a packet source node and a destination node of the network on chip.
Preferably, the global controller counts the congestion amount of router nodes in the area, and sends different data approximation rate information to the source node sending packets according to the congestion amount.
Preferably, the network congestion status monitoring unit converts the communication congestion status of the network into router waiting time, judges the network congestion status according to the waiting time, and sends congestion node information to the global controller.
Preferably, the network congestion status monitoring unit records the phenomenon that a plurality of ports compete for the same output port in the network topology, and counts the times that the input port does not acquire the transmission priority into the waiting time; recording the number of times that the input port does not acquire the transmission right in the waiting time when the data back pressure signal is zero, namely when the output port matched with the current input port has no idle virtual channel; and defining the router node whose waiting time exceeds the set threshold as a congestion node.
Preferably, congestion node information transmitted by the network congestion status supervision unit is transmitted to the global controller to match the corresponding data compression rate, and meanwhile, the original waiting time of the network congestion status supervision unit is cleared.
The invention adopts the network congestion supervision unit in the router to convert the communication competition congestion and the data back pressure signal into the waiting time, and effectively monitors the communication congestion degree of the network on chip as judging the congestion condition of the network area.
Preferably, the data screening unit steplessly adjusts the data compression rate transmitted to the network within an allowed approximation rate threshold. The method can well adapt to the change of the network congestion condition.
Preferably, the data filtering unit generates a pseudo random number through a linear feedback shift register to filter the data to be compressed, and the data packet without the acquired transmission right is compressed.
Preferably, the communication architecture employs a data transmission mode of approximate communication.
Preferably, the network is a two-dimensional or three-dimensional network. The global controller takes the number of the congestion nodes in the specific area as the basis for judging the network communication congestion condition, can support various network-on-chip dynamic routing algorithms, and is suitable for data-intensive applications running in two-dimensional and three-dimensional network topologies.
The communication system analyzes the communication competition condition and the data back pressure signal in the router, converts the communication congestion condition of the network into the waiting time of the router through the network congestion monitoring unit, transmits the information of the network congestion node to the global controller through the network, and dynamically compresses the data in real time according to the network congestion condition. The congestion state of network communication is judged by the accumulated waiting time of the router of the global controller, the data compression rate corresponding to the congestion information in the network area on chip is generated according to the congestion information, and different data approximation rate information is sent to the source node sending packets according to the number of the congestion nodes.
3. Advantageous effects
Compared with the prior art, the invention has the advantages that:
(1) The invention adopts the network congestion supervision unit in the router to convert the communication competition congestion and the data back pressure signal into the waiting time, and is used for judging the congestion condition of the network area, thereby effectively monitoring the communication congestion degree of the network on chip;
(2) The data to be compressed is screened by generating pseudo random numbers through the linear feedback shift register, and the compression rate of the data is steplessly adjusted within an error threshold value allowed by application, so that the method can be well adapted to the change of network congestion conditions;
(3) The global controller takes the number of congestion nodes in a specific area as the basis for judging the network communication congestion condition, and dynamically compresses data in real time according to the network congestion condition, so that a plurality of network-on-chip dynamic routing algorithms can be supported, and the method is suitable for data-intensive applications running in two-dimensional and three-dimensional network topologies;
(4) The method is oriented to any approximatable application, does not need to run the application to obtain a model, supports dynamic data approximation, can realize any data compression rate in a single router, automatically generates different compression forces for different applications and network conditions according to network conditions, and has wider applicability;
in summary, the invention is oriented to the congestion situation of data intensive application in the network, and can better fit the practical delay curve according to the dynamic reduced data precision of the congestion situation in the network. The method can be used for directly controlling the application which does not run before the network without obtaining a performance model, and has good use value and wide application prospect.
Drawings
FIG. 1 is a schematic cross-sectional view of a three-dimensional network topology;
FIG. 2 is a schematic diagram of a multi-core dynamic approximation communication architecture of the present invention;
FIG. 3 is a schematic diagram of a network congestion supervision unit of the present invention;
FIG. 4 is a schematic diagram of the control logic of the present invention;
FIG. 5 is a schematic diagram of the integrated flow pattern simulation delay results;
FIG. 6 is a schematic diagram of the results of integrated flow mode simulation power consumption.
Detailed Description
The invention will now be described in detail with reference to the drawings and the accompanying specific examples.
Examples
The invention discloses a dynamic approximate communication system of network on chip for data intensive application, which adopts a data transmission mode of approximate communication, and a global controller of data approximate rate is arranged in the network; a network congestion condition supervision unit is arranged in the router; the method comprises the steps that an on-chip network data compression and decompression unit and a screening unit for generating dynamic data compression probability are arranged in a port connected with a router and a processing core; the architecture analyzes communication competition conditions and data back pressure signals in a router, converts communication congestion conditions of a network into router waiting time through a monitoring unit, judges the network congestion conditions according to the waiting time, and sends congestion node information to a global controller. .
The data screening unit screens the approximatable data applied to the incoming network by generating a pseudo random number for the register through an 8-bit linear feedback, so that the data compression rate can be steplessly adjusted within an allowable maximum data approximation rate threshold. The control logic of the global controller judges the congestion status of network communication through the accumulated waiting time of each router in the area, and the area is formed by all possible routing paths of data in the packet source node and the destination node of the network on chip.
The network congestion status supervision unit in the router records the phenomenon that a plurality of ports compete for the same output port in a two-dimensional or three-dimensional network topology, and counts the times that the input port does not acquire transmission priority into waiting time; recording the number of times that the input port does not acquire the transmission right in the waiting time when the data back pressure signal is zero, namely when the output port matched with the current input port has no idle virtual channel; and defining the router node whose waiting time exceeds the set threshold as a congestion node.
In order to steplessly adjust the data compression rate of the incoming network within the error threshold allowed by the application, a hardware-implementable pseudo-random number is generated in the port by a linear shift feedback register and compared with the dynamic network data approximation rate fed back by the global controller, all approximable data are screened by the linear shift feedback register, and the data packet which does not win the transmission right is compressed, so that the traffic of the incoming network is dynamically controlled in real time.
The control logic of the global controller generates a data compression rate corresponding to the congestion information in the network-on-chip area according to the congestion information in the adjustable unit time interval, and the area is formed by all paths possibly passing through between a data packet source node and a destination node through a dynamic routing algorithm; the global controller counts the congestion quantity of router nodes in the area, and sends different data approximation rate information to the source node sending packets according to the congestion quantity.
As shown in fig. 1, which is a single-sided cross-sectional view of a three-dimensional network topology, the network-on-chip in this embodiment is a three-dimensional mesh structure, where core represents a processing core, router represents a router, NI (network interface) denotes a network interface, the three-dimensional network structure size in this embodiment is 4 x 4, each network node includes an approximate communication architecture as shown in fig. 2; the buffer area of the router is used for storing flits ready for transmission, in this embodiment, each buffer area is 8 flits, the virtual channel of each input port is 1, and the virtual channel can better utilize the link bandwidth for the physical link of the time division multiplexing network on chip, so as to avoid the condition that the link is continuously occupied.
The dynamic approximate communication architecture disclosed in this embodiment is shown in fig. 2, where the processing core includes a master control node, and the master control node is provided with a global controller; the Router in the system is provided with a network congestion condition monitoring unit, and the network congestion condition monitoring unit accumulates the delay into the waiting time according to the communication competition condition affecting the network communication delay and the condition that the input port cannot acquire the output right caused by the occupied output port, and defines a Router node with the accumulated waiting time of more than 5 in every 2000 clock cycles as a congestion node; the congestion node information is then transmitted to the global controller to match the corresponding data compression rate, and the original waiting time of the supervision unit is cleared.
The port NI (network interface) between the processing core and the Router is provided with a data compression and decompression unit and a data screening unit. The input data of the input network is firstly passed through a data screening unit, the data screening unit compares the pseudo random number generated by the linear shift feedback register with the data compression rate sent by the full-area controller, for example, the data compression rate sent by the global controller is 0.2, the random number between 0 and 1 generated by the linear feedback shift register is randomly compared with 0.2, when the data compression rate is smaller than the data compression rate, the data packet transmitted by the application is transmitted to a compression unit, and otherwise, the data packet is transmitted to the input network. When the data in the network reaches the port, the data is restored by a linear interpolation method through a decompression unit in the port, and then is transmitted into the processing core.
The specific structure of the network congestion status supervision unit is shown in fig. 3, and the accumulator statistics result in two cases of delay in transmission. The first case is that the tail end of the multi-directional input port buffer zone competes for the same output port, and the times of not obtaining the priority of the output port are accumulated into waiting time; the second case is that the output port backpressure signal to be sent to the end of the input port buffer is zero, i.e. the output port has no free buffer to hold data, and the number of data packet stalls is accumulated into the waiting time.
As shown in fig. 4, the control logic of the global controller is to count the number of congestion router nodes in an area formed by all paths possibly passing through in the network, wherein the area possibly passing through by the dynamic routing algorithm of all shortest paths is formed into a cube in the three-dimensional network, the approximation rate of network data linearly rises along with the number of congestion nodes until the maximum allowable error threshold is reached, and then the approximation rate of the data is maintained.
To verify the effect of the system to improve the data congestion condition, the communication architecture is simulated on a nonxim simulation platform. The simulation uses a 4 x 4 network topology, with a buffer depth of 8 flits, the number of virtual channels of each input port is 1, and the size of a data packet transmitted in the network is 8 flits. The routing algorithm adopted by the simulation comprises an XYZ reason algorithm and an OE_Z routing algorithm, and traffic patterns for comparison comprise Random, transfer and Shuffle. The simulation mainly verifies that the dynamic approximate communication architecture can improve the congestion condition of the network to reduce delay when the injection rate reaches saturation, and has good adaptability to a dynamic routing algorithm OE_Z.
FIG. 5 shows the delay results of the integrated traffic simulation, in which XYZ_original and OE_Z_original represent the delay results of the integrated traffic pattern run by a common three-dimensional network under routing algorithms XYZ and OE_Z; xyz_abdtr represents the delay result of the integrated traffic pattern run under the routing algorithm XYZ by the conventional approximate communication architecture suitable for the fixed routing algorithm, and xyz_dcbw and oe_z_dcbw represent the delay result of the integrated traffic pattern run under the routing algorithms XYZ and oe_z by the invention. It is not difficult to see that the invention can effectively improve the condition of saturation of injection rate when network communication approaches congestion, and can maintain good control effect under a dynamic routing algorithm.
FIG. 6 shows the power consumption results of integrated flow simulation, in which XYZ_original and OE_Z_original represent the power consumption results of a common three-dimensional network running integrated flow mode under routing algorithms XYZ and OE_Z; xyz_abdtr represents the power consumption result of the integrated traffic pattern run under the routing algorithm XYZ by the conventional approximate communication architecture suitable for the fixed routing algorithm, and xyz_dcbw and oe_z_dcbw represent the power consumption result of the integrated traffic pattern run under the routing algorithms XYZ and oe_z by the invention. The result shown in fig. 6 shows that the invention can effectively reduce the communication power consumption cost of the network on chip and maintain good control effect under the dynamic routing algorithm.
The foregoing has been described schematically the invention and embodiments thereof, which are not limiting, but are capable of other specific forms of implementing the invention without departing from its spirit or essential characteristics. The drawings are also intended to depict only one embodiment of the invention, and therefore the actual construction is not intended to limit the claims, any reference number in the claims not being intended to limit the claims. Therefore, if one of ordinary skill in the art is informed by this disclosure, a structural manner and an embodiment similar to the technical scheme are not creatively designed without departing from the gist of the present invention, and all the structural manners and the embodiment are considered to be within the protection scope of the present patent. In addition, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" preceding an element does not exclude the inclusion of a plurality of such elements. The various elements recited in the product claims may also be embodied in software or hardware. The terms first, second, etc. are used to denote a name, but not any particular order.
Claims (7)
1. A network-on-chip approximate communication system for data-intensive applications, wherein the system network comprises a plurality of approximate communication architectures, each approximate communication architecture comprises a processing core, a router and a network interface, and the network interface is respectively connected with the processing core and the router;
a main control node is arranged in the processing core, and a global controller is arranged in the main control node; the router is provided with a network congestion condition supervision unit which is used for transmitting congestion information of the network to the global controller in real time; the network interface is provided with a data screening unit and a data compression and decompression unit, and system data is approximately processed by the data screening unit and the data compression and decompression unit in the network port and then transmitted to the network through the router;
the global controller judges the congestion status of network communication through the accumulated waiting time of each router in the regional network, wherein the regional network comprises all router paths possibly passed through by data in a packet source node and a destination node of the network on chip;
the network congestion condition monitoring unit converts the communication congestion condition of the network into router waiting time, judges the network congestion condition according to the waiting time, sends congestion node information to the global controller, records the phenomenon that a plurality of ports in the network topology compete for the same output port, and counts the times that the input port does not acquire transmission priority into the waiting time; recording the number of times that the input port does not acquire the transmission right in the waiting time when the data back pressure signal is zero, namely when the output port matched with the current input port has no idle virtual channel; and defining the router node whose waiting time exceeds the set threshold as a congestion node.
2. The network-on-chip proximity communication system for data-intensive applications of claim 1 wherein the global controller counts the number of congestion of router nodes in the area and sends different data proximity information to the source node that sent the packet based on the number of congestion nodes.
3. The network-on-chip proximity communication system for data-intensive applications of claim 1, wherein congestion node information transmitted by the network congestion status supervision unit is transmitted to the global controller to match corresponding data compression rates, while clearing an original latency of the network congestion status supervision unit.
4. A network-on-chip proximity communication system for data-intensive applications as claimed in claim 1, wherein the data screening unit steplessly adjusts the data compression rate transmitted to the network within an allowed proximity rate threshold.
5. The network-on-chip proximity communication system of claim 4 wherein the data filtering unit generates pseudo-random numbers through a linear feedback shift register to filter data to be compressed, and data packets not having acquired transmission rights are compressed.
6. The network-on-chip proximity communication system for data-intensive applications of claim 1, wherein the communication architecture employs a data transmission mode of proximity communication.
7. A network-on-chip proximity communication system for data-intensive applications as claimed in claim 1, wherein said network is a two-dimensional or three-dimensional network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111126409.4A CN113839878B (en) | 2021-09-26 | 2021-09-26 | Network-on-chip approximate communication system for data intensive application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111126409.4A CN113839878B (en) | 2021-09-26 | 2021-09-26 | Network-on-chip approximate communication system for data intensive application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113839878A CN113839878A (en) | 2021-12-24 |
| CN113839878B true CN113839878B (en) | 2023-05-23 |
Family
ID=78970136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111126409.4A Active CN113839878B (en) | 2021-09-26 | 2021-09-26 | Network-on-chip approximate communication system for data intensive application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113839878B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114615208B (en) * | 2022-03-09 | 2024-02-23 | 新华三半导体技术有限公司 | Back pressure information transmission and request sending method and device and network chip |
| CN115277563B (en) * | 2022-06-07 | 2024-03-19 | 南京大学 | Network-on-chip approximation control system based on offline reinforcement learning |
| CN117411561A (en) * | 2023-12-05 | 2024-01-16 | 西安电子科技大学 | Optical network-on-chip communication method based on approximate calculation and optical network-on-chip |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010137572A1 (en) * | 2009-05-25 | 2010-12-02 | 日本電気株式会社 | Network-on-chip, network routing method, and system |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8020163B2 (en) * | 2003-06-02 | 2011-09-13 | Interuniversitair Microelektronica Centrum (Imec) | Heterogeneous multiprocessor network on chip devices, methods and operating systems for control thereof |
| KR100785472B1 (en) * | 2006-09-19 | 2007-12-13 | 삼성전자주식회사 | Apparatus for managing urgent packet latency of network-on-chip and method using the same |
| CN103401858B (en) * | 2013-07-29 | 2016-01-13 | 北京交通大学 | The expansion method of sheet upstream routers congestion aware scope |
| TWI531187B (en) * | 2014-11-24 | 2016-04-21 | 財團法人工業技術研究院 | Noc timing power estimating device and method thereof |
| CN104901899B (en) * | 2015-06-18 | 2018-03-06 | 西安电子科技大学 | The adaptive routing method of two-dimentional Survey on network-on-chip topology |
| CN109547263B (en) * | 2018-12-15 | 2021-08-20 | 华南理工大学 | Network-on-chip optimization method based on approximate calculation |
| CN112099614B (en) * | 2020-08-03 | 2022-04-22 | 华南理工大学 | Many-core system energy consumption and performance optimization method based on collaborative approximate calculation |
-
2021
- 2021-09-26 CN CN202111126409.4A patent/CN113839878B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010137572A1 (en) * | 2009-05-25 | 2010-12-02 | 日本電気株式会社 | Network-on-chip, network routing method, and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113839878A (en) | 2021-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113839878B (en) | Network-on-chip approximate communication system for data intensive application | |
| Duato et al. | Performance evaluation of adaptive routing algorithms for k-ary n-cubes | |
| CN111104775B (en) | Network-on-chip topological structure and implementation method thereof | |
| CN102823214B (en) | Know the heterogeneity internet of performance and traffic | |
| CN109873771B (en) | On-chip network system and communication method thereof | |
| CN106411770A (en) | Data center network energy-saving routing algorithm based on software defined network (SDN) architecture | |
| CN107920025B (en) | Dynamic routing method for CPU-GPU heterogeneous network on chip | |
| CN104022950B (en) | It is a kind of to share the router topology cached with self-configuring | |
| CN115277563B (en) | Network-on-chip approximation control system based on offline reinforcement learning | |
| CN113489617A (en) | Minimum network energy consumption optimization method and system based on traffic grooming | |
| Liu et al. | An encapsulated packet-selection routing for network on chip | |
| CN111245730B (en) | Routing system and communication method of network on chip | |
| Escamilla et al. | ICARO: Congestion isolation in networks-on-chip | |
| Dana et al. | Congestion aware routing algorithm for mesh network-on-chip platform | |
| Alonso et al. | Power saving in regular interconnection networks built with high-degree switches | |
| CN110784406B (en) | Dynamic self-adaptive on-chip network threshold routing method based on power perception | |
| Ueno et al. | VCSN: Virtual circuit-switching network for flexible and simple-to-operate communication in HPC FPGA cluster | |
| Whelihan et al. | Memory optimization in single chip network switch fabrics | |
| Fang et al. | Exploration on routing configuration of HNoC with intelligent on-chip resource management | |
| Zhang et al. | A cellular NoC architecture based on butterfly network coding (CBNoC) | |
| Salehi et al. | Minimal fully adaptive fuzzy-based routing algorithm for Networks-on-Chip | |
| Li et al. | Modeling and Simulation of Network-on-Chip Routing Algorithm Based on OPNET | |
| Conner et al. | Link shutdown opportunities during collective communications in 3-D torus nets | |
| Garcia et al. | Recn-dd: A memory-efficient congestion management technique for advanced switching | |
| CN117221212B (en) | Optical network on chip low congestion routing method and related equipment |
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 |