CN101626525A - Chip optical interconnect switching unit design based on micro-resonator - Google Patents
Chip optical interconnect switching unit design based on micro-resonator Download PDFInfo
- Publication number
- CN101626525A CN101626525A CN200910087680A CN200910087680A CN101626525A CN 101626525 A CN101626525 A CN 101626525A CN 200910087680 A CN200910087680 A CN 200910087680A CN 200910087680 A CN200910087680 A CN 200910087680A CN 101626525 A CN101626525 A CN 101626525A
- Authority
- CN
- China
- Prior art keywords
- crosspoint
- micro
- resonator
- omnirange
- switching unit
- 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
Landscapes
- Use Of Switch Circuits For Exchanges And Methods Of Control Of Multiplex Exchanges (AREA)
- Optical Communication System (AREA)
Abstract
The invention relates to a switching unit based on Micro-resonators and used for an optical interconnect and optical switching network among chips. The switching unit can be applied to a rectangle MESH optical switching network to complete a switching function without blockage under the condition of using fewer Micro-resonators, and besides, the switching unit can be divided into an omni-directional center node switching unit, an omni-directional edge node switching unit and a direction limitation center node switching unit for meeting different demands according to redundant backup capability and different network positions of the switching unit in a switching array. The dissymmetrical chip optical interconnect established on the basis of the switching unit has few using resources, low overall power dissipation, high bandwidth density, and the like.
Description
Technical field
The chip optical interconnect network exchange unit that the present invention relates to belongs to the chip optical interconnect field, is applied to building of chip chamber optical switching network.
Background technology
Current, the high-performance micro processor has obtained application more and more widely in every field, is the focus of industry at the research of high-performance processor always.Handle single core processor that frequency is a characteristic because its high energy consumption, expensively run into the development bottleneck that is difficult to overcome with height, and the development of multiple kernel construction processor chip by in recent years, become the main flow of high-performance micro processor gradually.Along with the polycaryon processor continuous advancement in technology, the number of handling nuclear obtains surprising growth, nuclear becomes one of the most key factor of polycaryon processor chip design gradually with the performance quality of data interaction network between the nuclear, and this means will the more and more effect of key of performance when the design of polycaryon processor at the research of chip chamber interference networks communication structure.
In the process that chip chamber interference networks communication structure develops rapidly, the miniaturization of chip and the rapid growth of bandwidth demand all propose harsh more requirement to the chip chamber interference networks.Traditional electrical interconnection network configuration be difficult to have solved the limitation of himself performance: low bandwidth, high latency, high energy consumption and crosstalking.By contrast, light interconnection is then because its almost unlimited bandwidth, the low delay, and low-power consumption, low crosstalking begins to be introduced into and is applied among the high performance chip chamber interference networks.In addition, the advantage that light interconnection also has an electrical interconnection to be equal to is: in case after the fiber waveguide link establishment because the low-loss of waveguide link, end to end data do not need as electrical interconnection regenerate, amplification and buffer memory.More than these all allow optical interconnection network be full of bright application prospect in the chip interconnect field.
But, still there is certain technical limitations in the development of light interconnection technique, and that main problem is to lack is low-cost, high efficiency optical storage and optical processor, so be difficult to realize the control to the chip chamber optical switching network.The research of relevant optical processor and design all are a focus not only in the light field of interconnects in the overall optical communications field always.And the research of expert's effort at home and abroad down, and gratifying achievement also constantly appears in optical processor in development.The development of nanocomposite optical element in recent years, make the optical microcavity that adopts usually in the photonic integrated circuits, especially micro-resonator can be used on traditional silicon and silicon-on-insulator (SOI) substrate at an easy rate, and this is just needed the chip chamber optical interconnection network of high density of integration to create a new developing direction.Because this class micro-resonator (diameter only is about 10 μ m) has low-power consumption and low insertion loss, can well be applied to optical filter, switch, optical add/drop multiplexer, and then be applied among the design of chip chamber optical interconnection network crosspoint.At present, a new direction has been opened up in the design of chip chamber optical interconnection network: on the basis of micro-resonator, the design crosspoint makes up the overall optical interference networks by introducing exchange algorithm again.The present invention's design just is being based on this thinking, designs the crosspoint and the high performance chip chamber optical interconnection network that have the top standard on this branch direction.
Summary of the invention
The present invention's design is based on micro-resonator, is applied to the brand-new network exchange unit of a class of chip chamber optical interconnection network.The design similar with other compared, this crosspoint passes through structure optimization design, can when satisfying basic clog-free function of exchange, reduce Primary Component---the use number of micro-resonator, thereby reduce the power consumption of crosspoint and shared area.The crosspoint of the present invention's design is applied to chip chamber optical interconnection network structure, can further realizes a kind of asymmetric core chip interconnect network configuration of novelty, reach the effect that reduces whole network power consumption, increases network bandwidth density.
The present invention is based on that micro-resonator designs, and at first does simple an introduction with regard to the function and the principle of micro-resonator:
Shown in the figure one is by fiber waveguide and the routing unit that micro-resonator constitutes of two intersections, and between micro-resonator and two fiber waveguides predetermined distance is arranged, so that the coupling of light signal.Can control the frequency of micro-resonator by the signal of telecommunication, realize the open and close state.When the operating frequency that is in out state from signal frequency in the horizon light waveguide and resonator is identical, light signal will be coupled on the resonator, realize being coupled to again in the vertical light waveguide after the turning to of 90 degree, realize the function of horizon light signal routing to the vertical light waveguide.And if resonator is when being in off status, the signal on the fiber waveguide will move ahead along former direction, can not be coupled on the optical resonator.
The present invention utilizes this routing function of micro-resonator, based on the model of rectangle MESH topology, and with reference to three kinds of nodes in this kind rectangular net topology: mid-side node, Centroid, corner node design the node switching unit.Because corner node does not need to carry out swap operation, so only need node switching unit, design centre, mid-side node crosspoint.Try one's best few micro-resonator and few crossing waveguide of trying one's best of employing is the basic principle that the present invention designs.Because the power consumption of whole crosspoint and area occupied depend primarily on the number of micro-resonator, to compare, power loss on the fiber waveguide and area take substantially and can ignore.And for the optical interconnection network of chip chamber, low-power consumption and small size mean more massive integrated and higher transmission bandwidth density, and these all are the key indexs of estimating the planned network quality, and few intersection of trying one's best is in order to optimize network configuration, to reduce the insertion loss that intersection is brought.
By design verification repeatedly, and relevant emulation testing, the structure that the present invention at first designs is called the omnirange crosspoint.Shown in Figure 2 is omnidirectional center/mid-side node crosspoint, and they have used 8/4 micro-resonator respectively, and the area that takies approximately is 16d
2(wherein d is the diameter of micro-resonator).This in addition structure has only been used two horizontal waveguide and two vertical waveguide, has used 4 crosspoints.
Function with regard to omnirange center crosspoint describes below:
1, when light signal by waveguide when left road enters crosspoint, if with No. 1 micro-resonator state of being changed to out, signal will be coupled on No. 1 micro-resonator, along setting out on a journey output; If No. 1 micro-resonator is changed to off status, and with No. 4 micro-resonator states of being changed to out, signal will be coupled to micro-resonator No. 4, along road output down; 1, No. 4 micro-resonators are off status, and signal is then exported along right wing;
2, when light signal by waveguide when right wing enters crosspoint, if with No. 8 micro-resonator states of being changed to out, signal will be coupled on No. 8 micro-resonators, along road output down; If No. 8 micro-resonators are changed to off status, and with No. 5 micro-resonator states of being changed to out, signal will be coupled to micro-resonator No. 5, along setting out on a journey output; 5, No. 8 micro-resonators are off status, and signal is then along the output of left road;
3, when light signal enters crosspoint by waveguide from setting out on a journey, if with No. 2 micro-resonator states of being changed to out, signal will be coupled on No. 2 micro-resonators, export along right wing; If No. 2 micro-resonators are changed to off status, and with No. 6 micro-resonator states of being changed to out, signal will be coupled to micro-resonator No. 6, along the output of left road; 2, No. 6 micro-resonators are off status, and signal is then along road output down;
4, when light signal by waveguide when down the road enters crosspoint, if with No. 7 micro-resonator states of being changed to out, signal will be coupled on No. 7 micro-resonators, along the output of left road; If No. 7 micro-resonators are changed to off status, and with No. 3 micro-resonator states of being changed to out, signal will be coupled to micro-resonator No. 3, export along right wing; 3, No. 7 micro-resonators are off status, and signal is then along setting out on a journey output.
Then, hereinafter again the function of omnirange limit crosspoint is described:
1, when light signal by waveguide when left road enters crosspoint, if with No. 2 micro-resonator states of being changed to out, signal will be coupled on No. 2 micro-resonators, along road output down, otherwise will export along right wing;
2, when light signal by waveguide when right wing enters crosspoint, if with No. 4 micro-resonator states of being changed to out, signal will be coupled on No. 4 micro-resonators, along road output down, otherwise will be along the output of left road;
3, when light signal by waveguide when down the road enters crosspoint, if with No. 3 micro-resonator states of being changed to out, signal will be coupled on No. 3 micro-resonators, along the output of left road; No. 3 micro-resonators are changed to off status, No. 1 micro-resonator is placed out state, signal will be exported along right wing.
Need should be mentioned that the difference of arranging according to waveguide input and output direction, structure shown in Figure 2 will change to some extent, because its principle also is based on design of the present invention, so this variation also should be within protection scope of the present invention.
According to the MESH network theory, in rectangle MESH structure, when a node will carry out exchanges data with another node, have eight kinds of directions and 12 kinds of link structures (as Fig. 3) between them.But the dotted portion among the figure is link not necessarily, the connection that it is realized can be replaced by other mode, possess the function of setting up the kind link of solid line part among the figure as long as that is to say each crosspoint, any 2 of MESH just can set up corresponding link.Then, the MESH network is controlled, can be realized the fairly simple optical switching network of chip chamber efficiently by routing algorithm.Therefore, the present invention invents again according to this principle and has designed the limited center crosspoint (as Fig. 4) of a kind of direction, by four redundant links of cutting, make performance save the micro-resonator of half under the situation of variation slightly, thereby reduce the power consumption of network and area occupied to satisfy the demand of different chip chamber optical interconnection networks.Redundant deleting of link can be selected, and can produce the structure that is different from Fig. 4 according to the difference of deleting link, so also should be within protection range of the present invention to the variation of the structure work of Fig. 4 according to this principle.
To describe with regard to the function of the limited center of direction crosspoint below:
1, when light signal by waveguide when left road enters crosspoint, if with No. 1 micro-resonator state of being changed to out, signal will be coupled on No. 1 micro-resonator, along setting out on a journey output; If No. 1 micro-resonator is changed to off status, and with No. 2 micro-resonator states of being changed to out, signal will be coupled to micro-resonator No. 2, along road output down; 1, No. 2 micro-resonators are off status, and signal is then exported along right wing;
2, when light signal by waveguide when right wing enters crosspoint, signal can only be from the output of left road;
3, when light signal enters crosspoint by waveguide from setting out on a journey, if with No. 3 micro-resonator states of being changed to out, signal will be coupled on No. 3 micro-resonators, along the output of left road; If No. 3 micro-resonators are changed to off status, signal will be along road output down;
4, when light signal by waveguide when down the road enters crosspoint, if with No. 4 micro-resonator states of being changed to out, signal will be coupled on No. 4 micro-resonators, along the output of left road; If No. 4 micro-resonators are changed to off status, signal will be along the output of setting out on a journey.
By using the node switching unit of above design, the present invention has designed a kind of light interconnection switch network architecture, and this structure is made up of two kinds of crosspoints, and a kind of is the node switching unit that is positioned at the MESH network node; Another kind is to be used for examining the transmitter/receiver unit that the I/O mouth links to each other with CPU.Wherein the node switching unit must possess the function of carrying out data interaction with other arbitrary nodes of MESH network, and it also will be connected with transmitter/receiver unit, make data on the network to handle nuclear by corresponding transmitter/receiver unit and corresponding C PU and carry out data interaction.And the effect of transmitter/receiver unit is exactly that the data that CPU nuclear will send or receive are exchanged on the relevant node switching unit by routing algorithm obviously.In the interference networks of the present invention's design, can realize transmitter/receiver unit with the mid-side node crosspoint.
Description of drawings
Fig. 1: micro-resonator is realized the schematic diagram of routing function in optical transmission;
Fig. 2: the structural representation of omnirange Centroid crosspoint and omnirange mid-side node crosspoint in the present invention's design, wherein annulus is represented micro-resonator, the tubular wire of intersection is represented fiber waveguide;
Fig. 3: node switching link direction schematic diagram;
Fig. 4: through deleting the structural representation of the limited Centroid crosspoint of the direction that designs behind the redundancy backup link;
Fig. 5: the Combination chip optical interconnect schematic network structure that automatically controlled preparative layer and optical transport layer make up;
Fig. 6: the asymmetric rectangle MESH chip optical interconnect structural representation that uses crosspoint to build;
Fig. 7: micro-resonator number comparison diagram in 9 * 9 networks of different research and design;
Fig. 8: OPNET performance simulation figure;
Embodiment
To the technical scheme of the embodiment of the present invention design be clearly and completely described below.Obviously, embodiment as described herein only is part embodiment of the present invention, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that is obtained under the creative work prerequisite.
In order to show the superiority of the crosspoint that the present invention designs, the present invention analyzes the superior function of the crosspoint of the present invention's design by a chip interconnect network struction embodiment.
As everyone knows, though optical interconnection network has high bandwidth, low delay, low-power consumption, low advantage such as crosstalk, but because the design of light buffer memory and optical processing unit is very complicated, simple optical waveguide interconnection network is difficult to realize complicated routing and switching function, is difficult to satisfy the demand that present polycaryon processor number increases.In contrast be, the electrical interconnection of chip chamber can be stored the signal of telecommunication and relevant processing very easily, can realize complicated network function, but but because limited bandwidth, and transmission performance not good to be difficult to carry out large-capacity data mutual.Therefore,, combine a kind of organic mixed structure of formation with optical transport layer,, can utilize the advantage of light interconnection transmission again, realize high performance chip chamber optical interconnection network with regard to the controlled function that promptly can utilize electrical interconnection to be easy to realize with simple automatically controlled preparative layer.Based on the blended network of this photoelectricity, present embodiment has designed a kind of two-layer rectangle MESH network (as Fig. 5), and adopts the crosspoint in the present invention's design to finish this Network Design.
Among this Combination network configuration, when a CPU processing nuclear need carry out data interaction with other nuclear, this nuclear will be to order of automatically controlled preparative layer, next automatically controlled preparative layer will be adjusted the on off state of micro-resonator in each crosspoint, set up an effective link by optical transport layer, realize the exchange of data.The design of automatically controlled preparative layer and design of the present invention invention relation is little, is not discussed here, next emphasis is applied to optical transport layer to the unit of the present invention's design and makes a detailed explanation.
The optical transport layer that present embodiment makes up is a kind of brand-new asymmetric rectangle MESH chip optical interconnect structure, here using 4 * 4 network configuration is example, as shown in Figure 6, this network configuration has been used 8 mid-side node crosspoints and 4 Centroid crosspoints, by at least 20 emissions/reception crosspoint, this network can provide 20 mouths as the CPU access port at least, and the number of CPU nuclear can also increase as required.The symmetrical rectangle MESH network of common design kind 4 * 4 then generally can only corresponding 4 * 4 CPU multinuclear array.Why the present invention can adopt this brand-new asymmetric rectangle MESH structure just because of can have benefited from the crosspoint area and significantly descend, and the reduction of whole network energy consumption.If this crosspoint can access further popularization, this novel asymmetric design network can lead the development in future direction.
Ensuing work will be carried out performance evaluation according to this asymmetric rectangle MESH structure:
Since in the crosspoint number of micro-resonator and structural design determined directly that the area of whole unit takies, power loss and transmission performance, so the design of crosspoint is most important factor for overall optical chip interconnect network.In the asymmetric rectangle MESH of the n * n network configuration based on the embodiments of the invention design, the number that micro-resonator uses is as follows:
N＝(n-2)
2×r
center+(n-2)×4×r
side+k
cores×r
inject/eject
R wherein
CenterRefer to the number of the micro-resonator that in Centroid, uses; r
SideRefer to the number of the micro-resonator that uses in the fringe node; k
CoresRefer to that then CPU handles the number of nuclear, comparing value for convenience is n * n.
In order further to be analyzed with other networks, in Fig. 7, with 9 * 9 rectangle MESH network configurations is that example is analyzed, wherein AD refers to the network that uses the omnirange crosspoint, LD refers to use the network of the limited crosspoint of direction, as can be seen from the figure, the designed network of this embodiment occupies suitable advantage in the number that micro-resonator uses.
Next will be to carrying out the calculating of power consumption based on the designed network of the present invention, owing to micro-resonator on crosspoint, the power consumption of other device and material is much smaller, and ignores their influence here for the time being, so total power consumption:
P wherein
OnThe power that refers to when micro-resonator is in out state to be consumed is approximately 0.5mW, and when off status, does not almost have consumed power, therefore also ignores here.I, j refer to center and limit crosspoint, m respectively
i, m
jThen be meant the number that is in out the micro-resonator of state in center and the limit crosspoint respectively.Here be example with 9 * 9 chip optical interconnect MESH network, through calculating, when using the omnirange crosspoint, total power consumption approximately is 126.15mW, and during the limited crosspoint of service orientation, total power consumption is 98.78mW.This result shows that this structure makes that power consumption has obtained reducing greatly.
Meanwhile, present embodiment uses OPNET to do simple emulation testing to analyze its performance with above-mentioned 9 * 9MESH network.In emulation, network is set up m bar link at random, and the packet that produces 1KB simultaneously transmits between each link, then in the whole transmission course of emulation, and when getting different m values, the average normalized data throughout that the average delay of data and CPU examine.Can find that from the result of Fig. 8 along with the increase of the number of links of setting up, throughput also increases in proportion; Increasing slowly also appears in average delay.But when link is increased to some (the limited crosspoint of direction is 10 links, and the omnirange crosspoint is 13 links), performance degradation is rapid.This explanation is in the crosspoint of the present invention's design, and the performance of omnirange crosspoint is better than the limited crosspoint of direction.Larger when network when data interaction is more frequent, should adopt the omnirange crosspoint to build network; And it is slightly little to work as network size, when data interaction is less demanding, adopts the limited crosspoint of direction better.
The explanation of above embodiment just is used to help to understand the method and the advantage place thereof of the present invention's design; for one of ordinary skill in the art; according to thought of the present invention; part in specific embodiments and applications all can change; in sum; this description should not be construed as limitation of the present invention, and the various conspicuous change of under the situation of spirit that does not deviate from the method for the invention and claim scope it being carried out is all within protection scope of the present invention.
Claims (10)
1. a class is applied to the crosspoint that the chip chamber optical interconnection network is built, and adopting micro-resonator and fiber waveguide is basic device, through the Combinatorial Optimization design, has clog-free function of exchange.According to the difference of function of exchange, crosspoint is divided into three kinds: omnirange Centroid crosspoint, and omnirange mid-side node crosspoint, the limited Centroid crosspoint of direction, its structure mainly comprises:
Each two parallel optical waveguide of horizontal direction and vertical direction;
Four or eight micro-resonators.
2. according to the requirement of right 1, omnirange Centroid crosspoint, in the limited Centroid crosspoint of direction, two fiber waveguides of horizontal direction and two fiber waveguides of vertical direction are carried out square crossing, have four crosspoints, total four direction up and down, each direction respectively has the passage of a pair of input and output.
3. according to the requirement of right 1, in the omnirange mid-side node crosspoint, after two fiber waveguides of horizontal direction and two fiber waveguides of vertical direction are carried out square crossing, have only three directions, the passage of respectively corresponding three pairs of input and output.
4. requiring micro-resonator according to right 1 is a kind of elementary cell with light routing function.When it was in opening state, micro-resonator can be finished with frequency light signal 90 degree and turn to function; When it was in closed condition, micro-resonator was not handled light signal.
5. according to the requirement of right 1 and right 4, omnirange Centroid crosspoint is controlled as routing by eight micro-resonators, makes that crosspoint can be to the exchange on the light signal travel direction of different directions input and output.
6. according to right 1 and right 4 requirements, limited Centroid crosspoint of direction and omnirange mid-side node crosspoint are controlled as routing by four micro-resonators, can be to the exchange on the light signal travel direction of different directions input and output.
7. according to the requirement of right 4, in order to finish routing function, micro-resonator is distributed in the certain location on every side of crossing waveguide, and is all equal apart from the position in each crosspoint.
8. according to the requirement of right 7, realize different particular exchange forms according to three kinds of different crosspoint needs, and the difference of fiber waveguide input and output direction, different distributed architectures had.
9. according to the requirement of right 8, the micro-resonator distributed architecture of omnirange Centroid crosspoint can make each road input all can exchange to the function of other three tunnel realizations outputs, and the limited center of direction crosspoint has been deleted several roads function of exchange on this basis and used four micro-resonators less.
10. according to the requirement of right 8, the micro-resonator distributed architecture of omnirange mid-side node crosspoint can make each road input all exchange to the function of other two tunnel realizations outputs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910087680A CN101626525A (en) | 2009-07-02 | 2009-07-02 | Chip optical interconnect switching unit design based on micro-resonator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910087680A CN101626525A (en) | 2009-07-02 | 2009-07-02 | Chip optical interconnect switching unit design based on micro-resonator |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101626525A true CN101626525A (en) | 2010-01-13 |
Family
ID=41522160
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910087680A Pending CN101626525A (en) | 2009-07-02 | 2009-07-02 | Chip optical interconnect switching unit design based on micro-resonator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101626525A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917333A (en) * | 2010-07-06 | 2010-12-15 | 西安电子科技大学 | Region-based photoelectric double-layer network-on-a-chip and routing method |
| CN103490820A (en) * | 2013-09-13 | 2014-01-01 | 中国人民解放军国防科学技术大学 | Network interconnecting device for cache-free optical interconnection networks |
| CN105814483A (en) * | 2013-12-13 | 2016-07-27 | 瑞典爱立信有限公司 | Parallel and WDM silicon photonics integration in information and communications technology systems |
-
2009
- 2009-07-02 CN CN200910087680A patent/CN101626525A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917333A (en) * | 2010-07-06 | 2010-12-15 | 西安电子科技大学 | Region-based photoelectric double-layer network-on-a-chip and routing method |
| CN101917333B (en) * | 2010-07-06 | 2012-10-24 | 西安电子科技大学 | Region-based photoelectric double-layer network-on-a-chip and routing method |
| CN103490820A (en) * | 2013-09-13 | 2014-01-01 | 中国人民解放军国防科学技术大学 | Network interconnecting device for cache-free optical interconnection networks |
| CN105814483A (en) * | 2013-12-13 | 2016-07-27 | 瑞典爱立信有限公司 | Parallel and WDM silicon photonics integration in information and communications technology systems |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101917333B (en) | Region-based photoelectric double-layer network-on-a-chip and routing method | |
| CN100495383C (en) | Three-dimensional multiprocessor system chip | |
| CN102027399B (en) | Optoelectronic switches using on-chip optical waveguides | |
| KR101956855B1 (en) | Fabric interconnection for memory banks based on network-on-chip methodology | |
| Ye et al. | A torus-based hierarchical optical-electronic network-on-chip for multiprocessor system-on-chip | |
| Shacham et al. | Photonic NoC for DMA communications in chip multiprocessors | |
| CN103091784B (en) | Low-loss four-port non-blocking optics router based on micro-ring resonator | |
| Guo et al. | Low insertion loss and non-blocking microring-based optical router for 3D optical network-on-chip | |
| CN102638311A (en) | Optical network-on-chip system based on wavelength allocation and communication method of system | |
| CN106416110B (en) | Expansible silicon photon count statistic framework for optical-fiber network | |
| CN102333250B (en) | Low-congestion communication method and router for realizing shared path transmission of optical network on chip | |
| CN105451103A (en) | Wavelength-allocation-based three-dimensional optical on-chip network router communication system and method | |
| CN102540345B (en) | Low-loss, low-crosstalk, four-port and non-blocking optical router based on micro-ring resonators | |
| CN102361467B (en) | Network structure on non-blocking optical section and communication method thereof | |
| CN102780936B (en) | Optical on-chip network system of non-blocking communication and communication method thereof | |
| CN110568560B (en) | Method for realizing large-scale optical switch topology array chip based on Benes structure and having balanced loss | |
| CN103353632B (en) | Optical switching unit based on micro-ring resonator | |
| CN101626525A (en) | Chip optical interconnect switching unit design based on micro-resonator | |
| Dang et al. | Mode-division-multiplexed photonic router for high performance network-on-chip | |
| CN105530206A (en) | Torus network based dual-access structures and working mode thereof | |
| CN110568559A (en) | Chip architecture based on large-scale optical switch topology array | |
| CN107220209B (en) | Three-dimensional optical network-on-chip architecture based on faults, communication method and optical router | |
| CN103888361A (en) | Optical router orienting fat-tree type topology structure | |
| CN110568552B (en) | Large-scale array crossed waveguide recombination and separation structure and design method thereof | |
| Seyednezhad et al. | Routing design in optical networks-on-chip based on gray code for optical loss reduction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20100113 |