CN109710314B - A method of based on graph structure distributed parallel mode construction figure - Google Patents
A method of based on graph structure distributed parallel mode construction figure Download PDFInfo
- Publication number
- CN109710314B CN109710314B CN201811568076.9A CN201811568076A CN109710314B CN 109710314 B CN109710314 B CN 109710314B CN 201811568076 A CN201811568076 A CN 201811568076A CN 109710314 B CN109710314 B CN 109710314B
- Authority
- CN
- China
- Prior art keywords
- composition
- node
- forming
- obtains
- parts
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of methods based on graph structure distributed parallel mode construction figure, it is related to the folding field of graph structure technical field in data structure, distributed parallel computing field and figure, the existing graph structure serially assembled is solved, the duration of calculating can increase with the increase of comparison number；The memory headroom for saving list can increase with the increase of node, and the node upper limit of assembling figure will be limited when the memory needed for memory node list reaches the calculator memory upper limit.The present invention includes step are as follows: the node of assembling figure needed for preparing；By node cutting or/and the cutting of composition figure, realize that parallel computation forms final figure.The present invention is used for the figure of distributed parallel mode construction graph structure.
Description
Technical field
A method of based on graph structure distributed parallel mode construction figure, being used for distributed parallel mode construction graph structure
Figure, be related to the folding field of graph structure technical field in data structure, distributed parallel computing field and figure.
Background technique
The generally serial assembling of existing figure (i.e. based on the figure of graph structure) assembling process, assembling substantially process are as follows:
1. pulling all nodes to be assembled, and node is retained in a manner of list；
2. node and other all nodes are done and matched two-by-two, if can be associated in since first node in list
It is then recorded as figure together；
3. graph structure is completed after node all in list is compared association all in accordance with the mode of step 2.
The building process of figure will be with increasing for number of nodes and compare and judge that number increases in the above scheme, comparison
Judge that the relationship of number and number of nodes is as follows:
Compare number=number of nodes * (number of nodes -1)；
The shortcomings that being calculated using the above method is as follows:
1. determine comparison number is the quantity of node, the duration of calculating can increase with the increase of comparison number；I.e.
The calculating that overlong time may cause first time is not completed, it is necessary to start secondary calculating.Such as: practical application structure
The time for building figure is usually once a day, it is possible to can exceed that 24 the time required to the quite huge calculating of the node of required structure figures
Hour, therefore first day figure calculates the figure for just needing to calculate second day not yet, and figure is caused to be unable to complete building forever.
2. the memory headroom for needing to save list then can due to needing in the form of a list to save node and in case using
Increase with the increase of node, assembling figure will be limited when the memory needed for memory node list reaches the calculator memory upper limit
The node upper limit.Such as: if a node saves the data volume for needing 1KB, if there is 100,000,000 node amount, required storage is empty
Between be only that node data inherently needs about 95GB, single machine has the server of 95GB actually rare in existing IT resource, if
Physical memory space needed for executing program in JVM then is about that 240G (calculates intermediate result to be also required to storage therefore also need
More memory spaces), in system architecture for single machine memory it is excessively high, be it is unreasonable, patrolled from the configuration of computer hardware
For volume, monomer resource is excessive to will cause the problem of high-speed equipment floods low-speed device, and IT resource is caused to run uncoordinated ask
Topic, is briefly exactly compatibility issue.
Summary of the invention
Aiming at the problem that the studies above, the purpose of the present invention is to provide one kind to be based on graph structure distributed parallel mode structure
The method for building figure solves the existing graph structure serially assembled, and the duration of calculating can increase with the increase of comparison number, may
It will cause the problem of figure is unable to complete building；The memory headroom for saving list can increase with the increase of node, work as memory node
Memory needed for list will limit the node upper limit of assembling figure when reaching the calculator memory upper limit, cause memory headroom insufficient,
There are compatibility issues etc. for configuration.
In order to achieve the above object, the present invention adopts the following technical scheme:
A method of based on graph structure distributed parallel mode construction figure, which is characterized in that following steps:
Step 1. prepares the node of required assembling figure；
The node that step 2. obtains step 1 is divided into more parts, it is parallel node in every part is carried out two-by-two between comparison
Match, obtain at least one composition figure or/and does not form the node of figure, the composition figure number judged, the number of nodes for not forming figure
Or calculate whether wheel number reaches given threshold value, if not up to, going to step 3, if reaching, going to step 5；
If step 3. not up to threshold value, and if step 2 obtain multiple composition figures, multiple composition figures that step 2 is obtained point
At more parts, and it is about to compare matching between the figure in every part carries out two-by-two, or/and
If step 2 obtains multiple nodes for not forming figure, and is about to multiple nodes for not forming figure that step 2 obtains and is divided into
More parts, parallel comparing between the node progress two-by-two in every part matches, and obtains at least one fusion composition figure or/and composition figure
Or/and the node of figure is not formed；
If step 4, multiple composition figures obtain at least one fusion composition figure or/and composition after being divided into more parts of parallel computations
Figure, multiple nodes for not forming figure obtain at least one composition figure after being divided into more parts of parallel computations or/and do not form the node of figure,
The former fusion composition figure or/and composition figure are compared with the composition figure of the latter then and matched, at least one fusion group is obtained
At figure or/and composition figure or/and the node for not forming figure；Judge obtained composition figure number, the number of nodes for not forming figure or calculating
Whether wheel number reaches given threshold value, if not up to, repeating step 3- step 4 if reaching and going to step 5；
Step 5, the node for obtaining composition figure or/and not forming figure, the composition figure in this step be fusion composition figure or/and
Otherwise composition figure obtains final composition figure or non-group if the obtained node for being composition figure and do not form figure, goes to step 6
At the node of figure；
Step 6 compares all nodes for not forming figure obtained in step 5 with the node of composition figure respectively
Match, obtains final composition figure or composition figure and the node for not forming figure.
Further, at least one composition figure is obtained in the step 2 or/and do not form the specific steps of the node of figure are as follows:
All nodes are divided into more parts by step 2.1, the server resource according to parallel computation；
Step 2.2 simultaneously is about to each part of node and compares with its internal node in order match；
Step 2.3, in step 2.2, the node of every portion and its internal node are all connected, and obtain multiple composition figures；
The node of every portion and its internal node are all not connected to, and obtain multiple nodes for not forming figure；
At least a node is connect with its internal node section, is obtained at least one composition figure or is not formed the section of figure
Point.
Further, at least one fusion composition figure or/and composition figure or/and the section for not forming figure are obtained in the step 3
The specific steps of point are as follows:
Step 3.1, the server resource according to parallel computation, by obtained multiple composition figures or/and the section for not forming figure
Point is divided into more parts；
Step 3.2 carries out the node in the composition figure in every portion with the node in the composition figure of its inside in order
Comparison matching one by one obtains at least one fusion composition figure and composition figure；Or/and
And the node for not forming figure for being about to each part compares with the node for not forming figure inside it match in order,
It obtains at least one composition figure or/and does not form the node of figure.
Further, the every number amount being divided into more parts are as follows: every two parts of data volume difference is no more than 10%.
The present invention compared with the existing technology, its advantages are shown in:
One, in the present invention, by the establishment time that the process of more parts of node cutting and parallel computation is reduced to figure significantly；
Two, in the present invention, there is no the limitations of the upper limit (can according to need for required memory source in calculating process
Configure resource and carry out parallel processing), calculating process can arrange more server resources according to the size of data volume.
Three, assembling process is carried out by way of traversal in the prior art, even space (i.e. the configuration of CPU) is high again, string
The mode of row assembling can not improve speed in addition to quicker in input or output, in specific assembling process at all, and configure
It is high also to waste source, and the present invention can carry out reasonable cutting according to the size and number of resource, then use parallel computation
Mode, not only improve N times than the prior art, per parallel primary, that is, improve the multiple of corresponding number of resources, and rationally
Each resource is utilized.
Four, in the present invention, when being divided into and (being cut into) more parts before each parallel computation, if quantity energy mean allocation
In the case of, certain method for selecting mean allocation, if in the case where being unable to mean allocation, best data volume difference had better not surpass
It crosses 10%, differs the smaller the better, be that the calculating in order to prevent more than quantity is slow, also limit and calculate fast resource apparatus.
Detailed description of the invention
Fig. 1 is the data variation block diagram in the specific embodiment in the present invention.
Specific embodiment
Below in conjunction with the drawings and the specific embodiments, the invention will be further described.
A method of based on graph structure distributed parallel mode construction figure, which is characterized in that following steps:
Step 1. prepares the node of required assembling figure；
The node that step 2. obtains step 1 is divided into more parts, it is parallel node in every part is carried out two-by-two between comparison
Match, obtain at least one composition figure or/and does not form the node of figure, the composition figure number judged, the number of nodes for not forming figure
Or calculate whether wheel number reaches given threshold value, if not up to, going to step 3, if reaching, going to step 5；At least one composition figure
Or/and do not form figure node refer at least one composition figure or/and at least one do not form the node of figure；
If step 3. not up to threshold value, and if step 2 obtain multiple composition figures, multiple composition figures that step 2 is obtained point
At more parts, and it is about to compare matching between the figure in every part carries out two-by-two, or/and
If step 2 obtains multiple nodes for not forming figure, and is about to multiple nodes for not forming figure that step 2 obtains and is divided into
More parts, parallel comparing between the node progress two-by-two in every part matches, and obtains at least one fusion composition figure or/and composition figure
Or/and the node of figure is not formed, at least one fusion composition figure or/and composition figure or/and the node of figure is not formed specifically refer to
At least one merges composition figure or/and at least one composition figure or/and at least one does not form the node of figure, herein otherly
The concrete meaning of the identical statement of side is also the same as herein；
If step 4, multiple composition figures obtain at least one fusion composition figure or/and composition after being divided into more parts of parallel computations
Figure, multiple nodes for not forming figure obtain at least one composition figure after being divided into more parts of parallel computations or/and do not form the node of figure,
Then the fusion of the former (multiple composition figures, which are divided into, obtains at least one fusion composition figure or/and composition figure after more parts of parallel computations)
(multiple nodes for not forming figure obtain at least one composition after being divided into more parts of parallel computations with the latter for composition figure or/and composition figure
Figure or/and do not form the node of figure) composition figure compare matching, obtain at least one and merge composition figure or/and composition figure
Or/and the node of figure is not formed；Judge composition figure number (as fusion composition figure or/and composition figure figure number), do not form
The number of nodes of figure calculates whether wheel number reaches given threshold value, if not up to, repeating step 3- step 4 if reaching and going to step
5；
Step 5, the node for obtaining composition figure or/and not forming figure, the composition figure in this step be fusion composition figure or/and
Otherwise composition figure obtains final composition figure or non-group if the obtained node for being composition figure and do not form figure, goes to step 6
At the node of figure；
Step 6 compares all nodes for not forming figure obtained in step 5 with the node of composition figure respectively
Match, obtains final composition figure or composition figure and the node for not forming figure.
Further, at least one composition figure is obtained in the step 2 or/and do not form the specific steps of the node of figure are as follows:
All nodes are divided into more parts by step 2.1, the server resource according to parallel computation；
Step 2.2 simultaneously is about to each part of node and compares with its internal node in order match；
Step 2.3, in step 2.2, the node of every portion and its internal node are all connected, and obtain multiple composition figures；
The node of every portion and its internal node are all not connected to, and obtain multiple nodes for not forming figure；
At least a node is connect with its internal node section, is obtained at least one composition figure or is not formed the section of figure
Point.
Further, at least one fusion composition figure or/and composition figure or/and the section for not forming figure are obtained in the step 3
The specific steps of point are as follows:
Step 3.1, the server resource according to parallel computation, by obtained multiple composition figures or/and the section for not forming figure
Point is divided into more parts；
Step 3.2 carries out the node in the composition figure in every portion with the node in the composition figure of its inside in order
Comparison matching one by one obtains at least one fusion composition figure and composition figure；Or/and
And the node for not forming figure for being about to each part compares with the node for not forming figure inside it match in order,
It obtains at least one composition figure or/and does not form the node of figure.
Further, the every number amount being divided into more parts are as follows: every two parts of data volume difference is no more than 10%.
Embodiment
As shown in Figure 1:
Step 1,1000 nodes for preparing required assembling figure；
1000 nodes are divided into 10 parts by step 2, the server resource for having 10 parallel computations；Every portion in 10 parts
It is carried out in respective server resource respective calculating (i.e. realization parallel computation), each server resource is will be every a
Node carries out comparison one by one with its internal node in order and matches；Amount to after calculating (i.e. comparison matching) and obtains 100
Composition figure, 200 do not form the point of figure, the composition figure number judged whether less than 15, the number of nodes that does not form figure it is whether small
In 45 or calculating wheel number whether less than 5, if one of them is less than, it is carried out step 5, by judgement, there is no less than setting
Threshold value, so to perform the next step rapid 3；
Step 3, the server resource for selecting 6 parallel computations, wherein calculating figure with 4, i.e., by 100 composition figures point
At 4 parts, its 25 internal composition is schemed, is successively selected in a figure by every part of 25 compositions figure, each server resource
Node carries out comparison one by one with the node in its internal composition figure in order and matches, and in many cases, does not have to all matchings
It is complete, so that it may it allows and establishes connection between figure and figure, the one point matching of a namely point of worst situation, if not worst condition,
Other situations can all realize Rapid matching, finally obtain 60 composition figures (comprising fusion composition figure and composition figure)；2 calculate
200 points of the node of composition figure are not divided into 2 parts, and every part of 100 points compare matching, finally in the way of in step 2
Obtain 20 composition figures and 40 points for not forming figure；
Step 4 is once compared 60 composition figures and 20 composition figures again, obtains 70 composition figures (comprising fusion
Composition figure and composition figure), this step obtains 70 composition figures, and 40 do not form the point of figure, then judge composition figure number (as
The figure number of fusion composition figure and composition figure) whether less than 15, the number of nodes that does not form figure whether less than 45 or whether calculate wheel number
Less than 5, if one of them is less than, rapid 5 are performed the next step, judging result is to execute step 5；
What step 5, the result of step 4 obtained is that (the composition figure obtained herein is to melt obtained in step 4 to 70 composition figures
It is combined into figure and composition figure) and 40 points for not forming figure, execute step 6；
Step 6 will form the points progress comparing calculation that Figure 70 composition figure and 40 do not form figure, obtain 68 groups
Cheng Tu, 20 do not form the point of figure, and 68 figures can be calculated by finally obtaining conclusion this 1000 points, 20 with any figure all
Not related acnode.
The above is only the representative embodiment in the numerous concrete application ranges of the present invention, to protection scope of the present invention not structure
At any restrictions.It is all using transformation or equivalence replacement and the technical solution that is formed, all fall within rights protection scope of the present invention it
It is interior.
Claims (4)
1. a kind of method based on graph structure distributed parallel mode construction figure, which is characterized in that following steps:
Step 1. prepares the node of required assembling figure；
Step 2. has the server resource of N number of parallel computation, and the node that step 1 is obtained is divided into more parts, and every portion is respective
Respective calculating is carried out in server resource, parallel the comparing between the node progress two-by-two in every part of server resource matches,
It obtains at least one composition figure or/and does not form the node of figure, the composition figure number judged, the number of nodes or meter that do not form figure
Calculate whether wheel number reaches given threshold value, if not up to, going to step 3, if reaching, going to step 5；Wherein, composition figure refers to often
A node connect with its internal node after figure；
If step 3. not up to threshold value, and if step 2 obtain multiple composition figures, multiple composition figures that step 2 is obtained are divided into more
Part, server resource simultaneously is about to compare matching between the figure in every part carries out two-by-two, or/and
If step 2 obtains multiple nodes for not forming figure, and be about to multiple nodes for not forming figure that step 2 obtains be divided into it is more
Part, parallel the comparing between the node progress two-by-two for not forming figure in every part of server resource matches, and obtains at least one and melts
It is combined into figure or/and composition figure or/and the node for not forming figure；Wherein, fusion composition figure refers to establishes between two composition figures
Figure after connection；
If step 4, multiple composition figures obtain at least one fusion composition figure or/and composition figure after being divided into more parts of parallel computations, more
A node for not forming figure obtains at least one composition figure after being divided into more parts of parallel computations or/and does not form the node of figure, service
Device resource then compares the former fusion composition figure or/and composition figure with the composition figure of the latter and matches, and obtains at least one
Merge composition figure or/and composition figure or/and the node for not forming figure；Composition figure number that server resource judges does not form
The number of nodes of figure calculates whether wheel number reaches given threshold value, if not up to, repeating step 3- step 4 if reaching and going to step
5；
Step 5, the node for obtaining composition figure or/and not forming figure, the composition figure in this step are fusion composition figure or/and composition
Figure, if the obtained node for being composition figure and do not form figure, goes to step 6, otherwise obtains final composition figure or does not form figure
Node；
Step 6, server resource carry out all nodes for not forming figure obtained in step 5 pair with the node of composition figure respectively
Than matching, final composition figure or composition figure and the node for not forming figure is obtained.
2. a kind of method based on graph structure distributed parallel mode construction figure according to claim 1, which is characterized in that
At least one composition figure is obtained in the step 2 or/and does not form the specific steps of the node of figure are as follows:
All nodes are divided into more parts by step 2.1, the server resource according to parallel computation；
Step 2.2 simultaneously is about to each part of node and compares with its internal node in order match；
Step 2.3, in step 2.2, the node of every portion and its internal node are all connected, and obtain multiple composition figures；
The node of every portion and its internal node are all not connected to, and obtain multiple nodes for not forming figure；
At least a node is connect with its internal node section, is obtained at least one composition figure or is not formed the node of figure.
3. a kind of method based on graph structure distributed parallel mode construction figure according to claim 2, which is characterized in that
At least one fusion composition figure or/and composition figure are obtained in the step 3 or/and do not form the specific steps of the node of figure are as follows:
Step 3.1, the server resource according to parallel computation, by obtained multiple composition figures or/and the node for not forming figure point
At more parts；
Step 3.2 carries out the node in the composition figure in every portion one by one with the node in the composition figure of its inside in order
Comparison matching obtains at least one fusion composition figure and composition figure；Or/and
And the node for not forming figure for being about to each part compares with the node for not forming figure inside it match in order, obtains
At least one composition figure or/and the node for not forming figure.
4. a kind of method based on graph structure distributed parallel mode construction figure according to claim 1-3,
It is characterized in that: the every number amount being divided into more parts are as follows: every two parts of data volume difference is no more than 10%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811568076.9A CN109710314B (en) | 2018-12-20 | 2018-12-20 | A method of based on graph structure distributed parallel mode construction figure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811568076.9A CN109710314B (en) | 2018-12-20 | 2018-12-20 | A method of based on graph structure distributed parallel mode construction figure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109710314A CN109710314A (en) | 2019-05-03 |
| CN109710314B true CN109710314B (en) | 2019-11-12 |
Family
ID=66257157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811568076.9A Active CN109710314B (en) | 2018-12-20 | 2018-12-20 | A method of based on graph structure distributed parallel mode construction figure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109710314B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114091686B (en) * | 2021-11-11 | 2022-10-14 | 北京百度网讯科技有限公司 | Data processing method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761453A (en) * | 2013-12-09 | 2014-04-30 | 天津工业大学 | Parallel gene splicing algorithm based on cluster map structure |
| CN104102699A (en) * | 2014-07-01 | 2014-10-15 | 华中科技大学 | Subgraph searching method in cluster graph set |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5511146A (en) * | 1991-06-26 | 1996-04-23 | Texas Instruments Incorporated | Excitory and inhibitory cellular automata for computational networks |
| CN101840329B (en) * | 2010-04-19 | 2012-10-03 | 浙江大学 | Data parallel processing method based on graph topological structure |
| CN103559016B (en) * | 2013-10-23 | 2016-09-07 | 江西理工大学 | A kind of Frequent tree mining method for digging based on graphic process unit parallel computation |
| CN106033342A (en) * | 2015-03-17 | 2016-10-19 | 阿里巴巴集团控股有限公司 | Method and system for generating random graph |
| CN105787865B (en) * | 2016-03-01 | 2018-09-07 | 西华大学 | Based on game engine and the graftal of GPU parallel processings generates and rendering intent |
| CN108073620A (en) * | 2016-11-14 | 2018-05-25 | 北京航天长峰科技工业集团有限公司 | A kind of method for quickly retrieving based on graph data structure |
| CN106599091B (en) * | 2016-11-24 | 2020-07-14 | 上海交通大学 | RDF graph structure storage and index method based on key value storage |
| CN106874422B (en) * | 2017-01-25 | 2019-07-26 | 东南大学 | A kind of figure querying method of facing relation type database |
| CN107330083B (en) * | 2017-07-03 | 2020-06-16 | 贵州大学 | Parallel construction method for equal-width histogram |
-
2018
- 2018-12-20 CN CN201811568076.9A patent/CN109710314B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103761453A (en) * | 2013-12-09 | 2014-04-30 | 天津工业大学 | Parallel gene splicing algorithm based on cluster map structure |
| CN104102699A (en) * | 2014-07-01 | 2014-10-15 | 华中科技大学 | Subgraph searching method in cluster graph set |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109710314A (en) | 2019-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110348571A (en) | A kind of neural network model training method, device, chip and system | |
| CN104298843B (en) | A kind of complex mechanism dynamic cascading Reliability Modeling based on colored stochastic Petri net | |
| CN103632001A (en) | Retention time sequence optimization method based on multiplexing of buffer unit | |
| CN103440364A (en) | Method and system for automatically generating WBS (Work Breakdown Structure) node basing on BIM (Building Information Modeling) | |
| CN102890735A (en) | Modeling method of strategy table and strategy table searching match of stable control device | |
| CN111181164B (en) | Improved master-slave split transmission and distribution cooperative power flow calculation method and system | |
| CN112100450A (en) | Graph calculation data segmentation method, terminal device and storage medium | |
| CN106786529B (en) | A kind of distribution static security analysis method | |
| CN105656039A (en) | Recognition method for vulnerable lines of power system | |
| CN103617226A (en) | Regular expression matching method and device | |
| CN115577675B (en) | Time sequence violation repairing method and device, server and readable storage medium | |
| CN109710314B (en) | A method of based on graph structure distributed parallel mode construction figure | |
| CN103258300A (en) | Method for designing cable bridge in cable passage of thermal power plant | |
| CN113886092A (en) | Computation graph execution method and device and related equipment | |
| CN107977504A (en) | A kind of asymmetric in-core fuel management computational methods, device and terminal device | |
| CN104036141B (en) | Open computing language (OpenCL)-based red-black tree acceleration method | |
| CN106326005A (en) | Automatic parameter tuning method for iterative MapReduce operation | |
| CN114239960A (en) | Distribution network project group progress management method and system based on dynamic resource optimization | |
| CN105958850A (en) | Optimized merging and sorting based modular multilevel converter capacitor voltage equalizing method | |
| CN102508720B (en) | Method for improving efficiency of preprocessing module and efficiency of post-processing module and system | |
| CN115718986B (en) | Multi-core parallel time domain simulation method based on distributed memory architecture | |
| CN104866611B (en) | A kind of practical method for fast searching of power system Active Splitting section | |
| CN105022896A (en) | Method and device for APDL modelling based on dynamic numbering | |
| CN102110151A (en) | Method for analyzing Petri network model of charge-off system | |
| CN110728299A (en) | Multi-extreme learning machine-based transient stability hierarchical evaluation method after power system fault |
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 |