CN107193650A - A kind of method and apparatus that video card resource is dispatched in distributed type assemblies - Google Patents
A kind of method and apparatus that video card resource is dispatched in distributed type assemblies Download PDFInfo
- Publication number
- CN107193650A CN107193650A CN201710250265.0A CN201710250265A CN107193650A CN 107193650 A CN107193650 A CN 107193650A CN 201710250265 A CN201710250265 A CN 201710250265A CN 107193650 A CN107193650 A CN 107193650A
- Authority
- CN
- China
- Prior art keywords
- video card
- pci
- buses
- resource
- job request
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/5038—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the execution order of a plurality of tasks, e.g. taking priority or time dependency constraints into consideration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/5021—Priority
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Television Signal Processing For Recording (AREA)
Abstract
The invention discloses a kind of method and apparatus that video card resource is dispatched in distributed type assemblies.Wherein method includes：The video card resource in distributed type assemblies is obtained, the available video card quantity in each PCI E buses is recorded in video card scheduling of resource table；The operation submitted is received, the video card quantity of the job request is included in the operation；Search the video card scheduling of resource table, when the available video card quantity in a PCI E bus meets the video card quantity of the job request, selection and the video card of the quantity of the video card quantity Matching of the job request are used as the video card resource for distributing to the operation from the PCI E buses.The technical scheme can ensure as much as possible the operation each submitted using need not across PCI E buses video cards for being communicated perform, avoid inefficient caused by across PCI E bus communications, there is larger improved efficiency to video card resource requirement very high homework type for deep learning operation etc., dispatch granularity very thin, meet the demand of distributed type assemblies.
Description
Technical field
The present invention relates to field of computer technology, and in particular to a kind of method that video card resource is dispatched in distributed type assemblies
And device.
Background technology
Resource Scheduler has had many kinds, such as k8s to explorer in current distributed type assemblies in other words,
Mesos, yarn etc..But they can not dispatch video card resource well, and for the need to video card resource such as deep learning
Ask for higher calculating task, the quality for the video card resource being assigned to will largely effect on the execution efficiency of calculating task.
The content of the invention
In view of the above problems, it is proposed that the present invention so as to provide one kind overcome above mentioned problem or at least in part solve on
State the method and apparatus that video card resource is dispatched in distributed type assemblies of problem.
According to one aspect of the present invention there is provided a kind of method that video card resource is dispatched in distributed type assemblies, including：
The video card resource in distributed type assemblies is obtained, records available in each PCI-E buses in video card scheduling of resource table
Video card quantity；
The operation submitted is received, the video card quantity of the job request is included in the operation；
The video card scheduling of resource table is searched, when the available video card quantity in a PCI-E bus meets the job request
Video card quantity when, from the PCI-E buses video card of selection and the quantity of the video card quantity Matching of the job request as point
The video card resource of the dispensing operation.
Alternatively, the video card resource obtained in distributed type assemblies includes：
The video card money on the computing device is read in the PCI-E buses for each computing device disposed from distributed type assemblies
Source.
Alternatively, the available video card quantity recorded in video card scheduling of resource table in each PCI-E buses includes：
The available video card ID in each PCI-E buses is recorded in open chain table, and by the available video card number in each PCI-E buses
Amount is ranked up.
Alternatively, described to be ordered as ascending order, the lookup video card scheduling of resource table includes：
The open chain table is traveled through by depth-priority-searching method, judges whether the available video card quantity in each PCI-E buses is full
The video card quantity of the foot job request.
Alternatively, when the available video card quantity in all PCI-E buses is unsatisfactory for the video card quantity of the job request,
Travel through the open chain table again by depth-priority-searching method, selection and the video card number of the job request from multiple PCI-E buses
The video card of flux matched quantity is used as the video card resource for distributing to the operation.
Alternatively, travel through the open chain table again by depth-priority-searching method, the selection from multiple PCI-E buses with
The video card of the quantity of the video card quantity Matching of the job request includes as the video card resource for distributing to the operation：
All available video cards in the first PCI-E buses found are distributed into the operation, judge that next PCI-E is total
Whether the available video card quantity on line meets the remaining video card quantity of the job request, is selected if meeting from the PCI-E buses
The video card with the quantity of the video card quantity Matching of the job request is taken as the video card resource for distributing to the operation, if being unsatisfactory for
All available video cards in the PCI-E buses are distributed into the operation, and judge the available video card number in next PCI-E buses
Amount whether meets the remaining video card quantity of the job request until the remaining video card quantity of the job request is satisfied.
Alternatively, this method also includes：
All available video cards distributed for the operation are deleted from the open chain table, and the open chain table carried out again
Sequence；
And/or,
The open chain table according to the video card resource modifying of release, and the open chain table is ranked up again.
According to another aspect of the present invention there is provided a kind of device that video card resource is dispatched in distributed type assemblies, including：
Recording unit, suitable for obtaining the video card resource in distributed type assemblies, each PCI-E is recorded in video card scheduling of resource table
Available video card quantity in bus；
Scheduling unit, suitable for receiving the operation submitted, includes the video card quantity of the job request, searches institute in the operation
Video card scheduling of resource table is stated, when the available video card quantity in a PCI-E bus meets the video card quantity of the job request, from
Selection and the video card of the quantity of the video card quantity Matching of the job request are used as the video card for distributing to the operation in the PCI-E buses
Resource.
Alternatively, the recording unit, suitable for being read in the PCI-E buses for each computing device disposed from distributed type assemblies
Take the video card resource on the computing device.
Alternatively, the recording unit, suitable for recording the available video card ID in each PCI-E buses in open chain table, and is pressed
Available video card quantity in each PCI-E buses is ranked up.
Alternatively, the recording unit carries out ascending sort in open chain table；
The scheduling unit, suitable for traveling through the open chain table by depth-priority-searching method, judge in each PCI-E buses can
The video card quantity of the job request whether is met with video card quantity.
Alternatively, the scheduling unit, is further adapted for when the available video card quantity in all PCI-E buses is unsatisfactory for the work
During the video card quantity of industry application, travel through the open chain table again by depth-priority-searching method, from multiple PCI-E buses selection with
The video card of the quantity of the video card quantity Matching of the job request is used as the video card resource for distributing to the operation.
Alternatively, the scheduling unit, suitable for all available video cards in the first PCI-E buses found are distributed to
The operation, judges whether the available video card quantity in next PCI-E buses meets the remaining video card quantity of the job request, if
Meet and then chosen from the PCI-E buses with the video card of the quantity of the video card quantity Matching of the job request as distributing to the work
All available video cards in the PCI-E buses are distributed to the operation, and judge next by the video card resource of industry if being unsatisfactory for
Available video card quantity in PCI-E buses whether meets the remaining video card quantity of the job request until the residue of the job request
Video card quantity is satisfied.
Alternatively, the recording unit, all available video cards for being adapted for operation distribution are deleted from the open chain table
Remove, and the open chain table is ranked up again；And/or, suitable for open chain table described in the video card resource modifying according to release, lay equal stress on
Newly the open chain table is ranked up.
From the foregoing, technical scheme, aobvious at one after the video card resource in getting distributed type assemblies
The available video card quantity in each PCI-E buses is recorded in card scheduling of resource table, when the work for receiving the video card quantity comprising application
During industry, video card scheduling of resource table is searched, therefrom selection can meet the PCI-E buses of the video card quantity of the job request, from this
The video card of respective numbers is distributed in bus to the operation.The technical scheme can ensure that the operation each submitted is sharp as much as possible
With need not the video card that is communicated of across PCI-E bus perform, it is to avoid it is inefficient caused by across PCI-E bus communication,
There is larger improved efficiency to video card resource requirement very high homework type for deep learning operation etc., scheduling granularity is very
Carefully, the demand of distributed type assemblies is met.
Described above is only the general introduction of technical solution of the present invention, in order to better understand the technological means of the present invention,
And can be practiced according to the content of specification, and in order to allow above and other objects of the present invention, feature and advantage can
Become apparent, below especially exemplified by the embodiment of the present invention.
Brief description of the drawings
By reading the detailed description of hereafter preferred embodiment, various other advantages and benefit is common for this area
Technical staff will be clear understanding.Accompanying drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And in whole accompanying drawing, identical part is denoted by the same reference numerals.In the accompanying drawings：
Fig. 1 shows a kind of method that video card resource is dispatched in distributed type assemblies according to an embodiment of the invention
Schematic flow sheet；
Fig. 2 shows a kind of device that video card resource is dispatched in distributed type assemblies according to an embodiment of the invention
Structural representation.
Embodiment
The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the disclosure in accompanying drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
Limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
Complete conveys to those skilled in the art.
Fig. 1 shows a kind of method that video card resource is dispatched in distributed type assemblies according to an embodiment of the invention
Schematic flow sheet, as shown in figure 1, this method includes：
Step S110, obtains the video card resource in distributed type assemblies, each PCI-E buses is recorded in video card scheduling of resource table
On available video card quantity.
PCI-E (PCI-Express) bus is bus protocol newer at present, in current most of computing devices, video card,
The equipment such as network interface card are all connected in PCI-E buses.
Step S120, receives the operation submitted, the video card quantity of the job request is included in operation.
Step S130, searches video card scheduling of resource table, when the available video card quantity in a PCI-E bus meets the operation
During the video card quantity of application, selection and the video card of the quantity of the video card quantity Matching of the job request are made from the PCI-E buses
To distribute to the video card resource of the operation.
If in practice, it has been found that multiple video cards that an operation is assigned to need across PCI-E bus communication, then effect
Rate will become especially low, and be connected to if multiple video cards in same PCI-E buses relatively high if efficiency.To avoid across total
The situation of line communication proposes the present embodiment.
It can be seen that, the method shown in Fig. 1 after the video card resource in getting distributed type assemblies, is adjusted in a video card resource
The available video card quantity in each PCI-E buses is recorded in degree table, when receiving the operation of the video card quantity comprising application, is searched
Video card scheduling of resource table, therefrom selection can meet the PCI-E buses of the video card quantity of the job request, be distributed from the bus
The video card of respective numbers gives the operation.The technical scheme can ensure the operation each submitted using need not be across as much as possible
The video card that PCI-E buses are communicated is performed, it is to avoid inefficient caused by across PCI-E bus communication, for depth
Exercises industry etc. has larger improved efficiency to video card resource requirement very high homework type, and scheduling granularity is very thin, meets distribution
The demand of formula cluster.
In one embodiment of the invention, in the above method, the video card resource obtained in distributed type assemblies includes：From point
The video card resource on the computing device is read in the PCI-E buses for each computing device disposed in cloth cluster.
For example, all devices checked by lspci orders in PCI-E buses, are therefrom filtering out video card resource.
In one embodiment of the invention, in the above method, recorded in video card scheduling of resource table in each PCI-E buses
Available video card quantity include：The available video card ID in each PCI-E buses is recorded in open chain table, and by each PCI-E buses
Available video card quantity be ranked up.
For example：Video card parent in the same PCI-E buses of PCI-E 0 [GPU0, GPU1], PCI-E 1 [GPU2, GPU3] ...
High, such as GPU0 and GPU1 with degree.A video card scheduling of resource table is thus obtained.Ensuing work is how to realize
Distribute affinity high video card for operation.In one embodiment of the invention, in the above method, ascending order is ordered as, is searched aobvious
Card scheduling of resource table includes：Open chain table is traveled through by depth-priority-searching method, judges that the available video card quantity in each PCI-E buses is
The no video card quantity for meeting the job request.
In above method, if operation needs 1 video card, then the video card on obvious PCI-E 0, PCI-E 1 all may be used
To meet condition, by taking order above as an example, the GPU 0 that the PCI-E 0 first found reaches the standard grade can just be assigned to the operation.
And if lower example：PCI-E 0 [GPU0], PCI-E 1 [GPU1, GPU2, GPU3], operation need 2 video cards, that
Video card on PCI-E 0 can't distribute to the GPU1 on the operation, PCI-E 1, and GPU2 can be assigned to the operation.Depth
Priority algorithm can be saved the time, fast dispatch to the video card for meeting job requirements.So problem is, the above method can expire
The less operation of sufficient video card quantity required, when the available video card quantity in all PCI-E buses is unsatisfactory for the job request
It can not just be handled during video card quantity.
Thus in one embodiment of the invention, in the above method, when the available video card quantity in all PCI-E buses
When being unsatisfactory for the video card quantity of the job request, open chain table is traveled through again by depth-priority-searching method, from multiple PCI-E buses
Upper selection and the video card of the quantity of the video card quantity Matching of the job request are used as the video card resource for distributing to the operation.Thus
Solve the problem.
However, the problem of second of traversal also brings along new.For example, when operation 4 pieces of video cards of needs, and it is current
PCI-E 0 on available video card be GPU0, available video card is that available video card is on GPU1, PCI-E 2 on PCI-E 1
Available video card is GPU4, GPU5 on GPU2, GPU3, PCI-E 3.So at this time, it is selection PCI-E 2, PCI-E 3 group
Conjunction mode more preferably, or selects PCI-E 0, PCI-E 1, PCI-E 2 combination to be more preferably worth what is probed into.
Due to above two mode be required for video card across bus communication, then in order that the fragment that must remain is less, I
Select PCI-E 0, PCI-E 1, PCI-E 2 combination.So for the selection for realizing this, at one of the present invention
In embodiment, in the above method, open chain table is traveled through again by depth-priority-searching method, selection is with being somebody's turn to do from multiple PCI-E buses
The video card of the quantity of the video card quantity Matching of job request includes as the video card resource for distributing to the operation：By the head found
All available video cards in individual PCI-E buses distribute to the operation, judge that the available video card quantity in next PCI-E buses is
The no remaining video card quantity for meeting the job request, chooses and the video card of the job request if meeting from the PCI-E buses
The video card of the quantity of quantity Matching is as the video card resource for distributing to the operation, by the institute in the PCI-E buses if being unsatisfactory for
There is available video card to distribute to the operation, and judge whether the available video card quantity in next PCI-E buses meets the operation Shen
Remaining video card quantity please is until the remaining video card quantity of the job request is satisfied.
To be operation distribution in the above method in one embodiment of the invention to ensure the accuracy of scheduling
All available video cards deleted from open chain table, and split chained list is ranked up again；And/or, according to the video card resource of release
Open chain table is changed, and split chained list is ranked up again.This assures the correct implementation of above-mentioned dispatching algorithm.
Fig. 2 shows a kind of device that video card resource is dispatched in distributed type assemblies according to an embodiment of the invention
Structural representation, as shown in Fig. 2 the device 200 of video card resource is dispatched in distributed type assemblies to be included：
Recording unit 210, suitable for obtaining the video card resource in distributed type assemblies, records each in video card scheduling of resource table
Available video card quantity in PCI-E buses.
Scheduling unit 220, suitable for receiving the operation submitted, includes the video card quantity of the job request in operation, search aobvious
Card scheduling of resource table, when the available video card quantity in a PCI-E bus meets the video card quantity of the job request, from this
Selection and the video card of the quantity of the video card quantity Matching of the job request are provided as the video card for distributing to the operation in PCI-E buses
Source.
If in practice, it has been found that multiple video cards that an operation is assigned to need across PCI-E bus communication, then effect
Rate will become especially low, and be connected to if multiple video cards in same PCI-E buses relatively high if efficiency.To avoid across total
The situation of line communication proposes the present embodiment.
It can be seen that, the device shown in Fig. 2, by the mutual cooperation of each unit, the video card money in distributed type assemblies are got
Behind source, the available video card quantity in each PCI-E buses is recorded in a video card scheduling of resource table, application is included when receiving
During the operation of video card quantity, video card scheduling of resource table is searched, therefrom selection can meet the video card quantity of the job request
PCI-E buses, distribute the video card of respective numbers to the operation from the bus.The technical scheme can ensure each as much as possible
The operation of submission all using need not across PCI-E bus video card for being communicated perform, it is to avoid because of across PCI-E bus communication
Caused poor efficiency, has larger efficiency to carry video card resource requirement very high homework type for deep learning operation etc.
Rise, scheduling granularity is very thin, meets the demand of distributed type assemblies.
In one embodiment of the invention, in said apparatus, recording unit 210, suitable for being disposed from distributed type assemblies
Each computing device PCI-E buses in read video card resource on the computing device.
For example, all devices checked by lspci orders in PCI-E buses, are therefrom filtering out video card resource.
In one embodiment of the invention, in said apparatus, recording unit 210, suitable for recording each in open chain table
Available video card ID in PCI-E buses, and be ranked up by the available video card quantity in each PCI-E buses.
For example：Video card parent in the same PCI-E buses of PCI-E 0 [GPU0, GPU1], PCI-E 1 [GPU2, GPU3] ...
High, such as GPU0 and GPU1 with degree.A video card scheduling of resource table is thus obtained.Ensuing work is how to realize
Distribute affinity high video card for operation.In one embodiment of the invention, in said apparatus, recording unit 210 is in open chain
Ascending sort is carried out in table；Scheduling unit 220, suitable for traveling through open chain table by depth-priority-searching method, judges in each PCI-E buses
Available video card quantity whether meet the video card quantity of the job request.
In above example, if operation needs 1 video card, then the video card on obvious PCI-E 0, PCI-E 1 all may be used
To meet condition, by taking order above as an example, the GPU 0 that the PCI-E 0 first found reaches the standard grade can just be assigned to the operation.
And if lower example：PCI-E 0 [GPU0], PCI-E 1 [GPU1, GPU2, GPU3], operation need 2 video cards, that
Video card on PCI-E 0 can't distribute to the GPU1 on the operation, PCI-E 1, and GPU2 can be assigned to the operation.Depth
Priority algorithm can be saved the time, fast dispatch to the video card for meeting job requirements.So problem is, the above method can expire
The less operation of sufficient video card quantity required, when the available video card quantity in all PCI-E buses is unsatisfactory for the job request
It can not just be handled during video card quantity.
Thus in one embodiment of the invention, in said apparatus, scheduling unit 220 is further adapted for when all PCI-E are total
When available video card quantity on line is unsatisfactory for the video card quantity of the job request, open chain is traveled through by depth-priority-searching method again
Table, the video card of selection and the quantity of the video card quantity Matching of the job request is as distributing to the work from multiple PCI-E buses
The video card resource of industry.
However, the problem of second of traversal also brings along new.For example, when operation 4 pieces of video cards of needs, and it is current
PCI-E 0 on available video card be GPU0, available video card is that available video card is on GPU1, PCI-E 2 on PCI-E 1
Available video card is GPU4, GPU5 on GPU2, GPU3, PCI-E 3.So at this time, it is selection PCI-E 2, PCI-E 3 group
Conjunction mode more preferably, or selects PCI-E 0, PCI-E 1, PCI-E 2 combination to be more preferably worth what is probed into.
Due to above two mode be required for video card across bus communication, then in order that the fragment that must remain is less, I
Select PCI-E 0, PCI-E 1, PCI-E 2 combination.So for the selection for realizing this, at one of the present invention
In embodiment, in said apparatus, scheduling unit 220, suitable for all available video cards in the first PCI-E buses found are divided
The dispensing operation, judges whether the available video card quantity in next PCI-E buses meets the remaining video card number of the job request
Amount, selection and the video card of the quantity of the video card quantity Matching of the job request are used as distribution from the PCI-E buses if meeting
To the video card resource of the operation, all available video cards in the PCI-E buses are distributed into the operation if being unsatisfactory for, and judge
Available video card quantity in next PCI-E buses whether meets the remaining video card quantity of the job request until the job request
Remaining video card quantity be satisfied.
In order to ensure the accuracy of scheduling, in one embodiment of the invention, in said apparatus, recording unit 210 is fitted
In will be deleted for all available video cards that the operation is distributed from open chain table, and split chained list is ranked up again；And/or, fit
In the video card resource modifying open chain table according to release, and split chained list is ranked up again.This assures above-mentioned scheduling calculation
The correct implementation of method.
In summary, technical scheme, after the video card resource in getting distributed type assemblies, in a video card
The available video card quantity in each PCI-E buses is recorded in scheduling of resource table, when the operation for receiving the video card quantity comprising application
When, video card scheduling of resource table is searched, therefrom selection can meet the PCI-E buses of the video card quantity of the job request, total from this
The video card of respective numbers is distributed on line to the operation.The technical scheme can ensure that the operation each submitted is utilized as much as possible
Need not the video card that is communicated of across PCI-E bus perform, it is to avoid it is inefficient caused by across PCI-E bus communication, it is right
There is larger improved efficiency to video card resource requirement very high homework type in deep learning operation etc., scheduling granularity is very thin,
Meet the demand of distributed type assemblies.
It should be noted that：
Algorithm and display be not inherently related to any certain computer, virtual bench or miscellaneous equipment provided herein.
Various fexible units can also be used together with based on teaching in this.As described above, construct required by this kind of device
Structure be obvious.In addition, the present invention is not also directed to any certain programmed language.It is understood that, it is possible to use it is various
Programming language realizes the content of invention described herein, and the description done above to language-specific is to disclose this hair
Bright preferred forms.
In the specification that this place is provided, numerous specific details are set forth.It is to be appreciated, however, that the implementation of the present invention
Example can be put into practice in the case of these no details.In some instances, known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this description.
Similarly, it will be appreciated that in order to simplify the disclosure and help to understand one or more of each inventive aspect, exist
Above in the description of the exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the method for the disclosure should be construed to reflect following intention：It is i.e. required to protect
The application claims of shield features more more than the feature being expressly recited in each claim.More precisely, such as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following embodiment are expressly incorporated in the embodiment, wherein each claim is in itself
All as the separate embodiments of the present invention.
Those skilled in the art, which are appreciated that, to be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more equipment different from the embodiment.Can be the module or list in embodiment
Member or component be combined into a module or unit or component, and can be divided into addition multiple submodule or subelement or
Sub-component.In addition at least some in such feature and/or process or unit exclude each other, it can use any
Combination is disclosed to all features disclosed in this specification (including adjoint claim, summary and accompanying drawing) and so to appoint
Where all processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (including adjoint power
Profit is required, summary and accompanying drawing) disclosed in each feature can or similar purpose identical, equivalent by offer alternative features come generation
Replace.
Although in addition, it will be appreciated by those of skill in the art that some embodiments described herein include other embodiments
In included some features rather than further feature, but the combination of the feature of be the same as Example does not mean in of the invention
Within the scope of and form different embodiments.For example, in the following claims, times of embodiment claimed
One of meaning mode can be used in any combination.
The present invention all parts embodiment can be realized with hardware, or with one or more processor run
Software module realize, or realized with combinations thereof.It will be understood by those of skill in the art that can use in practice
Microprocessor or digital signal processor (DSP) according to embodiments of the present invention dispatch video card to realize in distributed type assemblies
The some or all functions of some or all parts in the device of resource.The present invention is also implemented as being used to perform this
In described method some or all equipment or program of device (for example, computer program and computer program
Product).Such program for realizing the present invention can be stored on a computer-readable medium, or can have one or many
The form of individual signal.Such signal can be downloaded from internet website and obtained, either on carrier signal provide or with
Any other form is provided.
It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and ability
Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between bracket should not be configured to limitations on claims.Word "comprising" is not excluded the presence of not
Element or step listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of some different elements and coming real by means of properly programmed computer
It is existing.In if the unit claim of equipment for drying is listed, several in these devices can be by same hardware branch
To embody.The use of word first, second, and third does not indicate that any order.These words can be explained and run after fame
Claim.
Embodiment of the invention discloses that A1, a kind of method that video card resource is dispatched in distributed type assemblies, wherein, the party
Method includes：
The video card resource in distributed type assemblies is obtained, records available in each PCI-E buses in video card scheduling of resource table
Video card quantity；
The operation submitted is received, the video card quantity of the job request is included in the operation；
The video card scheduling of resource table is searched, when the available video card quantity in a PCI-E bus meets the job request
Video card quantity when, from the PCI-E buses video card of selection and the quantity of the video card quantity Matching of the job request as point
The video card resource of the dispensing operation.
A2, the method as described in A1, wherein, the video card resource obtained in distributed type assemblies includes：
The video card money on the computing device is read in the PCI-E buses for each computing device disposed from distributed type assemblies
Source.
A3, the method as described in A1, wherein, it is described recorded in video card scheduling of resource table in each PCI-E buses it is available
Video card quantity includes：
The available video card ID in each PCI-E buses is recorded in open chain table, and by the available video card number in each PCI-E buses
Amount is ranked up.
A4, the method as described in A3, wherein, described to be ordered as ascending order, the lookup video card scheduling of resource table includes：
The open chain table is traveled through by depth-priority-searching method, judges whether the available video card quantity in each PCI-E buses is full
The video card quantity of the foot job request.
A5, the method as described in A4, wherein, when the available video card quantity in all PCI-E buses is unsatisfactory for the operation
During the video card quantity of application, the open chain table is traveled through again by depth-priority-searching method, selection is with being somebody's turn to do from multiple PCI-E buses
The video card of the quantity of the video card quantity Matching of job request is used as the video card resource for distributing to the operation.
A6, the method as described in A5, wherein, travel through the open chain table again by depth-priority-searching method, it is described from multiple
Selection and the video card of the quantity of the video card quantity Matching of the job request are provided as the video card for distributing to the operation in PCI-E buses
Source includes：
All available video cards in the first PCI-E buses found are distributed into the operation, judge that next PCI-E is total
Whether the available video card quantity on line meets the remaining video card quantity of the job request, is selected if meeting from the PCI-E buses
The video card with the quantity of the video card quantity Matching of the job request is taken as the video card resource for distributing to the operation, if being unsatisfactory for
All available video cards in the PCI-E buses are distributed into the operation, and judge the available video card number in next PCI-E buses
Amount whether meets the remaining video card quantity of the job request until the remaining video card quantity of the job request is satisfied.
A7, the method as described in A3, wherein, this method also includes：
All available video cards distributed for the operation are deleted from the open chain table, and the open chain table carried out again
Sequence；
And/or,
The open chain table according to the video card resource modifying of release, and the open chain table is ranked up again.
Embodiments of the invention also disclose B8, a kind of device that video card resource is dispatched in distributed type assemblies, wherein, should
Device includes：
Recording unit, suitable for obtaining the video card resource in distributed type assemblies, each PCI-E is recorded in video card scheduling of resource table
Available video card quantity in bus；
Scheduling unit, suitable for receiving the operation submitted, includes the video card quantity of the job request, searches institute in the operation
Video card scheduling of resource table is stated, when the available video card quantity in a PCI-E bus meets the video card quantity of the job request, from
Selection and the video card of the quantity of the video card quantity Matching of the job request are used as the video card for distributing to the operation in the PCI-E buses
Resource.
B9, the device as described in B8, wherein,
The meter is read in the recording unit, the PCI-E buses suitable for each computing device disposed from distributed type assemblies
Calculate the video card resource in equipment.
B10, the device as described in B8, wherein,
The recording unit, suitable for recording the available video card ID in each PCI-E buses in open chain table, and by each PCI-E
Available video card quantity in bus is ranked up.
B11, the device as described in B10, wherein,
The recording unit carries out ascending sort in open chain table；
The scheduling unit, suitable for traveling through the open chain table by depth-priority-searching method, judge in each PCI-E buses can
The video card quantity of the job request whether is met with video card quantity.
B12, the device as described in B11, wherein,
The scheduling unit, is further adapted for when the available video card quantity in all PCI-E buses is unsatisfactory for the job request
Video card quantity when, travel through the open chain table again by depth-priority-searching method, from multiple PCI-E buses selection with the operation
The video card of the quantity of the video card quantity Matching of application is used as the video card resource for distributing to the operation.
B13, the device as described in B12, wherein,
The scheduling unit, suitable for all available video cards in the first PCI-E buses found are distributed into the operation,
Judge whether the available video card quantity in next PCI-E buses meets the remaining video card quantity of the job request, if meeting
Chosen from the PCI-E buses with the video card of the quantity of the video card quantity Matching of the job request as distributing to the aobvious of the operation
All available video cards in the PCI-E buses are distributed to the operation, and judge next PCI-E by card resource if being unsatisfactory for
Available video card quantity in bus whether meets the remaining video card quantity of the job request until the remaining video card of the job request
Quantity is satisfied.
B14, the device as described in B10, wherein,
The recording unit, all available video cards for being adapted for operation distribution are deleted from the open chain table, are laid equal stress on
Newly the open chain table is ranked up；And/or, suitable for open chain table described in the video card resource modifying according to release, and again to institute
Open chain table is stated to be ranked up.
Claims (10)
1. a kind of method that video card resource is dispatched in distributed type assemblies, wherein, this method includes：
The video card resource in distributed type assemblies is obtained, the available video card in each PCI-E buses is recorded in video card scheduling of resource table
Quantity；
The operation submitted is received, the video card quantity of the job request is included in the operation；
The video card scheduling of resource table is searched, when the available video card quantity in a PCI-E bus meets the aobvious of the job request
During card quantity, selection and the video card of the quantity of the video card quantity Matching of the job request are as distributing to from the PCI-E buses
The video card resource of the operation.
2. the method for claim 1, wherein the video card resource obtained in distributed type assemblies includes：
The video card resource on the computing device is read in the PCI-E buses for each computing device disposed from distributed type assemblies.
3. the method for claim 1, wherein it is described recorded in video card scheduling of resource table in each PCI-E buses can
Included with video card quantity：
The available video card ID in each PCI-E buses is recorded in open chain table, and is entered by the available video card quantity in each PCI-E buses
Row sequence.
4. method as claimed in claim 3, wherein, it is described to be ordered as ascending order, the lookup video card scheduling of resource table bag
Include：
The open chain table is traveled through by depth-priority-searching method, judges whether the available video card quantity in each PCI-E buses meets this
The video card quantity of job request.
5. method as claimed in claim 4, wherein, when the available video card quantity in all PCI-E buses is unsatisfactory for the work
During the video card quantity of industry application, travel through the open chain table again by depth-priority-searching method, from multiple PCI-E buses selection with
The video card of the quantity of the video card quantity Matching of the job request is used as the video card resource for distributing to the operation.
6. method as claimed in claim 5, wherein, travel through the open chain table again by depth-priority-searching method, it is described from many
Selection and the video card of the quantity of the video card quantity Matching of the job request are used as the video card for distributing to the operation in individual PCI-E buses
Resource includes：
All available video cards in the first PCI-E buses found are distributed into the operation, judged in next PCI-E buses
Available video card quantity whether meet the remaining video card quantity of the job request, chosen if meeting from the PCI-E buses with
The video card of the quantity of the video card quantity Matching of the job request, should if being unsatisfactory for as the video card resource for distributing to the operation
All available video cards in PCI-E buses distribute to the operation, and judge that the available video card quantity in next PCI-E buses is
The no remaining video card quantity for meeting the job request is until the remaining video card quantity of the job request is satisfied.
7. method as claimed in claim 3, wherein, this method also includes：
All available video cards distributed for the operation are deleted from the open chain table, and the open chain table arranged again
Sequence；
And/or,
The open chain table according to the video card resource modifying of release, and the open chain table is ranked up again.
8. a kind of device that video card resource is dispatched in distributed type assemblies, wherein, the device includes：
Recording unit, suitable for obtaining the video card resource in distributed type assemblies, records each PCI-E buses in video card scheduling of resource table
On available video card quantity；
Scheduling unit, suitable for receiving the operation submitted, includes the video card quantity of the job request in the operation, search described aobvious
Card scheduling of resource table, when the available video card quantity in a PCI-E bus meets the video card quantity of the job request, from this
Selection and the video card of the quantity of the video card quantity Matching of the job request are provided as the video card for distributing to the operation in PCI-E buses
Source.
9. device as claimed in claim 8, wherein,
The calculating is read in the recording unit, the PCI-E buses suitable for each computing device disposed from distributed type assemblies to set
Standby upper video card resource.
10. device as claimed in claim 8, wherein,
The recording unit, suitable for recording the available video card ID in each PCI-E buses in open chain table, and by each PCI-E buses
On available video card quantity be ranked up.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710250265.0A CN107193650B (en) | 2017-04-17 | 2017-04-17 | Method and device for scheduling display card resources in distributed cluster |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710250265.0A CN107193650B (en) | 2017-04-17 | 2017-04-17 | Method and device for scheduling display card resources in distributed cluster |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107193650A true CN107193650A (en) | 2017-09-22 |
| CN107193650B CN107193650B (en) | 2021-01-19 |
Family
ID=59871030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710250265.0A Active CN107193650B (en) | 2017-04-17 | 2017-04-17 | Method and device for scheduling display card resources in distributed cluster |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107193650B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109144578A (en) * | 2018-06-28 | 2019-01-04 | 中国船舶重工集团公司第七0九研究所 | A kind of video card resource allocation method and device based on Godson computer |
| CN115129483A (en) * | 2022-09-01 | 2022-09-30 | 武汉凌久微电子有限公司 | Multi-display-card cooperative display method based on display area division |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6421755B1 (en) * | 1999-05-26 | 2002-07-16 | Dell Usa, L.P. | System resource assignment for a hot inserted device |
| US20060195675A1 (en) * | 2005-02-25 | 2006-08-31 | International Business Machines Corporation | Association of host translations that are associated to an access control level on a PCI bridge that supports virtualization |
| CN101387993A (en) * | 2007-09-14 | 2009-03-18 | 凹凸科技（中国）有限公司 | Method and system for dynamically collocating resource for equipment in computer system |
| US20100007668A1 (en) * | 2008-07-08 | 2010-01-14 | Casparian Mark A | Systems and methods for providing scalable parallel graphics rendering capability for information handling systems |
| US20100036995A1 (en) * | 2008-08-05 | 2010-02-11 | Hitachi, Ltd. | Computer system and bus assignment method |
| CN101916209A (en) * | 2010-08-06 | 2010-12-15 | 华东交通大学 | Cluster task resource allocation method for multi-core processor |
| US20110157193A1 (en) * | 2009-12-29 | 2011-06-30 | Nvidia Corporation | Load balancing in a system with multi-graphics processors and multi-display systems |
| CN102609978A (en) * | 2012-01-13 | 2012-07-25 | 中国人民解放军信息工程大学 | Method for accelerating cone-beam CT (computerized tomography) image reconstruction by using GPU (graphics processing unit) based on CUDA (compute unified device architecture) architecture |
| CN102902589A (en) * | 2012-08-31 | 2013-01-30 | 浪潮电子信息产业股份有限公司 | Method for managing and scheduling cluster MIS (Many Integrated Core) job |
| US20130091500A1 (en) * | 2011-10-10 | 2013-04-11 | William J. Earl | Paravirtualized virtual gpu |
| CN103105895A (en) * | 2011-11-15 | 2013-05-15 | 辉达公司 | Computer system and display cards thereof and method for processing graphs of computer system |
| CN103248659A (en) * | 2012-02-13 | 2013-08-14 | 北京华胜天成科技股份有限公司 | Method and system for dispatching cloud computed resources |
| CN104954400A (en) * | 2014-03-27 | 2015-09-30 | 中国电信股份有限公司 | Cloud computing system and realizing method thereof |
| CN105718316A (en) * | 2014-12-01 | 2016-06-29 | 中国移动通信集团公司 | Job scheduling method and apparatus |
| CN106557366A (en) * | 2015-09-28 | 2017-04-05 | 阿里巴巴集团控股有限公司 | Task distribution method, apparatus and system |
-
2017
- 2017-04-17 CN CN201710250265.0A patent/CN107193650B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6421755B1 (en) * | 1999-05-26 | 2002-07-16 | Dell Usa, L.P. | System resource assignment for a hot inserted device |
| US20060195675A1 (en) * | 2005-02-25 | 2006-08-31 | International Business Machines Corporation | Association of host translations that are associated to an access control level on a PCI bridge that supports virtualization |
| CN101387993A (en) * | 2007-09-14 | 2009-03-18 | 凹凸科技（中国）有限公司 | Method and system for dynamically collocating resource for equipment in computer system |
| US20100007668A1 (en) * | 2008-07-08 | 2010-01-14 | Casparian Mark A | Systems and methods for providing scalable parallel graphics rendering capability for information handling systems |
| US20100036995A1 (en) * | 2008-08-05 | 2010-02-11 | Hitachi, Ltd. | Computer system and bus assignment method |
| US20110157193A1 (en) * | 2009-12-29 | 2011-06-30 | Nvidia Corporation | Load balancing in a system with multi-graphics processors and multi-display systems |
| CN101916209A (en) * | 2010-08-06 | 2010-12-15 | 华东交通大学 | Cluster task resource allocation method for multi-core processor |
| US20130091500A1 (en) * | 2011-10-10 | 2013-04-11 | William J. Earl | Paravirtualized virtual gpu |
| CN103105895A (en) * | 2011-11-15 | 2013-05-15 | 辉达公司 | Computer system and display cards thereof and method for processing graphs of computer system |
| CN102609978A (en) * | 2012-01-13 | 2012-07-25 | 中国人民解放军信息工程大学 | Method for accelerating cone-beam CT (computerized tomography) image reconstruction by using GPU (graphics processing unit) based on CUDA (compute unified device architecture) architecture |
| CN103248659A (en) * | 2012-02-13 | 2013-08-14 | 北京华胜天成科技股份有限公司 | Method and system for dispatching cloud computed resources |
| CN102902589A (en) * | 2012-08-31 | 2013-01-30 | 浪潮电子信息产业股份有限公司 | Method for managing and scheduling cluster MIS (Many Integrated Core) job |
| CN104954400A (en) * | 2014-03-27 | 2015-09-30 | 中国电信股份有限公司 | Cloud computing system and realizing method thereof |
| CN105718316A (en) * | 2014-12-01 | 2016-06-29 | 中国移动通信集团公司 | Job scheduling method and apparatus |
| CN106557366A (en) * | 2015-09-28 | 2017-04-05 | 阿里巴巴集团控股有限公司 | Task distribution method, apparatus and system |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109144578A (en) * | 2018-06-28 | 2019-01-04 | 中国船舶重工集团公司第七0九研究所 | A kind of video card resource allocation method and device based on Godson computer |
| CN109144578B (en) * | 2018-06-28 | 2021-09-03 | 中国船舶重工集团公司第七0九研究所 | Display card resource allocation method and device based on Loongson computer |
| CN115129483A (en) * | 2022-09-01 | 2022-09-30 | 武汉凌久微电子有限公司 | Multi-display-card cooperative display method based on display area division |
| CN115129483B (en) * | 2022-09-01 | 2022-12-02 | 武汉凌久微电子有限公司 | Multi-display-card cooperative display method based on display area division |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107193650B (en) | 2021-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106648664A (en) | Self-service terminal business process updating method and apparatus | |
| CN104239133B (en) | A kind of log processing method, device and server | |
| CN108829469A (en) | A kind of application program page methods of exhibiting and device | |
| US8839197B2 (en) | Automated analysis of composite applications | |
| CN103995908A (en) | Method and device for importing data | |
| CN106357738A (en) | Automatic arrangement method for server cluster and device as well as system thereof | |
| CN112306661B (en) | Task scheduling method and device, electronic equipment and readable storage medium | |
| CN109685301A (en) | Method for managing resource, device, equipment and readable storage medium storing program for executing | |
| CN108900434A (en) | Data collection distribution method and device | |
| CN111984426A (en) | Task scheduling method and device, electronic equipment and storage medium | |
| CN107341235A (en) | Report form inquiring method, device, medium and electronic equipment | |
| CN107193650A (en) | A kind of method and apparatus that video card resource is dispatched in distributed type assemblies | |
| CN106855862B (en) | Rapid comparison method and device | |
| CN106648823A (en) | Code release method and apparatus | |
| CN107864187A (en) | The online task executing method of terminal device and server | |
| CN112541688B (en) | Service data verification method and device, electronic equipment and computer storage medium | |
| CN109657240A (en) | Determine the method, apparatus, equipment and medium of fault type | |
| CN104572921B (en) | A kind of method of data synchronization and device across data center | |
| CN104199689B (en) | The installation method and device of comprehensive front-end system | |
| CN111612387A (en) | Flow direction distribution method, device, equipment and storage medium | |
| CN108205470A (en) | A kind of distribution ad data calculating task management system and method | |
| CN110865817B (en) | Cloud computing platform resource metering method and system | |
| Nakao et al. | Using agile software development methods to support human-centered design | |
| CN106372071B (en) | The information acquisition method and device of data warehouse | |
| CN111383037B (en) | Method and device for constructing advertisement materials |
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 |