CN101661406A - Processing unit dispatching device and method - Google Patents
Processing unit dispatching device and method Download PDFInfo
- Publication number
- CN101661406A CN101661406A CN200810214487A CN200810214487A CN101661406A CN 101661406 A CN101661406 A CN 101661406A CN 200810214487 A CN200810214487 A CN 200810214487A CN 200810214487 A CN200810214487 A CN 200810214487A CN 101661406 A CN101661406 A CN 101661406A
- Authority
- CN
- China
- Prior art keywords
- processing unit
- work
- control module
- request
- subtask
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Debugging And Monitoring (AREA)
Abstract
The invention provides processing unit dispatching device and method. In a system of processing units, the number of the processing units is dynamically regulated by monitoring the occupying conditionof shared data exchange resources. According to the provided technical scheme, the invention can avoid starting excessive processing units, thereby avoiding that the performance cannot be enhanced under the condition of consuming extra energy. Therefore, the energy consumption of the system is reduced on the premise of ensuring the performance of the system.
Description
Technical field
The present invention relates to computer technology, in particular, relate to the apparatus and method that processing unit is dispatched.
Background technology
In many systems in modern times, a calculation task is broken down into a plurality of subtasks, each subtask is represented that by a thread processing unit in one group of processing unit that each thread system that is assigned to comprises is finished, the result of the described calculation task of the comprehensive one-tenth of the result of each subtask.This method can shorten the deadline of task, thereby improves the performance of total system.
The processing power of supposing each processing unit all is the same, it is generally acknowledged that employed processing unit is many more, the calculated amount of the subtask that each processing unit is assigned to is just few more, the time of finishing the subtask calculated amount is just short more, sees that on the whole the time of finishing described calculation task is just short more.But the processing unit of use is many more, must bring bigger energy consumption.Therefore, need a kind of scheduling mechanism, improving system performance and reducing to reach between the energy consumption balance.
Summary of the invention
In view of this, the invention provides processing unit dispatching device and dispatching method, so that improving system performance and reducing to reach between the energy consumption balance.
According to technical scheme of the present invention, obtain the number of pending request of access to the data exchange resource, and according to the number of the processing unit of described request of access number adjustment work.
In the system of multiplied unit, the total system bottleneck of performance is the speed of exchanges data between the each processing unit, and does not lie in the computing velocity of single processing unit, does not also lie in the number of processing unit.When described request of access number is big, starts the overall performance that more processing unit can't correspondingly improve system, even may worsen overall performance.Therefore, be the number of the processing unit of standard adjustment work with the number of described request of access, can avoid under the situation that consumes additional energy, but can't obtaining corresponding performance boost, thereby under the prerequisite that guarantees system performance, reduce the energy consumption of system.
Description of drawings
Fig. 1 is the processing speed that records of experiment and the corresponding relation of processing unit number.
Fig. 2 is the energy consumption that records of experiment and the corresponding relation of processing unit number.
Fig. 3 is the block diagram according to the processing unit dispatching device of the embodiment of the invention.
Fig. 4 is the process flow diagram according to the processing unit dispatching method of the embodiment of the invention.
Embodiment
Following processing unit dispatching device and the dispatching method that provides with reference to the description of drawings embodiment of the invention.
Fig. 1 is the processing speed that records of experiment and the corresponding relation of processing unit number.
In this experiment, the calculation task that allows the processing unit group finish is to calculate fast fourier transform.Counting of fast fourier transform shown in the leftmost tabulation of Fig. 1.It will be understood by those skilled in the art that under the identical situation of other parameters it is many more to count, the calculated amount of calculating this fast fourier transform is just big more.
A uppermost line display is used to finish the number of the processing unit of this calculation task among Fig. 1.Processing unit here promptly can be a processing unit such as CPU (central processing unit), also can be a nuclear in the multinuclear processing unit.Needed deadline when other expressions use the processing unit of certain number to calculate specific fast fourier transform of counting among Fig. 1.Therefore need to prove that the unit of this time is a self-defining time of day in the system, should only be intended to illustrate the relativeness of required time under the various different situations time, required time of the big more expression of numeral is long more.
As can be seen from Figure 1, at the fast fourier transform point that calculates at 64, the processing unit number is that 2 o'clock used time is the shortest; The number that increases processing unit does not only shorten the deadline, has increased the deadline on the contrary.Calculating 128 and during 256 fast fourier transform, the processing unit number is that 4 o'clock used time is the shortest; The number that increases processing unit does not only shorten the deadline, has increased the deadline on the contrary.When the fast fourier transform of calculating at 512, the processing unit number is that 4 o'clock used time is the shortest; The number that increases processing unit does not only shorten the deadline, has increased the deadline on the contrary.Calculating 1024,2048 and during 4096 fast fourier transform, the processing unit number is that 8 o'clock used time is the shortest.Can obtain conclusion by Fig. 1, be not that employed processing unit number is many more, and it is just few more to finish the required time of whole calculation task.
Fig. 2 is the energy consumption that records of experiment and the corresponding relation of processing unit number.
Identical among the meaning of top line and the left side one row and Fig. 1 among Fig. 2.Other expressions use the processing unit of certain number to calculate the specific energy that fast fourier transform consumed of counting among Fig. 2.Need to prove that the unit of this energy value is a self-defining energy unit in the system, therefore only be intended to illustrate the relativeness of institute's energy requirement under the various different situations, the energy that the big more explanation of numeral consumes is many more.As can be seen from Figure 2, the calculation task that no matter will finish is a fast fourier transform of calculating how many points, and always can draw to draw a conclusion: employed processing unit number is many more, and it is just high more to finish the energy that this calculation task consumes.
Therefore,, use more processing unit, must mean to consume more energy for a specific calculation task.In some cases, use more processing unit, though can cause the deadline to shorten, the degree that shortens is compared and may be lost more than gain with the degree that energy consumption increases.Under even worse situation, use more processing unit can increase the deadline on the contrary.
The present inventor has carried out deep research and experiment to above problem, and a reason that causes above-mentioned phenomenon is the exchanges data between the processing unit.The subtask is being assigned to each computing unit and when each computing unit obtains the result of calculation of each subtask, is all needing the exchanges data between the processing unit.Each processing unit relies on external memory storage or bus to come swap data, in the present specification external memory storage or bus all is called the exchanges data resource.The exchanges data resource is generally all shared, then must there be certain mechanism to regulate the visit of each processing unit to shared exchanges data resource, allow some processing units elder generation visit data exchange resource, visit again the exchanges data resource after some processing unit waits for a period of time.The processing unit number is many more, and the average latency is just big more.When the average latency when finishing required time of subtask, the deadline of whole task is just depended on the deadline of average latency rather than subtask, even increase the deadline that the number of processing unit also can't shorten whole task this moment.
Fig. 3 is the block diagram according to the processing unit dispatching device of the embodiment of the invention.
Monitor module is configured to be connected to the exchanges data explorer, therefrom obtains the number of pending request of access to the data exchange resource, and judges whether that according to this number instruction control module adjusts the number of processing unit.
Control module is configured to the number according to the instruction adjustment processing unit of monitor module.
As mentioned above, the exchanges data resource is shared, need certain mechanism and regulate the visit that each processing unit is carried out the exchanges data resource of sharing, the device that will carry out this visit adjusting in the present specification is called the exchanges data explorer, and it is used for managing the request of access of each processing unit to the data exchange resource in the system of multiplied unit.Those skilled in the art can easily realize several data exchange resource manager.When two above processing units are all wanted the visit data exchange resource, come the request of access of these two above processing units is managed by the exchanges data explorer.The simplest Managed Solution is that the request of access of sending earlier obtains earlier carrying out, and that is to say, sends the processing unit elder generation visit data exchange resource of request of access earlier.At this moment, described request of access is placed in the request of access formation of exchanges data explorer inside according to the time order and function order.Queue heads is the request that the next one will be carried out; Rear of queue is last sent request of access.The length of formation can be represented the occupied situation of exchanges data resource.Certainly, also can set up formation respectively, read request queue and write request queue and be referred to as the request of access formation at the request of reading and the request that writes.
The request of access that current time sends if the length of formation is long, illustrates that a processing unit just can be performed after need be waited for the long duration.If the length of formation is increased to certain-length, then to calculate the required time of next subtask than this processing unit long the stand-by period, even this moment, this processing unit began to calculate next subtask at current time, it is among computation process or calculate also can't the visit data exchange resource after finishing.Such as this processing unit can't be from exchanges data resource reading of data among computation process, and perhaps this processing unit can't be to exchanges data resource output result after calculating is finished.Just can stop the work of this processing unit this time.When the length of formation is reduced to a certain degree, the work that can recover this processing unit.Certainly, the exchanges data explorer can manage request of access by other modes.No matter adopt which kind of mode, as long as the number of pending request of access such as can embody, then monitor module just can judge whether that instruction control module adjusts the number of processing unit according to this number.
From the angle of monitor module, if the number of finding the request of access waiting for greater than first threshold, then instruction control module reduces the number of processing unit.Described first threshold can be predefined according to experimental result before, also can dynamically be provided with.As mentioned above, when average deadline of subtask during, increase the processing unit number and just can not improve system performance in proportion less than average latency of visit data exchange resource; And the number of described average latency and waiting request of access has certain corresponding relation under the identical situation of other conditions, and in general the big more then average latency of the number of waiting request of access is just long more.Therefore, be increased to particular value from the request of access number and begin just to begin by measuring the average deadline to obtain the subtask, thereby this particular value is made as first threshold less than the average latency.Certainly, relation between the number that also can be by measuring waiting request of access and the bandwidth usage of exchanges data resource, obtain described bandwidth usage and be increased to particular value from the request of access number and begin just, thereby this particular value is made as first threshold near saturated.One of ordinary skill in the art will readily recognize that the present invention is not limited to any specific method that described first threshold is set.
It is exactly the work of at least one processing unit in the processing unit that stops to work that described control module reduces the processing unit number.Can directly stop the work of described at least one processing unit by means such as clock controls.But, if still the processing unit in work ceaselessly sends request of access, may cause the length of described request of access formation always can not be reduced to and make the processing unit that is stopped restart the degree of work, thereby cause the always intact subtask that does not become it to be responsible for calculating of processing unit that is stopped.
In order to prevent this situation, can when stopping processing unit work, the subtask that does not begin as yet in this processing unit to calculate be transferred to the processing unit that is not stopped, and stop follow-up to the newly-generated subtask of its distribution.It is known for those skilled in the art that processing unit is distributed in the subtask, thereby it also is to realize that easily the present invention is not limited to any specific transfer method for those skilled in the art that another processing unit is transferred to from a processing unit in the subtask.After transfer was finished, this processing unit continued to calculate current subtask of calculating.The processing unit that has possesses automatic out-of-work ability when not carrying out, and after finishing current subtask of calculating, these processing units just quit work automatically so.If processing unit does not possess this ability, can after finishing current subtask of calculating, processing unit stop the work of this processing unit by means such as clock controls.
If described monitor module is found the number of the described request of access of waiting for less than second threshold value, then instruction control module increases the number of processing unit.Described second threshold value can be identical identical with first threshold, so that reduce the complexity of described monitor module.Described second threshold value also can be other values less than first threshold, changes continually with the number that prevents processing unit.Promptly the difference between second threshold value and the first threshold can be according to the balance between complexity and the change frequency is provided with.Described monitor module can also from notify number that described control module reduces processing unit for the last time above certain hour situation at interval under, instruct described control module to increase the number of processing unit again.Also can by described control module from stop some processing unit work above certain hour situation at interval under, start the work of these processing units once more.The described time interval can be predefined, also can be to set according to the number of the request of access of waiting for.
Described control module is when increasing the processing unit number, also can be by distributing the subtask to start the processing unit that this is not worked to the processing unit of not working, the subtask of being distributed can be not begin the subtask of calculating in other processing units of working as yet, also can be newly-generated subtask.If processing unit does not possess the function of starting working automatically when being assigned with the subtask, then before distributing the subtask, start the processing unit that this is not worked by clock control earlier to the processing unit of not working.Certainly, processing units all in the possible system is all in running order, and this moment, control module can't increase the processing unit of work.Control module can be given monitoring modular with this advisory.
Fig. 4 is the process flow diagram according to the processing unit dispatching method of the embodiment of the invention.
As shown in Figure 4, in step 401, obtain the number of pending request of access to the data exchange resource.
Can obtain relevant information from the exchanges data explorer.The exchanges data explorer may have been safeguarded the request of access formation, and the request of access formation also can and write request queue by read request queue and form.The processing unit dispatching method that the embodiment of the invention provided does not also rely on the exchanges data explorer and how the request of access of each processing unit is managed, as long as can obtain wait pending request of access number from the exchanges data explorer.
As previously mentioned, reducing the number of the processing unit of work, can be the work that stops the processing unit of working more than simply by means such as clock controls.Thereby always can't the work of restarting to finish the task that it is responsible for calculating for fear of the processing unit that is stopped, the processing unit that other work on can be transferred in the subtask that does not begin as yet in the processing unit to be stopped to calculate, and stop follow-up newly-generated subtask to be assigned to this processing unit to be stopped.Processing unit described to be stopped to quit work after having calculated current subtask of calculating automatically; Perhaps allow it quit work by means such as clock controls.
Whether step 404 judges described request of access number less than second threshold value, if less than this second threshold value, then enter step 406, otherwise enters step 405.Described second threshold value can equal first threshold, also can be other values less than first threshold.
In above-mentioned steps 404 and the step 405 any one all can be omitted.When omitting step 404, be not more than under the situation of described first threshold at described request of access number, enter step 405.When omitting step 405, be not less than under the situation of second threshold value at described request of access number, return step 401.
As mentioned above, the number of the processing unit of increase work, can be by distributing the subtask to the processing unit of not working, the subtask of being distributed can be that a part is shifted to the new processing unit that starts in other processing units of having worked subtask of being responsible for calculating, and new subtask to be allocated such as also can be.If processing unit does not possess the function of starting working automatically when receiving the subtask, then earlier allow the processing unit of not working start working, and then distribute the subtask to this processing unit by means such as clock controls.If only be to stop the work of processing unit and do not have subsequent treatment in step 403, the processing unit that at first is stopped in the setting up procedure 403 then.Certainly, if all processing units are all worked in the system, then can not increase the processing unit number of work.
Though illustrated and described some exemplary embodiments of the present invention, but it will be understood by those skilled in the art that, under the prerequisite that does not depart from principle of the present invention and spirit, can change these embodiments, scope of the present invention is limited by claims and equivalence transformation thereof.
Claims (17)
1, a kind of processing unit dispatching device comprises:
Monitor module is configured to be connected to the exchanges data explorer, therefrom obtains the number of pending request of access to the data exchange resource, and judges whether the number of the processing unit of instruction control module adjustment work according to this number; With
Control module is configured to the number according to the processing unit of the instruction adjustment work of monitor module.
2, device according to claim 1, wherein said monitor module are configured under the situation of described request number greater than first threshold, instruct the number of processing unit of described control module minimizing work;
Described control module is configured to the described instruction according to monitor module, stops the processing unit of at least one work.
3, device according to claim 2, wherein said control module is configured to, and stops the processing unit of described at least one work by clock control.
4, according to claim 2 or 3 described devices, wherein said control module is configured to when stopping the processing unit of described at least one work, and the processing unit that works on is transferred in the subtask that this processing unit does not begin to calculate.
5, device according to claim 2, wherein said monitor module is configured to, from instructing the number of processing unit of described control module minimizing work to surpass under the situation in a time interval, instruct the number of processing unit of described control module increase work;
Described control module is configured to the described instruction according to monitor module, starts the processing unit that at least one is not worked.
6, device according to claim 2, wherein said control module is configured to, and from instructing the number of processing unit of described control module minimizing work to surpass under the situation in a time interval, starts the processing unit that at least one is not worked.
7, device according to claim 2, wherein said monitor module are configured under the situation of described request number less than second threshold value, instruct the number of processing unit of described control module increase work;
Described control module is configured to the described instruction according to monitor module, starts the processing unit that at least one is not worked.
8, according to any described device in the claim 5 to 7, wherein said control module is configured to, by distributing the subtask to start the processing unit that this is not worked to described at least one processing unit of not working.
9, according to any described device in the claim 5 to 7, wherein said control module is configured to, and starts described at least one processing unit of not working by clock control.
10, a kind of processing unit dispatching method comprises:
Obtain the number of pending request of access to the data exchange resource; With
Number according to the processing unit of described request of access number adjustment work.
11, method according to claim 10, the number of wherein said processing unit according to described request of access number adjustment work comprises:
Under the situation of described request of access number, stop the processing unit of at least one work greater than first threshold.
12, method according to claim 10, the wherein said step that stops the processing unit of at least one work comprises:
Stop the processing unit of described at least one work by clock control.
13, according to claim 11 or 12 described methods, the wherein said step that stops the processing unit of at least one work further comprises:
The processing unit that works on is transferred in the subtask that the processing unit of described at least one work does not begin to calculate.
14, method according to claim 11 further comprises: surpass under the situation in a time interval at the number from the processing unit that reduces work, increase the number of the processing unit of work.
15, method according to claim 11 further comprises: under the situation of described request number less than second threshold value, increase the number of the processing unit of work.
16, according to claim 14 or 15 described methods, the number of the processing unit of wherein said increase work comprises: distribute the subtask to start the processing unit that this is not worked by the processing unit of not working at least one, thereby increase the processing unit of work.
17, according to claim 14 or 15 described methods, the number of the processing unit of wherein said increase work comprises: start the processing unit that at least one is not worked by clock control, thereby increase the processing unit of work.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810214487A CN101661406A (en) | 2008-08-28 | 2008-08-28 | Processing unit dispatching device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810214487A CN101661406A (en) | 2008-08-28 | 2008-08-28 | Processing unit dispatching device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101661406A true CN101661406A (en) | 2010-03-03 |
Family
ID=41789452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200810214487A Pending CN101661406A (en) | 2008-08-28 | 2008-08-28 | Processing unit dispatching device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101661406A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102546946A (en) * | 2012-01-05 | 2012-07-04 | 中国联合网络通信集团有限公司 | Method and device for processing task on mobile terminal |
| CN103870340A (en) * | 2014-03-06 | 2014-06-18 | 华为技术有限公司 | Data processing method and control node in stream computation system and stream computation system |
| CN108805275A (en) * | 2017-06-16 | 2018-11-13 | 上海兆芯集成电路有限公司 | Programmable device and its operating method and computer usable medium |
| CN110574019A (en) * | 2017-04-26 | 2019-12-13 | 瑞典爱立信有限公司 | Robust data replication between data centers |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101076770A (en) * | 2004-09-28 | 2007-11-21 | 英特尔公司 | Method and apparatus for varying energy per instruction according to the amount of available parallelism |
| CN101159005A (en) * | 2006-07-31 | 2008-04-09 | Sap股份公司 | Method and system to adaptively manage the quality of service of interactions between smart item networks and enterprise applications |
-
2008
- 2008-08-28 CN CN200810214487A patent/CN101661406A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101076770A (en) * | 2004-09-28 | 2007-11-21 | 英特尔公司 | Method and apparatus for varying energy per instruction according to the amount of available parallelism |
| CN101159005A (en) * | 2006-07-31 | 2008-04-09 | Sap股份公司 | Method and system to adaptively manage the quality of service of interactions between smart item networks and enterprise applications |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102546946A (en) * | 2012-01-05 | 2012-07-04 | 中国联合网络通信集团有限公司 | Method and device for processing task on mobile terminal |
| CN103870340A (en) * | 2014-03-06 | 2014-06-18 | 华为技术有限公司 | Data processing method and control node in stream computation system and stream computation system |
| WO2015131721A1 (en) * | 2014-03-06 | 2015-09-11 | 华为技术有限公司 | Data processing method in stream computing system, control node and stream computing system |
| CN103870340B (en) * | 2014-03-06 | 2017-11-07 | 华为技术有限公司 | Data processing method, control node and stream calculation system in stream calculation system |
| US10097595B2 (en) | 2014-03-06 | 2018-10-09 | Huawei Technologies Co., Ltd. | Data processing method in stream computing system, control node, and stream computing system |
| US10630737B2 (en) | 2014-03-06 | 2020-04-21 | Huawei Technologies Co., Ltd. | Data processing method in stream computing system, control node, and stream computing system |
| CN110574019A (en) * | 2017-04-26 | 2019-12-13 | 瑞典爱立信有限公司 | Robust data replication between data centers |
| CN110574019B (en) * | 2017-04-26 | 2023-05-30 | 瑞典爱立信有限公司 | Robust data replication between data centers |
| CN108805275A (en) * | 2017-06-16 | 2018-11-13 | 上海兆芯集成电路有限公司 | Programmable device and its operating method and computer usable medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108984282B (en) | Scheduler for AMP architecture with closed-loop performance controller | |
| US10613876B2 (en) | Methods and apparatuses for controlling thread contention | |
| US9201816B2 (en) | Data processing apparatus and a method for setting priority levels for transactions | |
| EP1971924B1 (en) | Methods and system for interrupt distribution in a multiprocessor system | |
| CN104991830B (en) | YARN resource allocations and energy-saving scheduling method and system based on service-level agreement | |
| US9372526B2 (en) | Managing a power state of a processor | |
| US8639862B2 (en) | System-on-chip queue status power management | |
| JP6799947B2 (en) | Systems and methods for scheduling tasks using the sliding time window | |
| US20190042331A1 (en) | Power aware load balancing using a hardware queue manager | |
| US20100332869A1 (en) | Method and apparatus for performing energy-efficient network packet processing in a multi processor core system | |
| US20110161978A1 (en) | Job allocation method and apparatus for a multi-core system | |
| US20120297216A1 (en) | Dynamically selecting active polling or timed waits | |
| JP2003298599A (en) | Method and apparatus for distribution control | |
| US9244733B2 (en) | Apparatus and method for scheduling kernel execution order | |
| CN103460189A (en) | Techniques for managing power consumption state of a processor | |
| JP2008129846A (en) | Data processor, data processing method, and program | |
| CN106557369A (en) | A kind of management method and system of multithreading | |
| US20090178045A1 (en) | Scheduling Memory Usage Of A Workload | |
| US20150378782A1 (en) | Scheduling of tasks on idle processors without context switching | |
| WO2021012510A1 (en) | Adaptive adjustment method and device for cpu use ratio, terminal and storage medium | |
| CN101661406A (en) | Processing unit dispatching device and method | |
| CN113505084A (en) | Memory resource dynamic regulation and control method and system based on memory access and performance modeling | |
| US11907550B2 (en) | Method for dynamically assigning memory bandwidth | |
| CN115391031A (en) | Load balancing scheduling method and computing equipment | |
| US11704058B2 (en) | Systems and methods for resource-based scheduling of commands |
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: 20100303 |