CN102207851B - Method for calling immediate data in dynamically reconfigurable processor - Google Patents

Method for calling immediate data in dynamically reconfigurable processor Download PDF

Info

Publication number
CN102207851B
CN102207851B CN 201110140358 CN201110140358A CN102207851B CN 102207851 B CN102207851 B CN 102207851B CN 201110140358 CN201110140358 CN 201110140358 CN 201110140358 A CN201110140358 A CN 201110140358A CN 102207851 B CN102207851 B CN 102207851B
Authority
CN
China
Prior art keywords
immediately
subtask
subelement
optional
task
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.)
Expired - Fee Related
Application number
CN 201110140358
Other languages
Chinese (zh)
Other versions
CN102207851A (en
Inventor
王延升
刘雷波
朱敏
戚斌
杨军
曹鹏
时龙兴
尹首一
魏少军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuxi Research Institute of Applied Technologies of Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to CN 201110140358 priority Critical patent/CN102207851B/en
Publication of CN102207851A publication Critical patent/CN102207851A/en
Application granted granted Critical
Publication of CN102207851B publication Critical patent/CN102207851B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Logic Circuits (AREA)

Abstract

The invention provides a method for calling immediate data in a dynamically reconfigurable processor. The method comprises the following steps: popping up a sub-unit configuration word by a sub-unit configuration word buffer of the dynamically reconfigurable processor; loading a sub-task sequence in the sub-unit configuration word into a local sub-task sequence memory; respectively writing required immediate data and optional immediate data of the current task in the sub-unit configuration word into a required immediate data register and an optional immediate data register (1); partitioning the required immediate data and the optional immediate data, and reducing the immediate data transmission operation in each local sub-task only by changing the optional immediate data, thereby greatly improving the work efficiency.

Description

Call the method for counting immediately in a kind of dynamic reconfigurable processor
Technical field
The present invention relates to the dynamic reconfigurable processor technical field in the embedded system field, particularly relate to and call the method for counting immediately in a kind of dynamic reconfigurable processor.
Background technology
The dynamic reconfigurable processor is a kind of new life's processor framework, and its single core processor, special chip, field programmable logic array (FPLA) as compared with the past has significant advantage, is a direction of following circuit structure development.
At first, often contain a plurality of arithmetic logic unit in the dynamic reconfigurable processor, and enormous amount, be referred to as many nuclear arrays.Array inside is equipped with the high routing unit of flexibility ratio, realizes between the arithmetic logic unit diversified interconnected.Therefore, the crowd after the route unit connects examines array can realize high speed processing to data stream, and more traditional monokaryon and few core processor have huge advantage in performance.Simultaneously, the special circuit that solidifies also has huge advantage in dirigibility.
Secondly, more traditional static reconfigurable circuit---field programmable logic array (FPLA), the dynamic reconfigurable processor has dynamic characteristics, the i.e. function of commutation circuit dynamically in the circuit operational process, but not the heavy constant circuit function that do not change of static reconfigurable circuit one in the past, just programming circuit function before the circuit operation is carried out initialization to circuit.The benefit of doing like this is to have reduced by time-multiplexed mode the scale of circuit, and the full mapping of the circuit structure before reason is becomes the piecemeal mapping now, and has taked the mode of dynamic switching between piece and the piece just.
In the practical application, the demand of counting immediately of calling is often arranged in the dynamic reconfigurable processor, therefore, the present invention proposes to call in a kind of dynamic reconfigurable processor the method for counting immediately with innovating, to satisfy the demand of practical application.
Summary of the invention
Technical matters to be solved by this invention provides and calls the method for counting immediately in a kind of dynamic reconfigurable processor, realizes calling of counting immediately, greatly the work efficiency that improves.
In order to address the above problem, the invention discloses and call the method for counting immediately in a kind of dynamic reconfigurable processor, described method comprises:
A, dynamic reconfigurable processor subelement configuration words buffer eject the subelement configuration words;
Subtask sequence in B, the subelement configuration words is loaded into local subtask sequence memory;
Essential number immediately and the optional number immediately of the current task in the subelement configuration words write essential number register immediately and optional number register immediately 1 respectively;
Next task in C, the beginning subtasking sequence, the processing unit configuration information configuration process unit in the information of subtask;
Optional selection information of counting immediately in the information of the subtask optional number immediately that current subtask is used selects and writes optional number register immediately 2 from optional number register immediately 1;
The number immediately that D, processing unit call in essential number register immediately and the optional number register immediately 2 is carried out current subtask;
Whether also have the subtask not carry out in E, the judgement subtask sequence:
If then return C and carry out next subtask;
If not, then return A and eject next subelement configuration words.
Preferably, be cached with a plurality of subelement configuration words in the described subelement configuration words buffer, the task that the corresponding subelement of subelement configuration words will be carried out.
Preferably, the subtask sequence that comprises the task that subelement will carry out in the described subelement configuration words, required counting immediately of calling in the task that subtask task of sequence formation and subelement will be carried out, described number immediately comprise essential number immediately and optional several immediately.
Preferably, the performed task of subelement is divided into a plurality of subtasks and carries out at processing unit several times.
Preferably, described subtask information comprises the used optional selection information of counting immediately of processing unit configuration information and current subtask of current subtask.
Preferably, described number immediately needs the uncertain number that uses in calculating process for the processing unit in the subelement of dynamic reconfigurable processor.
Preferably, described essential number immediately is that reach is counting immediately of current subtask sequence, and optional number immediately is that reach is subtask several immediately in the sequence of subtask.
Preferably, the direct processed cell call of described essential number register immediately.
Preferably, it is optional several immediately to have many groups in the described optional number register immediately 1, and the size of optional number register immediately 1 depends on one group of optional number of counting immediately and the optional group number of counting immediately, and the size of register is more than or equal to this " one group of number * group number ".
Preferably, the size of described optional number register immediately 2 depends on one group of optional number of counting immediately, and the size of register is greater than this number.
Compared with prior art, the present invention has the following advantages:
Dynamic reconfigurable processor subelement configuration words buffer ejects the subelement configuration words among the present invention, subtask sequence in the subelement configuration words is loaded into local subtask sequence memory, essential number immediately and the optional number immediately of the current task in the subelement configuration words write essential number register immediately and optional number register immediately 1 respectively, divide the operation of number transmission immediately in essential number immediately and the each local subtask of minimizing, optional number immediately back, each optional number immediately of change that only needs gets final product, greatly the work efficiency that improves.
Description of drawings
Fig. 1 is the synoptic diagram of a kind of typical dynamic reconfigurable processor described in the specific embodiment of the invention;
Fig. 2 calls the method flow diagram of counting immediately in the described a kind of dynamic reconfigurable processor of the embodiment of the invention;
Fig. 3 is that the embodiment of the invention is described a kind of for the device synoptic diagram of realizing that use is counted immediately in the dynamic reconfigurable processor.
Embodiment
For above-mentioned purpose of the present invention, feature and advantage can be become apparent more, the present invention is further detailed explanation below in conjunction with the drawings and specific embodiments.
A kind of structural representation of typical dynamic reconfigurable processor as shown in Figure 1, the running flow process of dynamic reconfigurable processor can be summarized as following steps:
1. the dynamic reconfigurable processor is read in data pending in the processor external storage in the processor by the external data reader unit, is distributed to a plurality of subelements and handles interior external data buffer.
2. many subelements are read data from external data buffer separately, handle again.A plurality of subelements are handled simultaneously, to reach the effect of parallel processing.After each subelement is handled, result data is stored in separately internal data buffer.
3. external data is write out device result data is read from the internal data buffer of each subelement, and is written out to the processor external memory storage.
Except above-mentioned steps, also it should be noted that: the synchronization mechanism between subelement is realized by isochronous controller between subelement; Data interaction between subelement is by writing data by subelement A to data interaction working storage between subelement, and subelement B reads these data from data interaction working storage between subelement and realizes.Also need the synchronization mechanism between the zygote unit during use, namely subelement A is to after the data interaction working storage has been write data between subelement, and subelement B just can begin to read data interaction working storage between subelement.
The dynamic reconfigurable processor can go to be described from the angle of data stream and configuration flow.The angle of data stream refers to from the angle to the processing of data stream, ergodic data stream each module in the dynamic reconfigurable processor of process, the required function of each module is proposed.From the angle of data stream, the dynamic reconfigurable processor structurally can be divided into following four parts:
1. data shifting apparatus
A) data shifting apparatus is used for data are read from data storage device, and writes another data storage device, and data are shifted between data storage device.The final realization writes pe array with data, after pe array has moved, the result data of pe array read.
B) data shifting apparatus is divided into 6: the external data reader unit, and data interaction device between subelement, the inside and outside DTU (Data Transfer unit), the internal data reader unit, internal data is write out device, and external data is write out device.Flow direction by data stream specifies as follows.
C) external data reader unit:
I. be used for the outer pending data of dynamic reconfigurable processor are read in the dynamic reconfigurable processor, be cached in the external data buffer.
D) data interaction device between subelement:
I. read for the result data of other subelements that will store in the exchanges data working storage between subelement, and integrate to a certain extent, data after integrating the most at last are distributed to the internal data memory of current subelement or directly are distributed to pe array, and this depends on whether whether the result data of unit can direct processed cell array processing will be repeatedly processed with these data in addition.
E) inside and outside DTU (Data Transfer unit):
I. be used for the external data of buffer memory in the external data buffer is distributed to internal data memory or directly is distributed to pe array, this depends on whether whether these outside data can direct processed cell array processing will be repeatedly processed with these data.
F) internal data reader unit
I. be used for the internal data of the current subelement that will store in the internal data memory, and integrate to a certain extent, the data after integrating the most at last are written to pe array.
G) internal data is write out device
I. be used for the result data of pe array is read, writing to pe array again uses when and then moving pe array, or write to internal data memory with the data buffer memory, or write between subelement the data interaction working storage and use for data interaction between subelement, or write the internal data buffer and wait to export to outside the processor.
H) external data is write out device:
I. the result data for the subelement that will be cached in the internal data buffer exports to outside the processor.
2. data storage device
A) be divided into 4, specify as follows.
B) external data buffer:
I. be used for the external data buffer memory that reads in, be read out when to be needed.
C) internal data memory
I. be used for the result data of pe array is kept in, be read out when to be needed, also for using in processor inside.
D) internal data buffer
I. be used for the result data of pe array is carried out buffer memory, wait to export to outside the processor.
E) data interaction working storage between subelement
I. be used for the result data of temporary subelement A, treat that subelement B reads, to realize the data interaction between subelement.
3. arithmetic element
A) arithmetic element is pe array, specifically sees Fig. 1 for details.
B) processing unit is an arithmetic logical unit, can realize basic arithmetical operation function and the arithmetic logical operation function of logical operation function and some customizations.
C) connect by routing unit between the processing unit, to realize the flexible and fast data transmission between the processing unit.
D) the entire process cell array is controlled by timing control unit, with the sequential relationship between the input of Coordination Treatment cell array, computing, the output.
E) pe array can realize data in enormous quantities are carried out the fast processing of cycling.
4. sync control device
A) sync control device is isochronous controller between subelement.
B) it realizes that a plurality of subelements (1 to N, N is the subelement number) all finish a plurality of subelements behind its appointed task separately (what herein a plurality of can be with before is a plurality of inconsistent) and just can carry out its next task in task sequence separately.
C) isochronous controller has been realized certain comprehensive task after being divided into a plurality of tasks between subelement, is distributed to a plurality of subelement executed in parallel, after the intact a plurality of tasks of a plurality of subelement executed in parallel synchronously.The parallelization that just can begin next comprehensive task is synchronously carried out.
Embodiment:
With reference to Fig. 2, show and call the method flow diagram of counting immediately in a kind of dynamic reconfigurable processor of the present invention, described method comprises:
Step S201, dynamic reconfigurable processor subelement configuration words buffer ejects the subelement configuration words;
Step S202, the subtask sequence in the subelement configuration words is loaded into local subtask sequence memory;
Essential number immediately and the optional number immediately of the current task in the subelement configuration words write essential number register immediately and optional number register immediately 1 respectively;
Step S203, the next task in the beginning subtasking sequence, the processing unit configuration information configuration process unit in the information of subtask;
Optional selection information of counting immediately in the information of the subtask optional number immediately that current subtask is used selects and writes optional number register immediately 2 from optional number register immediately 1;
The number immediately that step S204, processing unit call in essential number register immediately and the optional number register immediately 2 is carried out current subtask;
Step S205, judge whether also have the subtask not carry out in the sequence of subtask:
If then return step S203 and carry out next subtask;
If not, then return step S201 and eject next subelement configuration words.
With reference to Fig. 3, show a kind of for the device synoptic diagram of realizing that use is counted immediately in the dynamic reconfigurable processor, in described device:
1. subelement configuration words buffer
A) be cached with a plurality of subelement configuration words in the subelement configuration words buffer, the task that the corresponding subelement of subelement configuration words will be carried out, the subelement configuration words is written in the subelement configuration words buffer and buffer memory;
B) comprise two parts content in the subelement configuration words:
I. the subtask sequence of the subelement task of will carrying out, the subtask sequence constitutes a task;
Ii. required counting immediately of calling in the task that subelement will be carried out, i.e. required counting immediately of calling in the sequence of subtask, this comprises essential number immediately and optional several immediately;
C) whenever a task begins to carry out, eject a configuration words in the subelement configuration words buffer:
I. the subtask sequence in the configuration words is loaded into local subtask sequence memory;
Ii. essential number immediately and optional number immediately are loaded into essential number register immediately and optional number register immediately 1 respectively;
2. local subtask sequence memory
A) during the subelement of dynamic reconfigurable processor was executed the task process, the performed task of subelement was divided into a plurality of subtasks and carries out at processing unit several times, and the subtask sequence is stored in the sequence memory of local subtask;
B) comprise two parts content in the information of subtask:
I. the processing unit configuration information of current subtask;
Ii. the used optional selection information of counting immediately in current subtask;
C) when a subtask begins to carry out:
I. the processing unit configuration information configuration process unit in the information of subtask;
Ii. the optional number immediately that current subtask is used of the optional selection information of counting immediately in the information of subtask selects and writes optional number register immediately 2 from optional number register immediately 1.
3. number register immediately
A) number needs the uncertain number that uses in calculating process for the processing unit in the subelement of dynamic reconfigurable processor immediately.
B) number is stored in the subelement configuration words immediately, is written into subelement configuration words buffer with the subelement configuration words and enters in the subelement, and when subelement configuration words buffer ejected configuration words, the information of counting immediately in the configuration words sent to number register immediately immediately;
C) number is divided into optional number immediately and essential several immediately immediately:
I. essential several immediately:
1. essential number immediately is that reach is counting immediately of current subtask sequence.
2. after being ejected, current configuration words is written into essential number register immediately immediately, after next configuration words ejects, by essential several covering the immediately of next task.
3. the direct processed cell call of essential number register immediately;
Ii. optional several immediately:
1. optional number immediately is that reach is the counting immediately of subtask in the sequence of subtask.
2. ejected the many groups of optional numbers of using each subtask, back immediately at current configuration words and be written into optional number register immediately 1 immediately.
3. in the process that current subtask sequence is carried out, whenever a subtask begins to carry out, optional selection information of counting immediately one group of optional number immediately that current subtask is used selects and writes optional number register immediately 2 from optional number register immediately 1.
4. will count immediately to be subdivided into and optionally count immediately, and from optional number register immediately 1, select one group of optional counting to immediately in the optional number register immediately 2, with making this step that is chosen in of the optional number of array is immediately finished, rather than go to finish at processing unit, this will reduce the memory space of the extra processing unit configuration information that brings thus.
D) number register immediately:
1. the size of essential number register immediately depends on the essential number of counting immediately that need use in the sequence of subtask, and the size of register needs more than or equal to this number.
2. it is optional several immediately to have many groups in the optional number register immediately 1, and the size of optional number register immediately 1 depends on one group of optional number of counting immediately and the optional group number of counting immediately, and the size of register needs more than or equal to this " one group of number * group number ".
3. the size of optional number register immediately 2 depends on one group of optional number of counting immediately, and the size of register needs greater than this number.
4. processing unit:
A) processing unit calls in every step task of subtask sequence and is loaded into essential number register immediately and the optional number register immediately 2 of number immediately.
B) processing unit is the data path in the subelement among Fig. 1, this comprises that inside and outside DTU (Data Transfer unit), internal data reader unit, internal data write out device, external data buffer, internal data memory, internal data buffer and pe array, and in fact using the device of counting immediately in the processing unit is pe array.
Number is example to use immediately in the dynamic reconfigurable processor below, illustrates to use the method for counting immediately in this dynamic reconfigurable processor.
For convenience of explanation, at first carry out some settings for parameter, specific as follows:
1. stored configuration word number is 1 in the subelement configuration words buffer, is labeled as configuration words 1 respectively, corresponding subelement task 1.The task of configuration words 1 corresponding subelement is divided into 2 subtasks.
2. in the configuration words 1, the information of counting immediately of storage is that required essential number immediately is 4 in the subelement task 1, is respectively 1011,1101,1111,1010, and the required optional number average immediately in each subtask is 1, is followed successively by 0101,0001.Count information such as following table immediately:
Task 1
Essential several immediately 1011 1101 1111 1010
Subtask 1 Subtask 2
Optional several immediately 0101 0001
3. it is as follows specifically to count call method immediately
A) dynamic reconfigurable processor subelement configuration words buffer ejects the subelement configuration words, and subelement configuration words 1 ejects;
B) the subtask sequence in the subelement configuration words 1 is loaded into the subtask sequence memory, and namely subtask 1 and subtask 2 are loaded in the sequence memory of subtask;
C) the essential of the current task in the subelement configuration words 1 several 1011,1101,1111,1011 writes essential number register immediately immediately, and be optional several 0101 immediately,, 0001 writes optional number register immediately 1;
D) sequence buffer in subtask ejects the subtask 1 of task 1, the processing unit configuration information configuration process unit in 1 information of subtask;
E) the optional number immediately that current subtask is used of the optional selection information of counting immediately in 1 information of subtask selects and writes optional number register immediately 2 from optional number register immediately 1, namely selects 0101 and write optional number register immediately 2 from optional number register immediately 1;
F) processing unit calls the several subtaskings 1 immediately in essential number register immediately and the optional number register immediately 2;
G) after subtask 1 was finished, it was true that pe array sends signal to subtask sequence buffer, and subtask sequence buffer is judged subtask 2 ejections;
H) sequence buffer in subtask ejects the subtask 2 of task 1, the processing unit configuration information configuration process unit in 2 information of subtask;
I) the optional number immediately that current subtask is used of the optional selection information of counting immediately in 2 information of subtask selects and writes optional number register immediately 2 from optional number register immediately 1, namely selects 0001 and write optional number register immediately 2 from optional number register immediately 1;
J) processing unit calls the several subtaskings 2 immediately in essential number register immediately and the optional number register immediately 2;
K) after subtask 2 was finished, it was true that pe array sends signal to subtask sequence buffer, and no subtask in the sequence of subtask waits for that subelement distributes new subtask sequence.
More than be described in detail calling immediately the method for counting in a kind of dynamic reconfigurable processor provided by the present invention, used specific case herein principle of the present invention and embodiment are set forth, the explanation of above embodiment just is used for helping to understand method of the present invention and core concept thereof; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims (9)

1. call the method for counting immediately in a dynamic reconfigurable processor, it is characterized in that described method comprises:
A, dynamic reconfigurable processor subelement configuration words buffer eject the subelement configuration words;
Subtask sequence in B, the subelement configuration words is loaded into local subtask sequence memory;
Essential number immediately and the optional number immediately of the current task in the subelement configuration words write essential number register immediately and optional number register immediately 1 respectively; Wherein, described essential number immediately is that reach is counting immediately of current subtask sequence, and optional number immediately is that reach is subtask several immediately in the sequence of subtask;
Next task in C, the beginning subtasking sequence, the processing unit configuration information configuration process unit in the information of subtask;
Optional selection information of counting immediately in the information of the subtask optional number immediately that current subtask is used selects and writes optional number register immediately 2 from optional number register immediately 1;
The number immediately that D, processing unit call in essential number register immediately and the optional number register immediately 2 is carried out current subtask;
Whether also have the subtask not carry out in E, the judgement subtask sequence:
If then return C and carry out next subtask;
If not, then return A and eject next subelement configuration words.
2. the method for claim 1 is characterized in that:
Be cached with a plurality of subelement configuration words in the described subelement configuration words buffer, the task that the corresponding subelement of subelement configuration words will be carried out.
3. the method for claim 1 is characterized in that:
Comprise two parts in the described subelement configuration words, the subtask sequence of the task that a part will be carried out for subelement, wherein, task of subtask sequence formation; Another part is required counting immediately of calling in the subelement task of will carrying out, and described number immediately comprises essential number immediately and optionally counts immediately.
4. the method for claim 1 is characterized in that:
The performed task of subelement is divided into a plurality of subtasks and carries out at processing unit several times.
5. the method for claim 1 is characterized in that:
Described subtask information comprises the used optional selection information of counting immediately of processing unit configuration information and current subtask of current subtask.
6. the method for claim 1 is characterized in that:
Described number immediately needs the uncertain number that uses for the processing unit in the subelement of dynamic reconfigurable processor in calculating process, wherein, described uncertain numerical table is shown in the number that dynamically produces in the described calculating process, be not fixed and use.
7. the method for claim 1 is characterized in that:
The direct processed cell call of described essential number register immediately.
8. the method for claim 1 is characterized in that:
It is optional several immediately to have many groups in the described optional number register immediately 1, and the size of optional number register immediately 1 depends on one group of optional number of counting immediately and the optional group number of counting immediately, and the size of register is more than or equal to this " one group of number * group number ".
9. the method for claim 1 is characterized in that:
The size of described optional number register immediately 2 depends on one group of optional number of counting immediately, and the size of register is greater than this number.
CN 201110140358 2011-05-27 2011-05-27 Method for calling immediate data in dynamically reconfigurable processor Expired - Fee Related CN102207851B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110140358 CN102207851B (en) 2011-05-27 2011-05-27 Method for calling immediate data in dynamically reconfigurable processor

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110140358 CN102207851B (en) 2011-05-27 2011-05-27 Method for calling immediate data in dynamically reconfigurable processor

Publications (2)

Publication Number Publication Date
CN102207851A CN102207851A (en) 2011-10-05
CN102207851B true CN102207851B (en) 2013-07-17

Family

ID=44696702

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110140358 Expired - Fee Related CN102207851B (en) 2011-05-27 2011-05-27 Method for calling immediate data in dynamically reconfigurable processor

Country Status (1)

Country Link
CN (1) CN102207851B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103942181B (en) * 2014-03-31 2017-06-06 清华大学 Method, device for generating the configuration information of dynamic reconfigurable processor

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1713133A (en) * 2004-06-25 2005-12-28 富士通株式会社 Reconfigurable processor and semiconductor device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101636377B1 (en) * 2009-10-23 2016-07-06 삼성전자주식회사 Configuration processor, configuration control apparatus and method, and Thread modeling method

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1713133A (en) * 2004-06-25 2005-12-28 富士通株式会社 Reconfigurable processor and semiconductor device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Register File Architecture Optimization in a Coarse-Grained Reconfigurable Architecture;Zion Kwok等;《13th Annual IEEE Symposium on Field-Programmable Custom Computing Machines》;20050420;35-44 *
Zion Kwok等.Register File Architecture Optimization in a Coarse-Grained Reconfigurable Architecture.《13th Annual IEEE Symposium on Field-Programmable Custom Computing Machines》.2005,35-44.

Also Published As

Publication number Publication date
CN102207851A (en) 2011-10-05

Similar Documents

Publication Publication Date Title
CN114168525B (en) Reconfigurable parallel processing
CN104899182B (en) A kind of Matrix Multiplication accelerated method for supporting variable partitioned blocks
CN100535850C (en) Registers for data transfers within a multithreaded processor
CN110222818B (en) Multi-bank row-column interleaving read-write method for convolutional neural network data storage
US11803508B2 (en) Systems and methods for implementing a machine perception and dense algorithm integrated circuit and enabling a flowing propagation of data within the integrated circuit
CN110738308A (en) neural network accelerators
US20200265546A1 (en) Task execution in a simd processing unit with parallel groups of processing lanes
CN102541809A (en) Dynamic reconfigurable processor
CN111324294B (en) Method and device for accessing tensor data
CN112487750A (en) Convolution acceleration computing system and method based on memory computing
CN102306141B (en) Method for describing configuration information of dynamic reconfigurable array
CN101093474A (en) Method for implementing matrix transpose by using vector processor, and processing system
EP3844610B1 (en) Method and system for performing parallel computation
CN102279729B (en) Method, buffer and processor for dynamic reconfigurable array to schedule configuration information
CN110991619A (en) Neural network processor, chip and electronic equipment
CN111047036A (en) Neural network processor, chip and electronic equipment
WO2022047403A1 (en) Memory processing unit architectures and configurations
CN102207851B (en) Method for calling immediate data in dynamically reconfigurable processor
CN102236632B (en) Method for hierarchically describing configuration information of dynamic reconfigurable processor
CN102184090B (en) Dynamic re reconfigurable processor and fixed number calling method thereof
CN104572519A (en) Multiport access and storage controller for multiprocessor and control method thereof
CN102207850B (en) Method for hierarchically executing configuration streams in dynamically reconfigurable processor
CN111047035A (en) Neural network processor, chip and electronic equipment
CN109685209B (en) Device and method for accelerating operation speed of neural network
CN101165805A (en) Multiple port memory access control module

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
ASS Succession or assignment of patent right

Owner name: WUXI APPLICATION TECHNOLOGY RESEARCH INSTITUTE OF

Free format text: FORMER OWNER: TSINGHUA UNIVERSITY

Effective date: 20141222

C41 Transfer of patent application or patent right or utility model
COR Change of bibliographic data

Free format text: CORRECT: ADDRESS; FROM: 100084 HAIDIAN, BEIJING TO: 214072 WUXI, JIANGSU PROVINCE

TR01 Transfer of patent right

Effective date of registration: 20141222

Address after: 214072, A3 building, No. 777 West Building Road, Binhu District, Jiangsu, Wuxi 4, China

Patentee after: Wuxi Research Institute of Applied Technologies Tsinghua University

Address before: 100084 Tsinghua Yuan, Beijing, Haidian District

Patentee before: Tsinghua University

CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20130717