WO2010061911A1 - 再構成可能回路生成装置、方法およびプログラム - Google Patents
再構成可能回路生成装置、方法およびプログラム Download PDFInfo
- Publication number
- WO2010061911A1 WO2010061911A1 PCT/JP2009/070009 JP2009070009W WO2010061911A1 WO 2010061911 A1 WO2010061911 A1 WO 2010061911A1 JP 2009070009 W JP2009070009 W JP 2009070009W WO 2010061911 A1 WO2010061911 A1 WO 2010061911A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- netlist
- shared
- reconfigurable circuit
- netlists
- groups
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/34—Circuit design for reconfigurable circuits, e.g. field programmable gate arrays [FPGA] or programmable logic devices [PLD]
Definitions
- Reconfigurable circuits such as FPGA (Field Programmable Gate Array) and PLD (Programmable Logic Device) are widely used because they can realize a wide variety of circuits.
- FPGA Field Programmable Gate Array
- PLD Programmable Logic Device
- switches and wirings are used to realize high flexibility, and there is a problem that the reconfigurable circuit has a large area and a high cost.
- a reconfigurable circuit is limited in its application range, and a reconfigurable circuit having sufficient flexibility to cover the application range (hereinafter, “restricted reconfigurable circuit”). Can be used.
- the limited use reconfigurable circuit has an advantage that the area can be greatly reduced although the flexibility is lower than that of a general-purpose reconfigurable circuit.
- Patent Documents 1 and 2 describe a method for generating a limited use reconfigurable circuit.
- FIG. 12 schematically shows a flowchart of a method for generating a limited use reconfigurable circuit described in Patent Documents 1 and 2.
- an initial architecture of a reconfigurable circuit is created (step S21).
- a placement and routing tool corresponding to the initial architecture is generated (step S22).
- the netlist for the application to be mounted on the reconfigurable circuit is placed and routed on the initial architecture using the placement and routing tool (step S23).
- the netlist placement and routing result is evaluated (step S24). As a typical evaluation, for example, it is evaluated whether or not a netlist has been placed and routed on the architecture.
- step S23 When multiple applications are mounted on the reconfigurable circuit, the placement and routing (step S23) and the evaluation (step S24) are repeated for the netlist for each application.
- step S25 When the evaluation of the netlist for all applications is completed (YES in step S25), it is determined whether or not to continue the architecture search (step S26). For example, the architecture search may be continued when all the applications can be arranged and routed, and the architecture search may be terminated otherwise.
- the architecture search is continued (YES in step S26)
- the architecture of the reconfigurable circuit is changed (step S27), and the steps S22 to S25 are performed again.
- the architecture change for example, the number of programmable wirings of the reconfigurable circuit may be changed, or the number of programmable switches may be changed.
- An object of the present invention is to provide a reconfigurable circuit generation device, method, and program for solving such problems.
- a reconfigurable circuit generation device includes a netlist generation unit that generates a netlist that can be shared among a plurality of netlists having a common portion as a shared netlist, and the shared A resource reduction unit that reduces resources of the reconfigurable circuit on which the plurality of netlists are to be mounted within a range in which the netlist can be mounted.
- a program according to a third aspect of the present invention implements a netlist generation process for generating a netlist that can be shared among a plurality of netlists having a common part as a shared netlist, and the shared netlist
- the computer is caused to execute resource reduction processing for reducing resources of the reconfigurable circuit on which the plurality of netlists are to be mounted to the extent possible.
- the reconfigurable circuit generating apparatus, method, and program according to the present invention can generate a limited-use reconfigurable circuit with a small area in a short time.
- FIG. 1 is a block diagram illustrating a configuration of a reconfigurable circuit generation device according to a first embodiment of the present invention. It is a flowchart of the reconfigurable circuit generation method which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the computer which performs the process in the program which concerns on the 3rd Embodiment of this invention. It is a block diagram which shows the structure of a typical reconfigurable circuit. It is a circuit diagram which shows the typical structure of the cell which comprises a reconfigurable circuit. It is a circuit diagram which shows the typical structure of the multiplexer which comprises a cell. It is a circuit diagram which shows the typical structure of the switch which comprises a multiplexer.
- the reconfigurable circuit generation device 60 includes a netlist generation unit 61 and a resource reduction unit 62.
- the net list generation unit 61 generates a net list that can be shared among a plurality of net lists having a common part as a shared net list.
- the resource reduction unit 62 reduces resources of a reconfigurable circuit that should mount a plurality of netlists within a range in which a shared netlist can be mounted.
- the resources of the reconfigurable circuit may include a function block and a routing resource for connecting the function blocks. At this time, it is preferable that the resource reduction part 62 reduces a routing resource.
- the netlist generation unit 61 further includes a classification unit that classifies a plurality of netlists into a plurality of groups, and a netlist that can be shared between a plurality of netlists included in each of the plurality of groups.
- a net list that can be shared between a first generation unit that generates a shared net list for each of a plurality of groups and a first shared net list for each of the plurality of groups is generated as a second shared net list. It is preferable to have the 2nd production
- the plurality of netlists and the shared netlist may each include a calculation block and a net connecting the calculation blocks.
- the shared net list is a net shared as much as possible among a plurality of net lists.
- the resources of the reconfigurable circuit may include functional blocks and routing resources that connect the functional blocks. At this time, it is preferable to reduce routing resources in the resource reduction step (step S32).
- the net list generation step (step S31) further includes a step of classifying the plurality of net lists into a plurality of groups, and a shared net list that can be shared between the plurality of net lists included in each of the plurality of groups. And generating for each of the plurality of groups.
- the resource reduction step (step S32) it is preferable to reduce the resources of the reconfigurable circuit within a range in which all the shared netlists for each of the plurality of groups can be mounted.
- the netlist generation step (step S31) further includes a step of classifying the plurality of netlists into a plurality of groups, and extracting a common part between the plurality of netlists included in each of the plurality of groups. Generating a first shared netlist for each of the plurality of groups, and a second netlist that can be shared between the first shared netlist for each of the plurality of groups. And generating a shared netlist.
- the resource reduction step (step S32) it is preferable to reduce resources of the reconfigurable circuit within a range in which the second shared netlist can be mounted.
- FIG. 3 is a block diagram illustrating a configuration of a computer that executes processing in the program according to the present embodiment.
- the computer 70 includes a CPU 71, a storage unit 72, an input device 73, and an output device 74.
- a netlist that can be shared among a plurality of netlists is provided so that a plurality of netlists can overlap each other as much as possible. It is preferable to synthesize. In addition, it is preferable that this shared net list is a target of placement and routing.
- a small area reconfigurable circuit with a limited application range can be generated in a short time. This is because a shared netlist that can be shared among a plurality of netlists is created in advance, and the resources of the reconfigurable circuit are reduced as long as the shared netlist can be placed and routed. Each of the plurality of netlists can be generated by removing a part of the shared netlist. Therefore, when the shared netlist can be arranged and wired for the reconfigurable circuit, any of a plurality of netlists sharing the shared netlist can be arranged and wired.
- FIG. 4 is a block diagram illustrating a configuration of a typical reconfigurable circuit.
- the reconfigurable circuit includes cells 1AA to 1CC.
- FIG. 4 shows a cell array of 3 rows and 3 columns as an example. However, the size of the array is arbitrary. Adjacent cells are connected by a horizontal wiring 30 and a vertical wiring 31.
- the horizontal wiring 30 in FIG. 4 connects the cells adjacent on the left and right, and the vertical wiring 31 connects the cells adjacent on the top and bottom.
- FIG. 4 shows a case where the number of wirings connecting adjacent cells is two as an example. However, the number of wirings is arbitrary.
- FIG. 5 is a circuit diagram showing a typical configuration of cells constituting the reconfigurable circuit.
- the cell 1 includes wirings 31A and 31B that transmit signals in the vertical direction, wirings 30A and 30B that transmit signals in the horizontal direction, and a functional block 50 that can set various functions according to configuration data.
- Multiplexers 20E, 20F that select input signals to the functional block 50 from the wiring groups 30A, 30B, 31A, 31B according to the configuration data, and between the wirings of the wiring groups 30A, 30B, 31A, 31B according to the configuration data
- multiplexers 20A, 20B, 20C, and 20D that connect the outputs of the functional block 50 and the wirings are included.
- FIG. 6 is a circuit diagram showing a typical configuration of the multiplexer 20 constituting the cell 1 of FIG.
- the plurality of input signal lines 7 of the multiplexer 20 are connected to one ends of the switches 2A to 2D, respectively.
- the other ends of the switches 2A to 2D are connected to the input of the buffer 8.
- the output of the buffer 8 is output to the output signal line 9.
- FIG. 7 is a circuit diagram showing a typical configuration of the switch 2 constituting the multiplexer 20 of FIG. Whether to connect the terminals 80 and 81 on both sides of the transmission gate 6 is determined by the configuration data recorded in the configuration memory 5.
- the above configuration of the reconfigurable circuit is only an example.
- the number of inputs of the multiplexer is arbitrary. Multiplexers with different numbers of inputs may be mounted together. Furthermore, many configurations other than the above are known as the configuration of the multiplexer, and any multiplexer may be used. Further, the number of input functional blocks, the number of wires, and the length of wires are not limited to the above-described example.
- step S11 common portions of a plurality of netlists are overlapped as much as possible to generate a shared netlist that can be shared among the plurality of netlists.
- step S12 an architecture of a reconfigurable circuit as a starting point, that is, an initial architecture is generated (step S12).
- the initial architecture is a reconfigurable circuit with abundant routing resources such as wiring and switches, and can implement an arbitrary netlist.
- a place and route tool that takes into account the configuration of the initial architecture is generated (step S13).
- the shared netlist generated in step S11 is placed and routed on the initial architecture (step S14).
- the placement and routing result of the shared net list is evaluated (step S15). The evaluation may be performed based on, for example, whether or not placement and routing have been completed. In addition, the evaluation is performed with the maximum delay and power consumption of the application circuit when the placement and routing is completed, the number of switches and the circuit area of the reconfigurable circuit when the placement and routing is completed, or a combination thereof (for example, circuit area and delay). May be performed based on the product). Which evaluation method is used is determined according to the purpose of the designer.
- step S16 it is determined whether or not to continue the architecture search for the reconfigurable circuit based on the evaluation result. Whether or not to continue searching the architecture is determined according to the conditions determined by the designer. For example, the search for architecture may be terminated (NO in step S16) when placement and routing cannot be performed or when improvement in delay is not observed.
- step S16 the architecture of the reconfigurable circuit is changed (step S17), and then the steps after step S13 are performed using the changed architecture.
- the architecture change for example, a part of the switches (2A to 2D in FIG. 6) of the multiplexer 20 in FIG. 5 may be deleted, or a part of the wiring connecting the cells in FIG. 4 may be deleted. .
- FIG. 9 is a diagram for explaining a method of generating a shared netlist by superimposing a plurality of netlists.
- FIG. 9A shows the first netlist 10.
- each operation block 10_1i (i is an integer) of the netlist 10 is assigned to a function block 50 (FIG. 5) of one cell 1 of the reconfigurable circuit.
- symbols A to E attached to the respective operation blocks 10_1i indicate the types of operations.
- the net 10_2i (i is an integer) corresponds to a wiring connecting operation blocks of the net list 10 or an input / output signal line of the net list 10.
- FIG. 9B is an example of the second netlist 11.
- each of the operation blocks 11_1i (i is an integer) of the netlist 11 is assigned to the function block 50 (FIG. 5) of one cell 1 of the reconfigurable circuit.
- symbols A to E attached to the respective calculation blocks 11_1i indicate the types of calculation.
- the net 11_2i (i is an integer) corresponds to a wiring connecting the operation blocks of the net list 11 or an input / output signal line of the net list 11.
- FIG. 9C shows an example of the shared net list 12 in which the first net list 10 and the second net list 11 are overlapped.
- each of the operation blocks 12_1i (i is an integer) of the shared netlist 12 is assigned to the function block 50 (FIG. 5) of one cell 1 of the reconfigurable circuit.
- the net 12_2i (i is an integer) corresponds to a wiring connecting the operation blocks of the shared netlist 12 or an input / output signal line of the shared netlist 12.
- the symbol X attached to the operation block 12_1i means an arbitrary operation that can be realized in the function block 50.
- the function of the operation block of the netlist may be any function as long as the function block of the reconfigurable circuit can be realized.
- the function of the calculation block is independent of whether or not it can be placed and routed.
- the shared netlist 12 includes a first netlist 10 and a second netlist 11.
- the first net list 10 is equivalent.
- the calculation block 12_13 and the nets 12_23 and 12_25 are deleted from the shared net list 12, and the function of each calculation block is appropriately set, the result is equivalent to the net list 11.
- both the first netlist 10 and the second netlist 11 can be arranged and wired to the reconfigurable circuit.
- the network structure formed by them is the same in both netlists.
- the shared netlist 12 has a structure in which a common portion having the same network structure in a plurality of netlists is used as it is and a difference between the plurality of netlists and the common portion is added to the common portion. By deleting a portion corresponding to the difference from the shared net list 12, one of the plurality of net lists is obtained.
- FIG. 10C shows a state where the shared netlist 12 of FIG. 9C is arranged and wired in the reconfigurable circuit of FIG.
- a thick arrow corresponds to the net 12_2i (i is an integer).
- the arrows inside the cells 1AB, 1CC, and 1CA indicate that the function of transmitting signals from the cell input (thick line) to the output (thick line) is mapped.
- a cell to which a symbol is attached is a cell in which the functional block is used (any operation is performed).
- a cell that is not given a symbol inside the cell is a cell whose functional block is not used.
- An arrow indicated by a thin line represents an unused wiring.
- the netlist 10 in FIG. 9A is obtained by appropriately setting the function of the arithmetic block by excluding the arithmetic block 12_15 and the net 12_28 from the shared netlist 12 in FIG. 9C.
- the result of the placement and routing of the netlist 10 shown in FIG. 10A is that the functional block of the cell 1CA and the net 31CA are unused in FIG. 10C, and other X symbols are written. It is obtained by appropriately setting the function of the functional block of the cell.
- the present invention by using the shared netlist for the placement and routing, there is no variation in the common part, so that the number of necessary resources can be reduced and a reconfigurable circuit having a smaller area than the conventional one can be generated. it can.
- a matching method that maximizes the scale may be searched.
- a plurality of netlists may be classified into a plurality of groups, and a common netlist may be generated for each group. In this case, as many common netlists as the number of groups are generated. In this case, when reducing resources in the reconfigurable circuit, it is necessary to evaluate whether or not all the common netlists can be arranged and routed. However, compared to the case where all netlists are placed and routed individually, the number of times of placement and routing can be reduced, so the time required to generate a reconfigurable circuit with reduced resources can be reduced. .
- a net list that can be shared between these shared net lists may be further generated.
- a shared net list is generated at once from a large number of net lists, it may take a long time for matching between net lists.
- the netlist is divided into a plurality of groups, a shared netlist for the netlist belonging to each group is generated, and then a netlist that can be shared among these shared netlists is generated. It may be.
- the embodiments and examples can be changed and adjusted based on the basic technical concept. Various combinations and selections of various disclosed elements are possible within the scope of the claims of the present invention. That is, the present invention of course includes various variations and modifications that could be made by those skilled in the art according to the entire disclosure including the claims and the technical idea.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Design And Manufacture Of Integrated Circuits (AREA)
- Logic Circuits (AREA)
Abstract
Description
本発明は、日本国特許出願:特願2008-304164号(2008年11月28日出願)の優先権主張に基づくものであり、同出願の全記載内容は引用をもって本書に組み込み記載されているものとする。
本発明は、再構成可能回路を生成する再構成可能回路生成装置、方法およびプログラムに関し、特に、適用範囲を限定した再構成可能回路を生成する再構成可能回路生成装置、方法およびプログラムに関する。
図12は、特許文献1および2に記載された用途限定再構成可能回路の生成方法のフローチャートを概略的に示す。
本発明の実施形態に係る再構成可能回路生成装置について、図面を参照して説明する。図1は本実施形態に係る再構成可能回路生成装置の構成を示すブロック図である。
本発明の実施形態に係る再構成可能回路生成方法について、図面を参照して説明する。図2は本実施形態に係る再構成可能回路生成方法のフローチャートである。
本発明の第3の実施形態に係る再構成可能回路を生成するためのプログラムについて図面を参照して説明する。図3は、本実施形態に係るプログラムにおける処理を実行させるコンピュータの構成を示すブロック図である。図3を参照すると、コンピュータ70は、CPU71、記憶部72、入力装置73および出力装置74を有する。
本発明の全開示(請求の範囲を含む)の枠内において、さらにその基本的技術思想に基づいて、実施形態ないし実施例の変更・調整が可能である。また、本発明の請求の範囲の枠内において種々の開示要素の多様な組み合わせないし選択が可能である。すなわち、本発明は、請求の範囲を含む全開示、技術的思想にしたがって当業者であればなし得るであろう各種変形、修正を含むことは勿論である。
2、2A~2D スイッチ
5 コンフィギュレーションメモリ
6 トランスミッションゲート
7 入力信号線
8 バッファ
9 出力信号線
10、11 ネットリスト
12 共用ネットリスト
10_10~10_14、11_10~11_12、11_14、11_15、12_10~12_15 演算ブロック
10_20~10_27、11_20~11_22、11_24、11_26~11_28、12_20~12_28 ネット
20、20A~20F マルチプレクサ
30、30A、30B 横方向配線
31、31A、31B、31CA 縦方向配線
50 機能ブロック
60 再構成可能回路生成装置
61 ネットリスト生成部
62 リソース削減部
70 コンピュータ
71 CPU
72 記憶部
73 入力装置
74 出力装置
80、81 端子
Claims (16)
- 共通部分を有する複数のネットリストの間で共用することができるネットリストを共用ネットリストとして生成するネットリスト生成部と、
前記共用ネットリストを実装することができる範囲で前記複数のネットリストを実装すべき再構成可能回路のリソースを削減するリソース削減部とを備えていることを特徴とする再構成可能回路生成装置。 - 前記複数のネットリスト及び前記共用ネットリストは、それぞれ演算ブロックと該演算ブロック間を接続するネットとを含み、
前記共用ネットリストは、前記複数のネットリストの間で可能な限りネットを共用化したものであることを特徴とする、請求項1に記載の再構成可能回路生成装置。 - 前記再構成可能回路のリソースは、機能ブロックと該機能ブロック間を接続するルーティングリソースとを含み、
前記リソース削減部は、前記ルーティングリソースを削減することを特徴とする、請求項1又は2に記載の再構成可能回路生成装置。 - 前記共用ネットリストは、前記複数のネットリストが包含する最小のネットリストであることを特徴とする、請求項1乃至3のいずれか1項に記載の再構成可能回路生成装置。
- 前記ネットリスト生成部は、さらに、複数のネットリストを複数のグループに分類する分類部と、
前記複数のグループのそれぞれに含まれる複数のネットリストの間で共用することができるネットリストを共用ネットリストとして前記複数のグループのそれぞれについて生成する生成部とを備え、
前記リソース削減部は、前記複数のグループのそれぞれに対する共用ネットリストをいずれも実装することができる範囲で前記再構成可能回路のリソースを削減することを特徴とする、請求項1乃至4のいずれか1項に記載の再構成可能回路生成装置。 - 前記ネットリスト生成部は、さらに、複数のネットリストを複数のグループに分類する分類部と、
前記複数のグループのそれぞれに含まれる複数のネットリストに共用することができるネットリストを第1の共用ネットリストとして前記複数のグループのそれぞれについて生成する第1の生成部と、
前記複数のグループのそれぞれに対する第1の共用ネットリストの間で共用することができるネットリストを第2の共用ネットリストとして生成する第2の生成部とを備え、
前記リソース削減部は、前記第2の共用ネットリストを実装することができる範囲で前記再構成可能回路のリソースを削減することを特徴とする、請求項1乃至4のいずれか1項に記載の再構成可能回路生成装置。 - 共通部分を有する複数のネットリストの間で共用することができるネットリストを共用ネットリストとして生成するネットリスト生成工程と、
前記共用ネットリストを実装することができる範囲で前記複数のネットリストを実装すべき再構成可能回路のリソースを削減するリソース削減工程とを含むことを特徴とする再構成可能回路生成方法。 - 前記複数のネットリスト及び前記共用ネットリストは、それぞれ演算ブロックと該演算ブロック間を接続するネットとを含み、
前記共用ネットリストは、前記複数のネットリストの間で可能な限りネットを共用化したものであることを特徴とする、請求項7に記載の再構成可能回路生成方法。 - 前記再構成可能回路のリソースは、機能ブロックと該機能ブロック間を接続するルーティングリソースとを含み、
前記リソース削減工程において、前記ルーティングリソースを削減することを特徴とする、請求項7又は8に記載の再構成可能回路生成方法。 - 前記ネットリスト生成工程は、さらに、複数のネットリストを複数のグループに分類する工程と、
前記複数のグループのそれぞれに含まれる複数のネットリストの間で共用することができる共用ネットリストを前記複数のグループのそれぞれについて生成する工程とを含み、
前記リソース削減工程において、前記複数のグループのそれぞれに対する共用ネットリストをいずれも実装することができる範囲で前記再構成可能回路のリソースを削減することを特徴とする、請求項7乃至9のいずれか1項に記載の再構成可能回路生成方法。 - 前記ネットリスト生成工程は、さらに、複数のネットリストを複数のグループに分類する工程と、
前記複数のグループのそれぞれに含まれる複数のネットリストの間で共通部分を抽出することにより、該複数のネットリストを包含する第1の共用ネットリストを前記複数のグループのそれぞれについて生成する工程と、
前記複数のグループのそれぞれに対する第1の共用ネットリストの間で共用することができるネットリストを第2の共用ネットリストとして生成する工程とを含み、
前記リソース削減工程において、前記第2の共用ネットリストを実装することができる範囲で前記再構成可能回路のリソースを削減することを特徴とする、請求項7乃至9のいずれか1項に記載の再構成可能回路生成方法。 - 共通部分を有する複数のネットリストの間で共用することができるネットリストを共用ネットリストとして生成するネットリスト生成処理と、
前記共用ネットリストを実装することができる範囲で前記複数のネットリストを実装すべき再構成可能回路のリソースを削減するリソース削減処理とをコンピュータに実行させることを特徴とするプログラム。 - 前記複数のネットリスト及び前記共用ネットリストは、それぞれ演算ブロックと該演算ブロック間を接続するネットとを含み、
前記共用ネットリストは、前記複数のネットリストの間で可能な限りネットを共用化したものであることを特徴とする、請求項12に記載のプログラム。 - 前記再構成可能回路のリソースは、機能ブロックと該機能ブロック間を接続するルーティングリソースとを含み、
前記リソース削減処理において、前記ルーティングリソースを削減する処理をコンピュータに実行させることを特徴とする、請求項12又は13に記載のプログラム。 - 前記ネットリスト生成処理において、さらに、複数のネットリストを複数のグループに分類する処理と、
前記複数のグループのそれぞれに含まれる複数のネットリストの間で共用することができる共用ネットリストを前記複数のグループのそれぞれについて生成する処理とをコンピュータに実行させるとともに、
前記リソース削減処理において、前記複数のグループのそれぞれに対する共用ネットリストをいずれも実装することができる範囲で前記再構成可能回路のリソースを削減する処理をコンピュータに実行させることを特徴とする、請求項12乃至14のいずれか1項に記載のプログラム。 - 前記ネットリスト生成処理において、さらに、複数のネットリストを複数のグループに分類する処理と、
前記複数のグループのそれぞれに含まれる複数のネットリストの間で共用することができるネットリストを第1の共用ネットリストとして前記複数のグループのそれぞれについて生成する処理と、
前記複数のグループのそれぞれに対する第1の共用ネットリストの間で共用することができるネットリストを第2の共用ネットリストとして生成する処理とをコンピュータに実行させるとともに、
前記リソース削減処理において、前記第2の共用ネットリストを実装することができる範囲で前記再構成可能回路のリソースを削減する処理をコンピュータに実行させることを特徴とする、請求項12乃至14のいずれか1項に記載のプログラム。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/130,215 US8694948B2 (en) | 2008-11-28 | 2009-11-27 | Reconfigurable circuit generation device, method, and program |
JP2010540520A JP5402940B2 (ja) | 2008-11-28 | 2009-11-27 | 再構成可能回路生成装置、方法およびプログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008304164 | 2008-11-28 | ||
JP2008-304164 | 2008-11-28 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010061911A1 true WO2010061911A1 (ja) | 2010-06-03 |
Family
ID=42225781
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2009/070009 WO2010061911A1 (ja) | 2008-11-28 | 2009-11-27 | 再構成可能回路生成装置、方法およびプログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US8694948B2 (ja) |
JP (1) | JP5402940B2 (ja) |
WO (1) | WO2010061911A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2018526714A (ja) * | 2016-01-12 | 2018-09-13 | ▲騰▼▲訊▼科技(深▲セン▼)有限公司 | Cnn処理方法およびデバイス |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014238617A (ja) * | 2013-06-05 | 2014-12-18 | 富士通株式会社 | データ処理装置、データ処理方法及び情報処理装置 |
EP3336727A1 (en) * | 2016-12-19 | 2018-06-20 | Menta | System and method for defining a programmable logic architecture |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7051313B1 (en) * | 1999-10-29 | 2006-05-23 | Altera Toronto Co. | Automatic generation of programmable logic device architectures |
JP2006172219A (ja) * | 2004-12-16 | 2006-06-29 | Toshiba Corp | 半導体集積回路の自動設計装置、半導体集積回路の自動設計方法、及び半導体集積回路の自動設計プログラム |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH1041475A (ja) | 1996-07-23 | 1998-02-13 | Mitsubishi Electric Corp | 半導体集積回路装置及びその製造方法 |
US6473726B1 (en) * | 1999-09-24 | 2002-10-29 | Frederic Reblewski | Method and apparatus for concurrent emulation of multiple circuit designs on an emulation system |
JP2002007497A (ja) | 2000-06-22 | 2002-01-11 | Fujitsu Ltd | 半導体装置設計支援装置 |
JP4534384B2 (ja) | 2001-05-18 | 2010-09-01 | 日本電気株式会社 | 半導体集積回路の自動フロアプランシステムとそのフロアプラン方法及びそのコンピュータプログラム |
US6496058B1 (en) * | 2001-07-24 | 2002-12-17 | Virtual Ip Group | Method for designing an integrated circuit containing multiple integrated circuit designs and an integrated circuit so designed |
US6938236B1 (en) * | 2002-03-29 | 2005-08-30 | Altera Corporation | Method of creating a mask-programmed logic device from a pre-existing circuit design |
US7093225B2 (en) * | 2002-07-17 | 2006-08-15 | Osann Robert Jr | FPGA with hybrid interconnect |
US6829756B1 (en) * | 2002-09-23 | 2004-12-07 | Xilinx, Inc. | Programmable logic device with time-multiplexed interconnect |
JP2009020786A (ja) | 2007-07-13 | 2009-01-29 | Renesas Technology Corp | ネットリストの配線難易度計算方法 |
JP5396781B2 (ja) | 2008-09-05 | 2014-01-22 | 富士通セミコンダクター株式会社 | 半導体集積回路の設計方法及び設計装置 |
-
2009
- 2009-11-27 US US13/130,215 patent/US8694948B2/en active Active
- 2009-11-27 JP JP2010540520A patent/JP5402940B2/ja active Active
- 2009-11-27 WO PCT/JP2009/070009 patent/WO2010061911A1/ja active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7051313B1 (en) * | 1999-10-29 | 2006-05-23 | Altera Toronto Co. | Automatic generation of programmable logic device architectures |
JP2006172219A (ja) * | 2004-12-16 | 2006-06-29 | Toshiba Corp | 半導体集積回路の自動設計装置、半導体集積回路の自動設計方法、及び半導体集積回路の自動設計プログラム |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2018526714A (ja) * | 2016-01-12 | 2018-09-13 | ▲騰▼▲訊▼科技(深▲セン▼)有限公司 | Cnn処理方法およびデバイス |
Also Published As
Publication number | Publication date |
---|---|
JP5402940B2 (ja) | 2014-01-29 |
US20110225558A1 (en) | 2011-09-15 |
JPWO2010061911A1 (ja) | 2012-04-26 |
US8694948B2 (en) | 2014-04-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10587269B2 (en) | Integrated circuit including an array of logic tiles, each logic tile including a configurable switch interconnect network | |
JP4275013B2 (ja) | データフローグラフ処理装置、処理装置、リコンフィギュラブル回路。 | |
Weaver et al. | The SFRA: a corner-turn FPGA architecture | |
JP6250548B2 (ja) | 再構成可能な半導体装置の論理構成方法 | |
US8316337B2 (en) | Method and system for optimally placing and assigning interfaces in a cross-fabric design environment | |
US8181139B1 (en) | Multi-priority placement for configuring programmable logic devices | |
US8527929B2 (en) | Method and system for optimally connecting interfaces across multiple fabrics | |
JP5402940B2 (ja) | 再構成可能回路生成装置、方法およびプログラム | |
Kung et al. | Mapping systolic arrays onto 3d circuit structures: Accelerating convolutional neural network inference | |
US8479134B2 (en) | Method and system for specifying system level constraints in a cross-fabric design environment | |
US8954906B1 (en) | Method and apparatus for performing parallel synthesis on a field programmable gate array | |
JPWO2010106738A1 (ja) | 再構成可能な論理回路 | |
Alexander et al. | Placement and routing for three-dimensional FPGAs | |
CN102893282A (zh) | 用于在综合期间执行异步和同步复位去除的方法以及设备 | |
US7409650B2 (en) | Low power consumption designing method of semiconductor integrated circuit | |
CN112131813B (zh) | 基于端口交换技术的用于提升布线速度的fpga布线方法 | |
JPH09305633A (ja) | 論理回路最適化方法 | |
US8074198B2 (en) | Apparatus and method for circuit layout using longest path and shortest path search elements | |
US7185299B1 (en) | Methods of estimating routing delays during the placement process in programmable logic devices | |
US20090210850A1 (en) | Method for Simplifying Tie Net Modeling for Router Performance | |
CN114722770A (zh) | 创建fpga电路的时延模型和获取时延的方法及设备 | |
Petelin et al. | Wotan: A tool for rapid evaluation of FPGA architecture routability without benchmarks | |
Taghavi et al. | Routing algorithms: architecture driven rerouting enhancement for FPGAs | |
Mittal et al. | Simulation analysis and performance evaluation of four innovative routers for network on chip | |
Mukherjee et al. | Nearly-2-SAT solutions for segmented-channel routing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 09829156 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13130215 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 2010540520 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 09829156 Country of ref document: EP Kind code of ref document: A1 |