Information processing device, dynamic reconfiguration element, and recording medium

By using dynamic reconstruction components and cross switches in FPGAs, the write position is determined according to the requirements of circuit resource and circuit reconstruction is solved, the problem of low resource utilization efficiency is achieved, the effective reconstruction of circuits and efficient utilization of resources is promoted, and the Miniaturization of FPGA.

CN112449070BActive Publication Date: 2024-12-06FUJIFILM BUSINESS INNOVATION CORP
View PDF 11 Cites 0 Cited by

Patent Information

Application Number
CN202010150556.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-08-29
Filing Date
2020-03-06
Publication Date
2024-12-06
Estimated Expiration
2040-03-06

AI Technical Summary

Technical Problem

In Field Programmable Gate Array (FPGA), the resource utilization efficiency of the variable constituting areas outside the fixed area is low, and the resource requirements of multiple circuits cannot be effectively met, resulting in the larger circuit size and the inability to write small-sized circuits into In a small area.

Method used

An information processing device is designed, including a dynamic reconstruction element, having a fixed first area and a variable second area, through the coordinated work of the cross switch and the processor, the writing position is determined according to the resource requirements of the circuit, and Switch and reconstruct the circuit during pipeline processing to achieve effective utilization of resources.

Benefits of technology

The resource utilization rate of the variable constituent areas is improved, the freedom of the circuit and the flexibility of pipeline processing are enhanced, the effective reconstruction of the circuit and the efficient use of resources are realized, resource waste is avoided, and the miniaturization of FPGAs is promoted.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112449070B_ABST
    Figure CN112449070B_ABST
Patent Text Reader

Abstract

Unlike the case where a region with a changeable configuration is prepared in such a way as to satisfy the total amount of resources required for a plurality of circuits selectively written into one region, the present invention seeks to effectively utilize the resources of the region with a changeable configuration. The present invention provides an information processing device, a dynamic reconfiguration element, and a recording medium, the information processing device comprising: a dynamic reconfiguration element having a first region with a fixed configuration, a second region with a changeable configuration, an input / output switch that bypasses the second region, and a cross switch that switches the connection between the first region and the second region; and a processor that, when writing a circuit to be reconfigured into the second region, determines the writing position of the circuit according to the resources used by the circuit.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to an information processing device, a dynamic reconstruction element and a recording medium. Background Art

[0002] Predetermined processing actions can be executed at high speed by using logic circuits. On the other hand, it is not cost-effective to prepare dedicated logic circuits for each processing action. Therefore, a field programmable gate array (FPGA) is used, which allows the configuration of logic circuits to be freely edited. An area is provided in the FPGA where the purchaser or designer can freely change the configuration of the circuit.

[0003] [Prior art literature]

[0004] [Patent Document]

[0005] Patent Document 1: Japanese Patent Application Publication No. 2016-035692 Summary of the invention

[0006] [Problems to be solved by the invention]

[0007] The configuration or wiring of the circuit is fixed outside the area where the configuration can be changed. Therefore, the circuit to be reconstructed must be written into the same area. On the other hand, the size of the area where the configuration can be changed in the FPGA varies. Of course, it is impossible to write the circuit to be reconstructed into an area that is smaller than the size of the circuit. Moreover, when there are multiple circuits that are selectively written into one area, the resources of each area must meet the sum of the resources required in the multiple circuits written.

[0008] However, if a rewritable area that satisfies the total resources required in a plurality of circuits is prepared in the FPGA, the circuit size of the FPGA increases.

[0009] Unlike the case where a region with a changeable configuration is prepared so as to satisfy the total amount of resources required for a plurality of circuits selectively written in one region, the present invention aims at achieving effective utilization of resources in the region with a changeable configuration.

[0010] [Technical means to solve the problem]

[0011] The invention described in technical solution 1 is an information processing device, comprising: a dynamic reconstruction element having a first area with a fixed structure, a second area with a changeable structure, an input / output switch that bypasses the second area, and a cross switch that switches the connection between the first area and the second area; and a processor that, when writing a circuit that is an object of reconstruction into the second area, determines the write position of the circuit based on the resources used by the circuit.

[0012] The invention described in claim 2 is the information processing device described in claim 1, wherein the crossbar switch is provided in the second area.

[0013] The invention described in claim 3 is the information processing device described in claim 1, wherein the processor instructs the dynamic reconfiguration element to switch the crossbar switch in accordance with the position of the second region written in the circuit.

[0014] The invention described in claim 4 is that in the information processing device described in claim 2 or claim 3, the switching of the crossbar switch is performed between the current pipeline processing and the next pipeline processing.

[0015] The invention described in Technical Solution 5 is that in the information processing device described in Technical Solution 1, when a circuit that is not used in the pipeline processing that starts next has been written into an area of ​​a part of the second area, the processor controls the switch corresponding to the part of the area to a bypass state during the execution of the next pipeline processing.

[0016] The invention described in Technical Solution 6 is that in the information processing device described in Technical Solution 5, when a part of the second area is controlled to be in a bypass state so that a circuit used in a pipeline processing to be started next can be written to the part of the area, the processor instructs the writing of the circuit in the part of the area in the current pipeline processing.

[0017] The invention described in Technical Solution 7 is a dynamic reconstruction element, which includes a first area with a fixed structure, a second area with a changeable structure, an input / output switch that bypasses the second area, and a cross switch that switches the connection between the first area and the second area, and through external control, the circuit to be reconstructed is written into the second area corresponding to the resources used by the circuit.

[0018] The invention described in claim 8 is that, in the dynamic reconfiguration element described in claim 7, the crossbar switch is switched so that the circuit written in the second area executes processing in a predetermined order.

[0019] The invention described in technical solution 9 is a recording medium storing a program, wherein the program enables a computer controlling a dynamic reconstruction element to execute the following functions, wherein the dynamic reconstruction element has a first area whose structure is fixed, a second area whose structure can be changed, an input / output switch that bypasses the second area, and a cross switch that switches the connection between the first area and the second area: a function of obtaining resources used by a circuit that is an object of reconstruction; and a function of determining a write position of the circuit based on the resources used by the circuit.

[0020] [Effects of the Invention]

[0021] According to the invention described in claim 1, unlike the case where a region with a changeable configuration is prepared so as to satisfy the total amount of resources required for a plurality of circuits selectively written in one region, it is possible to effectively utilize the resources of the region with a changeable configuration.

[0022] According to the invention described in claim 2, the degree of freedom of the circuit written in the region whose structure can be changed is improved.

[0023] According to the invention described in claim 3, it is possible to rewrite pipeline processing in a way that effectively utilizes resources.

[0024] According to the invention described in claim 4, the degree of freedom of the circuit written in the region whose structure can be changed can be improved.

[0025] According to the invention described in claim 5, the configuration can be reconfigured during the execution of pipeline processing.

[0026] According to the invention described in claim 6, the configuration can be reconfigured during the execution of pipeline processing.

[0027] According to the invention described in claim 7, it is possible to realize a device capable of effectively utilizing resources in a region where the configuration can be changed.

[0028] According to the invention described in claim 8, it is possible to realize an element that can perform pipeline processing even when the write position is determined according to the resources.

[0029] According to the invention described in claim 9, unlike the case where a region with a changeable configuration is prepared in a manner that satisfies the total amount of resources required for a plurality of circuits selectively written in one region, effective use of resources in the region with a changeable configuration can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a diagram for explaining a configuration example of an image forming apparatus used in Embodiment 1.

[0031] Figure 2This is a diagram for explaining a conceptual configuration example of an image processing unit including an FPGA.

[0032] Figure 3 (A) and Figure 3 (B) is a diagram illustrating a switching circuit configured in a static region. Figure 3 (A) indicates that the switch circuit has been controlled to bypass state. Figure 3 (B) represents a switch circuit that has been controlled to a non-bypass state.

[0033] Figure 4 (A) and Figure 4 (B) is a diagram for explaining the change in the execution order of pipeline processing by rewriting the crossbar switch. Figure 4 (A) shows the flow of data between areas before changing the connection of the crossbar switch. Figure 4 (B) shows the flow of data between areas after the connection of the crossbar switch is changed.

[0034] Figure 5 This is a flowchart for explaining a part of the processing operation executed by the processor used in the first embodiment.

[0035] Figure 6 This is a table that explains the relationship between the characteristics of resources used by the elements constituting the processing of each job and the reconfigurable area for writing positions.

[0036] Figure 7 This is a flowchart for explaining an example of processing operations executed by the processor used in the first embodiment after a job starts.

[0037] Figure 8 This is a diagram for explaining the elements constituting the processing of each job and the order of execution when three jobs are arranged in the image forming apparatus.

[0038] Fig. 9 This is a diagram for explaining the timing of executing the reconstruction of the image processing unit and the content of the reconstruction in relation to the execution of three jobs.

[0039] Fig.10 This is a diagram for explaining a case where the image processing unit is reconfigured between jobs.

[0040] Fig.11 This is a diagram for explaining another case where the image processing unit is reconfigured between jobs.

[0041] [Explanation of Symbols]

[0042] 1: Image forming device

[0043] 2: Processor

[0044] 3: Image processing unit

[0045] 4: Image forming unit

[0046] 5: Storage

[0047] 6: Operation and display unit

[0048] 31: Memory area

[0049] 32: DSP area

[0050] 33: Switching Circuit

[0051] 34: Reconfigurable Area DETAILED DESCRIPTION

[0052] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0053] <Implementation method 1>

[0054] <System Configuration>

[0055] Figure 1 1 is a diagram for explaining an example of the structure of an image forming apparatus 1 used in Embodiment 1. In this embodiment, the image forming apparatus 1 is an apparatus for forming an image on a medium such as paper. The medium such as paper is determined according to the purpose of the image forming apparatus 1 or the choice of the user. Furthermore, the image forming apparatus 1 is not limited to an apparatus for business use, but may also be an apparatus for individuals.

[0056] Figure 1 The image forming device 1 shown in the figure has: a processor 2 for controlling each part in the device; an image processing part 3 for applying color correction or harmonization processing to image data; an image forming part 4 for forming an image on a medium; a storage part 5 for storing image data, etc.; an operation and display part 6 for accepting operations performed by an operator or displaying information; and a network interface (IF) 7 for performing network communication. The above-mentioned parts are connected by signal lines 8 such as a data bus, an address bus, and a peripheral component interconnect (PCI) bus.

[0057] The processor 2 implements various functions by executing programs. The program in this embodiment is not limited to firmware or operating system, but also includes application programs.

[0058] The image processing unit 3 used in this embodiment is a circuit element that sequentially applies one or more predetermined processes to the image data input from the storage unit 5. That is, the circuits corresponding to the respective processes are connected in series inside the image processing unit 3. In other words, the output of the circuit corresponding to a certain process can be used as the input of the circuit corresponding to the next process to be performed. This form of processing is called pipeline processing.

[0059] Furthermore, in a part of the image processing unit 3 in the present embodiment, an area is provided in which the content of the circuit corresponding to each process or the execution order between the circuits can be changed. In the present embodiment, a circuit element having such a configuration is referred to as a reconfigurable circuit element. In the case of the present embodiment, the image processing unit 3 includes, for example, an FPGA. The image processing unit 3 here is an example of a dynamically reconfigurable element. In the present embodiment, the so-called dynamically reconfigurable element refers to an element that can reconfigure a part of the circuit configuration during its operation.

[0060] The image forming unit 4 is a device that forms an image on a medium such as paper by an electrophotographic method or an inkjet method, etc. The image forming unit 4 also includes a device that conveys paper.

[0061] The storage unit 5 includes, for example, a read-only memory (ROM) for storing firmware or a basic input output system (BIOS), a random access memory (RAM) used as a work area, a hard disk device or a semiconductor memory for storing programs or image data, and the like.

[0062] In addition, the processor 2 constitutes a computer by cooperating with the storage unit 5 .

[0063] The operation and display unit 6 includes switches or buttons arranged on an operation panel, a liquid crystal display or an organic electroluminescence (EL) display for displaying information, a touch sensor for detecting an operator's operation on a soft key displayed on the display, and the like.

[0064] The network IF 7 is used for communication with external devices via a network.

[0065] <Detailed structure of each part>

[0066] Figure 2 This is a diagram for explaining a conceptual configuration example of the image processing unit 3 including an FPGA. Figure 2 The image processing unit 3 shown in FIG. 1 includes an area where the circuit configuration is fixed (hereinafter referred to as a “static area”) and an area where the circuit configuration can be rewritten later (hereinafter referred to as a “reconfigurable area 34”).

[0067] The static region is an example of a first region whose configuration is fixed, and the reconfigurable region 34 is an example of a second region whose configuration can be changed.

[0068] In the static area, there are circuits corresponding to one or more processes (hereinafter referred to as "static circuits"), a memory area, a digital signal processor (DSP) area, and a switch circuit 33 (see FIG. 1 ) that connects the input and output of the reconfigurable area 34. Figure 3 (A) and Figure 3 (B)), data line, address line, etc.

[0069] Figure 3 (A) and Figure 3 (B) is a diagram for explaining the switch circuit 33 arranged in the static region. Figure 3 (A) indicates the switch circuit 33 which has been controlled to the bypass state. Figure 3 (B) shows the switch circuit 33 which is controlled to be in a non-bypass state. The control line for switching the switch of the switch circuit 33 also constitutes a part of the static region. The switch circuit 33 is an example of a switch that bypasses the second region.

[0070] like Figure 3 (A) and Figure 3 As shown in (B), one input of the switch circuit 33 is connected to the input side of the reconfigurable area 34 , and the other input of the switch circuit 33 is connected to the output side of the reconfigurable area 34 .

[0071] In the case of the bypass state, the switch of the switch circuit 33 is connected to the input side of the reconfigurable area 34. Therefore, the data input to the reconfigurable area 34 appears at the output of the switch circuit 33 unchanged.

[0072] In the case of the non-bypass state, the switch of the switch circuit 33 is connected to the output side of the reconfigurable area 34. Therefore, the data processed in the reconfigurable area 34 appears in the output of the switch circuit 33.

[0073] Back to Figure 2 Description.

[0074] exist Figure 2 In the case of , a plurality of reconfigurable areas 34 are arranged in the static area. Figure 2 In the case of , the number of reconfigurable areas 34 is four. Figure 2 In FIG. 1 , the reconfigurable area 34 is divided by using the characteristics of the resources on the hardware, and the characteristics of each resource are expressed as area A, area B, area C, and area D.

[0075] The area A is connected to the memory area 31 but is not connected to the DSP area 32. The area A is suitable for use in a circuit that has memory access but little arithmetic processing.

[0076] The area B is an area connected to both the memory area 31 and the DSP area 32. The area B is suitable for use in a circuit that performs a lot of memory access and arithmetic processing.

[0077] The area C is connected to the DSP area 32 but is not connected to the memory area 31. The area C is suitable for use in a circuit that does not require memory access and performs a lot of calculation processing.

[0078] The area D is an area that is not connected to both the memory area 31 and the DSP area 32. The area D is used to set a crossbar switch for a path connecting the static area and the reconfigurable area 34.

[0079] exist Figure 2 In FIG. 1 , four representative regions A, B, C, and D are depicted one each, but in reality, each region may exist in multiples. Figure 2 In the case of , regarding the area of ​​the reconfigurable region 34, the areas of region A and region B are substantially the same and the largest, the area of ​​region C is the second largest, and the area of ​​region D is the smallest, but the relationship between the areas is not limited to Figure 2 Moreover, even within the same region, there can be a relationship between the sizes of the areas.

[0080] Figure 4 (A) and Figure 4 (B) is a diagram for explaining the change in the execution order of pipeline processing by rewriting the crossbar switch. Figure 4 (A) shows the flow of data between areas before changing the connection of the crossbar switch. Figure 4 (B) shows the flow of data between areas after the connection of the crossbar switch is changed.

[0081] exist Figure 4 (A) and Figure 4 In (B), the elements constituting the static region are distinguished and represented as element S1, element S2, and element S3. In the case of this embodiment, element S1, element S2, and element S3 are, for example, buffer circuits. Of course, element S1, element S2, and element S3 may also be elements other than buffer circuits, and may correspond to different processes, respectively.

[0082] exist Figure 4 (A) and Figure 4 In the case of the example (B), the connection of the cross switch is changed, and the input to the image processing unit 3 (refer to Figure 1 ) is switched from “element S1→area A→element S2→area B→area C→element S3” to “element S1→area B→element S2→area C→area A→element S3”.

[0083] In addition, the cross switch cannot be changed during the execution of a job by the image processing unit 3. The job here is defined by the combination of the content of the processing and the order in which each processing is executed. The combination of the content of the processing constituting the job and the order in which the processing is executed varies, for example, depending on the type of image data to be processed. For example, the content of the job is different when the image data is a photo, when the image data is text, and when the image data is a mixed image of a photo and text. In the mixed image of a photo or text, there is, for example, a material for display.

[0084] <Processing Action>

[0085] Next, the image forming apparatus 1 (see Figure 1 ) will be described below. In addition, the image forming apparatus 1 is an example of an information processing apparatus.

[0086] Figure 5 This is an explanation of the processor 2 (see Figure 1 ) is a flowchart of a portion of the processing actions performed by the image processing unit 3 (refer to Figure 1 )'s circuit configuration can also be rewritten using a processor specifically prepared for the rewriting.

[0087] Figure 5 The processing shown in is implemented by executing a program by the processor 2. The program here may be a part of firmware or a part of an application program.

[0088] When the processor 2 receives the job (step 1), it determines the process to be reconstructed among the processes executed by the image processing unit 3 (step 2). The process here corresponds to the configuration of the circuit for realizing the process.

[0089] Next, the processor 2 determines the write position of the circuit corresponding to the process to be reconfigured according to the resources used by each circuit (step 3). In the case of this embodiment, the so-called resources mainly refer to the memory area 31 (see Figure 2 ) and DSP area 32 (refer to Figure 2 ).

[0090] Figure 6 The following describes the characteristics of resources used by the circuits corresponding to the processing of each operation and the reconfigurable area 34 (see Figure 2 )’s relationship table.

[0091] exist Figure 6 In the example of , expansion processing, coordination processing, screen processing, and edge processing are shown as examples of processing.

[0092] For example, in a circuit corresponding to the expansion processing, there is a resource feature that memory access exists but the calculation processing is small.

[0093] Furthermore, the circuit corresponding to the coordination process has a resource characteristic that both memory access and arithmetic processing are much more.

[0094] Furthermore, the circuit corresponding to the screen processing has the resource characteristics of no memory access and much calculation processing.

[0095] Furthermore, circuits corresponding to edge processing have a resource characteristic in that both memory access and arithmetic processing are high.

[0096] Therefore, area A is determined as the writing position of the circuit corresponding to the expansion process. Similarly, area B is determined as the writing position of the circuit corresponding to the coordination process, area C is determined as the writing position of the circuit corresponding to the screen process, and area B is determined as the writing position of the circuit corresponding to the edge process.

[0097] In the case of the present embodiment, since the writing position is determined according to the characteristics of the resources used by each circuit, the resources provided in the reconfigurable area 34 of the image processing unit 3 can be effectively utilized.

[0098] Incidentally, if the writing position of the circuit corresponding to each processing element is determined based only on the area of ​​the circuit, there is a possibility that the resource of the area serving as the writing position is not utilized. In other words, there is a possibility that the resource is wasted.

[0099] However, as described above, the processor 2 in this embodiment determines the reconfigurable area 34 as the write location according to the characteristics of the resources used by the circuit corresponding to the processing element, so there is no waste of resources. This is also conducive to the miniaturization of the FPGA constituting the image processing unit 3.

[0100] Next, the image processing unit 3 (see Figure 1 ) is explained below.

[0101] Figure 7 This is a description of the processor 2 used in the first embodiment (see Figure 1 ) is a flowchart of an example of processing actions performed after the job starts. S in the figure represents a step.

[0102] First, the processor 2 determines whether the current job has been completed (step 11). If the completion of the job has not been confirmed, a negative result is obtained in step 11. In this case, the processor 2 repeats the determination until the completion of the job is confirmed. On the other hand, if the completion of the job has been confirmed, a positive result is obtained in step 11.

[0103] If a positive result is obtained in step 11, the processor 2 determines whether there is a next job (step 12). In addition, the determination of whether there is a next job may be performed before the job being executed ends.

[0104] If there is no next job, a negative result is obtained in step 12. In this case, the processor 2 ends the process of controlling the reconstruction of the image processing unit 3. The original state is that there is no rewrite reconfigurable area 34 (refer to Figure 2 )’s necessity.

[0105] On the other hand, when there is a next job, a positive result is obtained in step 12. In this case, processor 2 determines whether circuit switching is required in the installed image processing (step 13).

[0106] Here, the case where the circuit switching is required includes, for example, the case where the crossbar switch needs to be reconfigured, the case where the reconfigurable area 34 corresponding to the characteristics of the resources used by the circuit needs to be determined, etc. In other words, the case where the circuit switching is required means that a rewrite that cannot be performed during the execution of the job is required.

[0107] When a positive result is obtained in step 13, processor 2 executes reconstruction of image processing unit 3 (step 14). The reconstruction here is executed between the completed job and the next job.

[0108] When the reconstruction of the image processing unit 3 is started, the processor 2 determines whether the reconstruction is completed (step 15). Before the completion of the reconstruction is confirmed, a negative result is obtained in step 15. In this case, the processor 2 repeats the determination of step 15.

[0109] On the other hand, in the case where the completion of the reconstruction has been confirmed, a positive result is obtained in step 15 .

[0110] When a negative result is obtained in step 13 or a positive result is obtained in step 15 , processor 2 determines whether or not there is a circuit in the installed image processing that is not required for the next operation (step 16 ).

[0111] In the case where there is an unnecessary circuit, a positive result is obtained in step 16. In this case, the processor 2 sets the target circuit as a bypass target (step 17). In this case, as in Figure 3 As described in (A) of FIG. 1 , the processor 2 switches the switch so that the output of the target circuit is not provided to the subsequent circuit. On the other hand, if there is no unnecessary circuit, a negative result is obtained in step 16 .

[0112] When a negative result is obtained in step 16 or after executing step 17, processor 2 starts a job (step 18). The job here is equivalent to the next job in step 12.

[0113] When the operation is started, the processor 2 determines whether the circuit used in the next operation needs to be switched in the area of ​​the circuit set as the bypass object (step 19). The next operation means the next operation of the operation started in step 18 and being executed.

[0114] In the case where switching is not required, a negative result is obtained in step 19. In this case, the processor 2 returns to step 11.

[0115] If switching is necessary, a positive result is obtained in step 19. In this case, processor 2 performs the reconfiguration of the area of ​​the circuit set as the bypass target (step 20).

[0116] For example, in the switch circuit 33 (see Figure 3 (A) and Figure 3 In the state where (B)) is controlled to be in a bypass state, the circuit of area B is rewritten from a circuit corresponding to the coordination process to a circuit corresponding to the edge process. The rewriting is performed simultaneously during the execution of the operation. Even during the execution of the operation, the area of ​​the corresponding circuit is in a bypass state, so the circuit can be reconfigured. In addition, in the reconstruction of the circuit, a data bus or an address bus configured for reconstruction can be used.

[0117] After performing the reconstruction, the processor 2 returns to step 11 .

[0118] <Specific example of processing action>

[0119] Hereinafter, reconstruction by the image processing unit 3 will be described using a specific example of the operation.

[0120] Figure 8 The image forming apparatus 1 (see Figure 1 ) is a diagram showing the elements that constitute the processing of each job and the order of execution when three jobs are arranged in the program.

[0121] exist Figure 8 In the case of , there are jobs #1, #2, and #3. The execution order is job #1, job #2, and job #3.

[0122] Job #1 includes processing using static area S, expansion processing using area A, processing using static area S, adjustment processing using area B, screen processing using area C, and processing using static area S. Area A, area B, and area C are respectively Figure 2The reconfigurable area 34 shown in FIG.

[0123] Job #2 includes processing using static area S, expansion processing using area A, processing using static area S, screen processing using area C, and processing using static area S. The difference from job #1 is that the coordination processing is not performed.

[0124] Job #3 includes processing using static area S, expansion processing using area A, processing using static area S, edge processing using area B, screen processing using area C, and processing using static area S. The difference from job #2 is that edge processing using area B is added.

[0125] Before starting each operation, the circuit configuration of the image processing unit 3 must be reconfigured in accordance with the operation to be started.

[0126] Fig. 9 This is a diagram for explaining the timing of executing reconstruction by the image processing unit 3 and the content of reconstruction in relation to the execution of three jobs.

[0127] Homework #1, Homework #2, Homework #3 and Figure 8 In addition, Fig. 9 In FIG. 1 , the symbol S is used to represent a static circuit using the static area, and the combination of the symbol and number of the area where the writing position is located is used to represent a circuit using the reconfigurable area 34. For example, A1 represents a circuit 1 written in area A. Fig. 9 In the case of , the circuit corresponding to the coordination processing is represented by B1, and the circuit corresponding to the edge processing is represented by B2.

[0128] In job #1 for processing image data corresponding to a photograph, since there is no other job executed first, the image processing unit 3 is reconfigured before starting job #1.

[0129] exist Fig. 9 In the case of, job #2 is executed after job #1, so processor 2 performs the following operations in step 12 (refer to Figure 7 ) obtains a positive result. The elements of the processing required for job #2 are in a relationship where the harmonization processing is removed from the image processing installed in job #1. Here, the harmonization processing is installed in area B. Therefore, processor 2 performs step 17 (refer to Figure 7 ), the area B in which the circuit B1 corresponding to the tuning process is written is set as a bypass target.

[0130] When job #2 is started, processor 2 confirms the configuration of job #3 to be executed next. Fig. 9 In the case of job #3 shown in FIG, edge processing must be performed in region B. Therefore, processor 2 performs step 19 (refer to Figure 7 ) obtains a positive result, and during the execution of job #2, the circuit B1 in area B that has been set to the bypass state is rewritten into the circuit B2 corresponding to the edge processing.

[0131] exist Fig. 9 In this case, after the completion of task #2, the processor 2 can start task #3 by controlling the switch circuit 33 to change the area B from the bypass state to the non-bypass state. The switching of the switch circuit 33 is different from the reconfiguration of the reconfigurable area 34 and is completed in a short time.

[0132] So, in Fig. 9 In the example shown in , the image processing unit 3 may not be reconfigured between job #1 and job #2, and between job #2 and job #3. In other words, the reconfiguration of the image processing unit 3 is hidden from the user.

[0133] Therefore, the time from the end of job #1 to the start of job #2 is shortened compared to the case where the image processing unit 3 is reconstructed during the period from the end of job #1 to the start of job #2. Similarly, the time from the end of job #2 to the start of job #3 is shortened compared to the case where the image processing unit 3 is reconstructed during the period from the end of job #2 to the start of job #3.

[0134] The more the number of jobs executed increases, the more significant the effect of shortening the time from the end of one job to the start of the next job becomes. In other words, by using the image forming apparatus 1 in this embodiment, the degradation of the image forming performance is suppressed compared to the case where the image processing unit 3 is not used. In other words, by using the image forming apparatus 1 in this embodiment, the number of jobs executed per unit time can be increased compared to the case where the image processing unit 3 is not used.

[0135] Fig.10 This is a diagram for explaining a state where the image processing unit 3 performs reconstruction between jobs.

[0136] exist Fig.10 In the case of job #11, job #12 executed after job #11 must write to the reconfigurable area 34 (refer to Figure 2 ) in area A, the installed circuit A1 is rewritten into circuit A2. Fig.10 In the case of , job #11 must execute circuit A1 which has been written into area A. Therefore, if Fig. 9 As in the case of , the reconstruction of area A cannot be performed during the execution of job #11.

[0137] In this case, the processor 2 performs step 13 (refer to Figure 7) obtains a positive result, and after the completion of job #11, the reconstruction of the image processing unit 3 is performed. Fig.10 In the case of the operation #11 and the operation #12 shown in FIG, the configuration other than the circuit A2 is the same. Therefore, the crossbar switch does not need to be reconfigured. In other words, the configuration of the crossbar switch remains unchanged and is used.

[0138] Fig.11 This is a diagram for explaining another case where the image processing unit 3 performs reconstruction between jobs.

[0139] exist Fig.11 In the case of , the content of the circuits constituting job #13 executed after job #11 is the same as job #11. That is, job #13 also uses circuit A1, circuit B1, and circuit C1. However, the execution order of each circuit in job #13 is different from the execution order of each circuit in job #11. Therefore, the crossbar switch needs to be reconfigured. As described above, the crossbar switch cannot be reconfigured during the execution of a job.

[0140] In this case, the processor 2 also performs step 13 (refer to Figure 7 ) obtains a positive result, and after the completion of job #11, the image processing unit 3 is reconfigured. Specifically, only the crossbar switch is reconfigured, and the execution order of each circuit is rearranged.

[0141] <Other embodiments>

[0142] Although the embodiments of the present invention have been described above, the technical scope of the present invention is not limited to the scope described in the embodiments. As is clear from the description of the claims, various changes or improvements made to the embodiments are also included in the technical scope of the present invention.

[0143] For example, in the first embodiment, as an example of an information processing device, an image forming device 1 (see Figure 1 ) has been described, but a device that executes multiple operations and requires reconstructing the circuit of the image processing unit 3 in each operation is included in the information processing device.

[0144] In the first embodiment, the memory area 31 (see Figure 2 ) and DSP area 32 (refer to Figure 2 ), but it is also possible to focus on features on other resources. Furthermore, even when focusing on the memory area 31 and the DSP area 32, when there are differences in performance and other aspects of the areas that can be used in the reconfigurable area 34, the write position of the circuit to be reconfigured can be determined by considering not only the resource but also the performance and other aspects.

[0145] In the first embodiment, the processor 2 of the image forming apparatus 1 controls the reconstruction of the image processing unit 3 . However, a processor in an external device such as a server that externally controls the image forming apparatus 1 may control the reconstruction in the image processing unit 3 .

[0146] In addition, the processor in the embodiment 1 refers to a processor in a broad sense, and includes special-purpose processors (such as a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic element, etc.) in addition to general-purpose processors (such as a central processing unit (CPU), etc.).

[0147] Furthermore, the actions of the processor in the first embodiment may be performed by a single processor alone or by a plurality of processors located in physically separate locations in cooperation. Furthermore, the execution order of the actions in the processor is not limited to the order described in the first embodiment, but may be changed individually.

Claims

1. An information processing device, comprising: A dynamic reconstruction element having a first area whose structure is fixed, a second area whose structure can be changed, an input / output switch that bypasses the second area, and a cross switch that switches the connection between the first area and the second area; and a processor, when writing a circuit that is an object of reconstruction into the second area, determines the write position of the circuit according to the characteristics of resources used by the circuit corresponding to the processing of the configuration job, wherein when a circuit that is not used in the pipeline processing to be started next has been written into an area of ​​a part of the second area, the processor controls the switch corresponding to the area of ​​the part to a bypass state during the execution of the next pipeline processing. 2 . The information processing device according to claim 1 , wherein the crossbar switch is provided in the second area. 3 . The information processing device according to claim 1 , wherein the processor instructs the dynamic reconfiguration element to switch the crossbar switch in accordance with the position of the second region written into the circuit. 4 . The information processing apparatus according to claim 2 , wherein the switching of the crossbar switch is performed between a current pipeline processing and a next pipeline processing.

5. An information processing device according to claim 1, wherein when a part of the second area is controlled to be in a bypass state so that a circuit used in a pipeline processing to be started next can be written to the part of the area, the processor instructs writing to the circuit in the part of the area in the current pipeline processing.

6. A dynamic reconstruction element, comprising a first area whose structure is fixed, a second area whose structure can be changed, an input / output switch that bypasses the second area, and a cross switch that switches the connection between the first area and the second area, and through external control, according to the characteristics of resources used by the circuit corresponding to the processing of the configuration job, the circuit that is the object of reconstruction is written into the second area corresponding to the characteristics, wherein when a circuit that is not used in the pipeline processing that starts next has been written into an area of ​​a part of the second area, during the execution of the next pipeline processing, the switch corresponding to the area of ​​the part is controlled to be in a bypass state.

7. The dynamic reconfiguration element according to claim 6, wherein the crossbar switch is switched so that the circuits written in the second area execute processing in a predetermined order.

8. A recording medium storing a program for causing a computer controlling a dynamic reconfiguration element having a first region whose configuration is fixed, a second region whose configuration can be changed, a switch between inputs and outputs for bypassing the second region, and a crossbar switch for switching connection between the first region and the second region to execute the following functions: a function of acquiring resources used by a circuit to be reconfigured; and A function for determining a write position of the circuit according to characteristics of resources used by the circuit corresponding to the processing constituting the job, When a circuit not used in the pipeline processing to be started next has been written into a part of the second area, during the execution of the next pipeline processing, the switch corresponding to the part of the area is controlled to a bypass state.

Citation Information

Patent Citations

  • Flexible reconfigurable task processing unit structure

    CN101620587A

  • Signal processing device, signal processing method, integrated circuit for signal processing, and television receiver

    CN102057575A

  • Business realization method and device for programmable devices and communication terminal

    CN107544819A

  • Integrated digital signal processing system supporting dynamic reconfiguration

    CN108182168A

  • Semiconductor integrated circuit capable of facilitating layout modification

    CN1501496A