JP5614266B2 - Print document processing system, cache device, data processing device, and program - Google Patents

Description

  The present invention relates to a print document processing system, a cache device, a data processing device, and a program.

  In general, print document data sent from a personal computer to a printing apparatus is called a page description language (hereinafter referred to as PDL) such as PostScript (registered trademark) or PDF (Portable Data Format: ISO 32000-1). Abbreviation). In a printing apparatus, print document data is converted into image data in a bitmap format (also called a raster format) by a data processing device called RIP (Raster Image Processor), and the image data is printed by a print engine. In addition to the method of directly converting PDL data to a raster image, PDL data is converted into intermediate language format data such as a display list having a smaller granularity than the PDL command and buffered. A printing apparatus that performs a two-stage conversion of a bitmap format is also well known.

  2. Description of the Related Art Conventionally, there is also known a system in which conversion from PDL data to image data in a bitmap format or an intermediate language format is executed in parallel by a plurality of data processing devices for each predetermined unit such as a page unit.

  Conventionally, when a data processing apparatus converts a document element (object) described in PDL in print document data into a bitmap format or an intermediate language format, the converted data is used as document element identification information. If the same or different data processing device needs to convert the same document element later, the conversion process can be omitted by using the cached data. It has been broken.

  For example, Patent Document 1 discloses an apparatus that performs print data drawing processing in parallel by a plurality of drawing processors. This apparatus is configured to cache bitmap data such as forms drawn by each drawing processor in a common cache memory in the printing mechanism control processor.

Japanese Patent Laid-Open No. 2000-33522

  By the way, in order to effectively utilize the limited capacity of the cache memory, the value of storing the image data of the document element in the cache memory is evaluated, and the one with a high evaluation value (cache priority) is preferentially stored in the cache memory. Management to leave it in is considered. When such management is performed, there are two situations when image data of a certain document element is not stored in the cache memory. One is a situation where it is not stored in the cache memory because the value to cache is low, and the other is not stored in the cache memory due to a reason that there is no free space although the value is sufficiently high to be cached. This is the situation. When the image data of the document element inquired from the data processing apparatus is not in the cache memory, the conventional apparatus that manages the cache memory caches the target image data without distinguishing between the two situations. A response (Cache Miss) is returned to indicate that it is not. Therefore, the data processing apparatus caches the created image data in the cache memory even if the image data of the document element is not worth caching.

  According to the present invention, in a system in which a plurality of data processing apparatuses share a cache memory, the data processing apparatus knows whether the image data of the document element is not stored in the cache memory because the value stored in the cache memory is low. The purpose is to be able to cope with it.

  The invention according to claim 1 includes a cache device and a plurality of data processing devices, wherein the cache device stores image data in a bitmap format or an intermediate language format created by each of the data processing devices. A state storage unit that stores, for each document element of the cache memory and the print document data, which of the plurality of predefined states is the cache state of the image data of the document element in the cache memory. In the plurality of states, there is a first state indicating that there is no image data in the cache memory despite the worth of storing the image data in the cache memory, and an image in the cache memory. A second indicating that the image data is not in the cache memory because the data is not worth storing. And a cache of image data of the document element stored in the state storage means when an inquiry about image data of the document element is received from the data processing device. A first response is provided to the data processing apparatus when the state is the first state, and a second response is provided when the image data cache state of the document element is the second state. Response providing means, and each of the plurality of data processing devices processes page description language data of each document element in the print document data to create bitmap format or intermediate language format image data For each document element in the print document data, before the image data creation means creates the image data of the document element, An inquiry means for making an inquiry as to whether or not the data is in the cache device; and if the response from the cache device to the inquiry by the inquiry means is the first response, the image data creation means sends the document Create the image data of the element, generate a print image using the image data created as a result, execute a process for registering the image data in the cache device, and inquire by the inquiry means If the response from the cache device to the second response is the second response, the image data creation unit is caused to create image data of the document element, and the image data created as a result is used for printing. Control means for generating an image and controlling the image data to be discarded without being registered in the cache device. A print document processing system characterized by that.

  According to a second aspect of the present invention, when the response providing means receives an inquiry about image data of a document element from the data processing device, the document element is stored even if the image data exists in the cache memory. 2. The print document processing system according to claim 1, wherein if it is determined that the image data is not worth storing in the cache memory, the second response is provided to the data processing device. It is.

  In the invention according to claim 3, the plurality of states include a third state indicating that image data of the document element is stored in the cache memory, and the cache device includes the data When a reservation request for securing a storage area for image data of a document element is received from a processing device, if the free space of the cache memory is insufficient to store the image data, A process for expelling at least one of the document elements in the third state having a cache priority lower than that of the target document element from the cache memory, and stored in the state storage unit; Eviction processing means for changing the cache state of the document element that is the target of the eviction processing to the second state, and the response provider Is received from the data processing device, and when the cache state of the document element stored in the state storage means is the second state, the document element is 3. The print document processing system according to claim 2, wherein it is determined that the image data is not worth storing in the cache memory.

  According to a fourth aspect of the present invention, when the eviction processing unit receives a securing request for securing a storage area for image data of a document element from the data processing device, the cache priority is higher than that of the document element. Even if the image data of all the document elements in the third state in which the image data is low and the image data is not being used by any data processing device is expelled, the free space of the cache memory is sufficient to store the image data. If not, the cache priority is lower than that of the document element and at least one of the image data being used by any of the data processing devices is changed to the second state and changed to the second state. Expelling the image data from the cache memory after the image data of the document element is no longer used by any data processing device. A print document processing system according to claim 3, characterized.

  In the invention according to claim 5, the plurality of states include a third state indicating that image data of the document element is stored in the cache memory, and the response providing means includes When an inquiry about image data of a document element is received from the data processing apparatus, when the cache priority of the document element is lower than a first threshold value for determining a document element for evicting image data from the cache memory Determines that the document element is not worth storing image data in the cache memory even if the cache state of the document element stored in the state storage means is the third state. The print document processing system according to claim 2, wherein a response of 2 is provided to the data processing apparatus.

  In the invention according to claim 6, when the response providing unit determines that the document element to be queried is worth storing image data in the cache memory, the cache priority is higher than the document element. If the free space of the cache memory is insufficient to store the image data even if the image data of all the document elements that are low and the image data is not in use from any data processing device is expelled, the first data A threshold value changing means for changing the threshold value of the document element to a value lower than the cache priority of the document element and higher than at least one cache priority of the document element in which image data is stored in the cache memory. A print document processing system according to claim 5.

In the invention according to claim 7, when the cache device receives a secure request for securing a storage area for image data of a document element from the data processing device, the cache device When the cache priority is less than a second threshold for determining a document element for registering image data in the cache memory, the cache state of the document element that is the target of the reservation request stored in the state storage unit Is changed to the second state , and further includes first state changing means for responding to the data processing apparatus that has made the reservation request that the reservation has failed. In response to a request for securing failure from the cache device in response to the request, the image data creation unit executes creation of image data of the document element, 7. The print document processing system according to claim 5, further comprising a unit that generates a print image using data and discards the image data without registering the image data in the cache device. 8. .

The invention according to claim 8 is characterized in that the second threshold value used by the first state change means is a minimum value among the cache priorities of the document elements in the third state. A print document processing system according to claim 7.

  According to a ninth aspect of the present invention, in the case where the cache priority of the document element to be secured is equal to or higher than the second threshold value, any data processing in which the cache priority is lower than the document element and the image data is processed. Even if the image data of all the document elements in the third state that are not in use by the apparatus is purged, if the free space of the cache memory is insufficient to store the image data, the second threshold value is set. Second threshold value changing means for changing to a value lower than the cache priority of the document element and higher than at least one cache priority of the document element in which image data is stored in the cache memory. A print document processing system according to claim 7 or 8.

  In the invention according to claim 10, the image data of the document element that is in the third state and whose cache priority is lower than the first threshold is not used by any data processing apparatus. 10. The image processing apparatus according to claim 5, further comprising: eviction means for expelling the image data from the cache memory and changing the cache state of the document element stored in the state storage means to a second state. The print document processing system according to any one of the above.

  According to an eleventh aspect of the present invention, when the cache device receives a securing request for securing a storage area for image data of a document element from the data processing device, the cache priority is higher than that of the document element. Even if the image data of all the document elements in the third state that are low and the image data is not being used by any data processing device is expelled, the free space of the cache memory is not sufficient to store the image data. And a means for responding to the data processing device that has made the reservation request that the reservation has failed, the data processing device responding to the reservation request from the cache device in response to the reservation request. When the image data is received, the image data creation means executes creation of image data of the document element, and prints an image for printing using the image data created as a result. While formed, comprising means discards without registering the image data in the cache unit, a print document processing system according to any one of claims 3-10, characterized in that.

The invention according to claim 12 stores the cache priority of each of the document elements in the state storage unit, and updates the cache priority so as to increase as the number of times the document element has received the inquiry increases. When the cache priority of the document element whose cache state is the second state exceeds a predetermined state change threshold, the cache state of the document element is changed to the first state . The print document processing system according to claim 1, further comprising a state change unit.

  According to a thirteenth aspect of the present invention, there is provided a cache memory for storing image data in a bitmap format or an intermediate language format created by each data processing device, and an image of each document element for each document element of print document data. State storage means for storing which of a plurality of predefined states the cache state of the data in the cache memory is, wherein image data is stored in the cache memory in the plurality of states A first state indicating that the image data is not in the cache memory even though the cache priority is higher than each document element, and more than each document element in which the image data is stored in the cache memory A second state indicating that no image data is present in the cache memory due to a low cache priority; And a cache state of the image data of the document element stored in the state storage means when the inquiry about the image data of the document element is received from the data processing device. A response providing a first response to the data processing apparatus when the data processing apparatus is in the first state, and a second response when the cache state of the image data of the document element is the second state. And a cache device.

According to a fourteenth aspect of the present invention, there is provided image data generating means for processing page description language data of each document element in the print document data to generate image data in a bitmap format or an intermediate language format, and in the print document data For each document element, before making the image data creation means create image data of the document element, an inquiry means for making an inquiry as to whether the image data is in the cache device, and an inquiry by the inquiry means A response from the cache device indicates that the document element to be inquired has higher cache priority than each document element in which image data is stored in the cache device , but there is no image data in the cache device. Is the first response indicating that the document element is in the first state, the image data creating means The image data creation is executed, and a print image is generated using the image data created as a result, and a process for registering the image data in the cache device is executed. A response from the cache device indicates that there is no image data in the cache device because the document element to be queried has a lower cache priority than each document element in which image data is stored in the cache device . In the case of the second response indicating the state 2, the image data creation unit executes creation of image data of the document element, and a print image is generated using the image data created as a result. And a control means for controlling to discard the image data without registering it in the cache device. It is the location.

  According to the fifteenth aspect of the present invention, there is provided a computer, a cache memory for storing image data in a bitmap format or an intermediate language format created by each data processing device, and each document element of print document data. State storage means for storing one of a plurality of predefined states of the cache state of the image data in the cache memory, the image data being stored in the cache memory in the plurality of states In a first state indicating that image data does not exist in the cache memory even though the cache priority is higher than each document element in which the image data is stored, than each document element in which the image data is stored in the cache memory Indicates that there is no image data in the cache memory because the cache priority is low. State storage means characterized in that the image data of the document element stored in the state storage means is received when an inquiry about the image data of the document element is received from the data processing device. When the cache state is the first state, a first response is sent to the data processing device, and when the cache state of the image data of the document element is the second state, a second response is sent to the data processing device. It is a program for functioning as response providing means to be provided.

According to a sixteenth aspect of the present invention, there is provided an image data generating means for processing a page description language data of each document element in print document data to generate image data in a bitmap format or an intermediate language format, and the print document. For each document element in the data, an inquiry means for making an inquiry as to whether or not the image data is in the cache device before causing the image data creation means to create image data of the document element, an inquiry by the inquiry means In response to the response from the cache device , image data is not in the cache device even though the document element to be queried has a higher cache priority than each document element in which image data is stored in the cache device . If the response is the first response indicating that the image is in the first state indicating that the The image data of the document element is created, a print image is generated using the image data created as a result, and a process for registering the image data in the cache device is executed. response from the cache apparatus to the inquiry is that the image data is not in the cache device to a low cache priority than the document element document element of the subject of the query image data in the cache unit are stored In the case of the second response indicating the second state indicating the image, the image data generating unit executes the generation of the image data of the document element, and the print image is generated using the image data generated as a result. And a control means for performing control to discard the image data without registering it in the cache device. Is a program of the order to.

  According to the inventions according to claims 1, 12, 13, 14, 15, and 16, it is determined whether or not the fact that the image data of the document element is not in the cache memory is because the value stored in the cache memory is low. Can know and deal with it.

  According to the second aspect of the present invention, it is possible to avoid that image data of a document element that is not worth storing in the cache memory continues to be used by any data processing device and cannot be deleted indefinitely.

  According to the third aspect of the present invention, whether or not it is worth storing the image data in the cache memory can be easily determined from the cache state.

  According to the fourth aspect of the present invention, the image data of the document element whose cache state is changed to the second state when the image data is being used from any of the data processing apparatuses is cached after it is no longer used. Can be kicked out of.

  According to the fifth aspect of the present invention, it is possible to determine whether or not there is image data in the cache memory, while determining whether or not there is image data in the cache memory, while determining whether it is worth storing the image data in the cache memory. it can.

  According to the sixth aspect of the present invention, image data of another document element having a lower cache priority than the document element to be newly stored in the cache memory can be evicted from the cache memory.

  According to the seventh aspect of the present invention, if a document element that is newly storing image data in the cache memory is not worth storing in the cache memory, the storage of the image data in the cache memory can be canceled. it can.

  According to the invention according to claim 8, if the document element for which image data is newly stored in the cache memory has a lower cache priority than all the document elements in which the image data is stored in the cache memory, The storage of the image data in the cache memory can be canceled.

  According to the ninth aspect of the present invention, it is possible to prevent image data of another document element having a lower cache priority than a document element to be newly stored in the cache memory from being stored in the cache memory.

  According to the tenth aspect of the present invention, the image data of the document element determined not to be cached can be expelled from the cache memory after it is no longer used.

  According to the eleventh aspect of the present invention, it is possible to cause the data processing apparatus to proceed without waiting for the necessary space in the cache memory to be created.

It is a figure which shows an example of a structure of the printing document conversion system of a 1st example, and a structure of the printing system to which this is connected. It is a figure which shows an example of the data content of an entry management table. It is a figure which shows the state transition between five stages of the cache entry in a 1st example. It is a flowchart which shows an example of the whole processing procedure of a RIP part. It is a flowchart which shows an example of the process sequence of a RIP part when response "Miss" is received from a cache management part. It is a flowchart which shows an example of the process sequence of a RIP part when response "Creating" is received from a cache management part. It is a flowchart which shows an example of the process sequence of a cache management part when an inquiry is received from a RIP part. It is a flowchart which shows an example of the process sequence of a cache management part when a use start request | requirement is received from a RIP part. It is a flowchart which shows an example of the process sequence of a cache management part when a use completion notification is received from a RIP part. It is a flowchart which shows a part of example of the process sequence of a cache management part when the area reservation request | requirement is received from a RIP part. It is a flowchart which shows the remaining part of an example of the process sequence of a cache management part when an area reservation request | requirement is received from a RIP part. It is a flowchart which shows an example of the process sequence of a cache management part when a registration request is received from a RIP part. It is a flowchart which shows an example of the process which ejects the cache data of the cache entry by which the eviction reservation was made. 10 is a flowchart illustrating an example in which cache data of a cache entry for which eviction reservation has been made is evicted during processing corresponding to a use end notification. It is a flowchart which shows an example of the procedure of the review process of the evicted cache entry. 12 is a flowchart illustrating a part of a processing procedure of a modified example 1-1 in which the cache management unit determines whether the RIP unit waits for completion of cache data being created in the first example. 12 is a flowchart illustrating a remaining part of a processing procedure of a modified example 1-1 in which the cache management unit determines whether the RIP unit waits for completion of cache data being created in the first example. In the first example, the RIP unit is a flowchart showing a main part of a processing procedure of a modified example 1-2 in which determination of whether to wait for completion of cache data being created is skipped according to its own load. It is a figure which shows the state transition between five stages of the cache entry in a 2nd example. It is a flowchart which shows an example of the process sequence of a cache management part when an inquiry is received from the RIP part in a 2nd example. It is a flowchart which shows an example of the process sequence of a cache management part when a use completion notification is received from the RIP part in a 2nd example. It is a flowchart which shows a part of example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 2nd example. It is a flowchart which shows the remaining part of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 2nd example. It is a flowchart which shows a part of example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in the modification 2-1 of a 2nd example. It is a flowchart which shows an example of the process sequence of a cache management part when it receives the inquiry from the RIP part in modification 2-2 of the 2nd example. It is a flowchart which shows an example of the process sequence of a cache management part when the use completion notification is received from the RIP part in modification 2-2 of the 2nd example. It is a flowchart which shows a part of example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in the modification 2-2 of a 2nd example. It is a flowchart which shows the remaining part of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in the modification 2-2 of a 2nd example. It is a flowchart which shows an example of the procedure of the review process of the evicted cache entry in the modification 2-2 of a 2nd example. It is a figure which shows an example of the data content of the entry management table in a 3rd example. It is a figure which shows the state transition between five stages of the cache entry in a 3rd example. It is a flowchart which shows a part of example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 3rd example. It is a flowchart which shows another part of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 3rd example. It is a flowchart which shows the principal part of the 1st system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 3rd example. It is a flowchart which shows an example of the process sequence of a cache management part when a registration request | requirement is received from the RIP part in the 1st system of a 3rd example. It is a flowchart which shows the principal part of the 2nd system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 3rd example. It is a flowchart which shows the principal part of the 3rd system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 3rd example. It is a flowchart which shows an example of the process sequence of a cache management part when a registration request | requirement is received from the RIP part in the 3rd system of a 3rd example. In a 3rd example, it is a flowchart which shows a part of example of the process sequence of the cache management part in the case of switching a 1st system-a 3rd system dynamically. In the third example, it is a flowchart showing the remaining portion of an example of the processing procedure of the cache management unit when dynamically switching the first method to the third method. It is a flowchart which shows the principal part of the 1st system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 4th example. It is a figure which shows an example of the data content of the entry management table used with the 2nd system of a 4th example. It is a flowchart which shows an example of the process sequence of a cache management part when the use start request | requirement is received in the 2nd system of a 4th example. It is a flowchart which shows an example of the process sequence of a cache management part when the use completion notification is received in the 2nd system of a 4th example. It is a flowchart which shows the principal part of the 2nd system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 4th example. It is a flowchart which shows the principal part of the 3rd system of an example of the process sequence of a cache management part when the area reservation request | requirement is received from the RIP part in a 4th example. In the fourth example, it is a flowchart showing a part of an example of the processing procedure of the cache management unit when dynamically switching the first method to the third method. In the fourth example, it is a flowchart showing the remaining portion of an example of the processing procedure of the cache management unit when dynamically switching the first method to the third method.

  Hereinafter, embodiments of the present invention (hereinafter referred to as “embodiments”) will be described with reference to the drawings. In the drawings, similar constituent elements or steps are denoted by the same reference numerals.

[First example]
"Device configuration"
An example of the configuration of the print document conversion system according to the first example of the present embodiment and the configuration of a print system to which the system is connected will be described with reference to FIG.

  The printing system shown in FIG. 1 includes a print document conversion system 100, a host computer 200, a print control device 210, and a printer engine 220.

  The print document conversion system 100 is an apparatus that converts print document data described in a page description language (PDL) into print image data such as a bitmap format (also called a raster format) that can be handled by the printer engine 220. Although not shown, the host computer 200 is connected to the print document conversion system 100 via a network such as a LAN (local area network). The host computer 200 sends print document data to the print document conversion system 100 via the network, and instructs printing of the print document data. Although only one host computer 200 is shown in FIG. 1, a plurality of host computers 200 may exist on the network.

  The print document conversion system 100 includes a job management unit 110, n (n is an integer of 2 or more) RIP units 120-1, 120-2, ..., 120-n, a cache management unit 130, and a cache memory 140.

  RIP (Raster Image Processor) units 120-1, 120-2,..., 120-n (hereinafter collectively referred to as RIP unit 120 when it is not necessary to distinguish between them) are data processing devices that perform RIP processing. . Here, the RIP process is a conversion process for converting print document data described in PDL into print image data in a bitmap (raster) format that can be handled by the printer engine 220. The print document conversion system 100 performs RIP processing in parallel on a page basis, that is, a maximum of n pages in parallel by a plurality (n) of RIP units 120. RIP processing is also called drawing or rendering. The RIP unit 120 examines the given PDL data in order from the beginning, for example, interprets the PDL data of each object (ie, document element) found in the process, and outputs an intermediate language format or bitmap format image. Generate data. Here, there are several types of objects such as text (character string), figure (graphics), and image (continuous tone image). Each object is represented by, for example, a PDL drawing command.

  In one example, the RIP unit 120 may directly convert PDL data into print image data such as a bitmap format. In another example, the RIP unit 120 converts the PDL data into intermediate language data such as a display list having a smaller granularity than the PDL command and buffers the buffered intermediate language data into a bitmap format. The two-stage conversion may be performed.

  The job management unit 110 receives print document data from the host computer 200 and manages print processing of the print document data. Hereinafter, one piece of print document data received from the host computer 200 is referred to as a job. Jobs (printed document data) from the host computer 200 include so-called variable print jobs in which data for each target stored in the database is combined with standard data and printed for each target (for example, a customer). obtain.

  The management of print processing performed by the job management unit 110 includes page allocation management in which each page in one job is assigned to each RIP unit 120 in addition to general print order management in units of jobs. The page allocation unit 112 is in charge of the latter page allocation management.

  For example, when the print document data (job) is page-independent (that is, a method of print document data in which information necessary for drawing one page is completely described in the PDL data of the page), the page allocation unit 112 May sequentially supply PDL data of different pages to each RIP unit 120 and cause the pages to be RIP processed. In another example, the job management unit 110 provides the entire print document data to all the RIP units 120, and the page allocation unit 112 assigns the page number to be processed in the print document data to each RIP unit 120. You may make it instruct | indicate in order (this system is applicable regardless of page independence and page non-independence).

  Also, the page allocation method for each RIP unit 120 is not particularly limited. For example, the page assignment unit 112 assigns pages to the RIP units 120-1, 120-2,..., 120-n in order of 1, 2,. The operation may be repeated. As another example, a method is also possible in which pages are assigned to each RIP unit 120 in order, and then the pages are assigned in the order in which the RIP processing of the assigned pages is completed.

  As yet another example, a method of estimating the RIP processing load amount of each page in advance (before the start of RIP processing) and scheduling pages to be allocated to each RIP unit 120 according to the estimation result is also conceivable. The RIP processing load amount of the page is obtained by, for example, summing the RIP processing load amounts of the individual objects included in the page over all the objects. The RIP processing load amount of each object is obtained from, for example, the complexity and size (area) of the object. The complexity of an object is determined according to the type of the object (image, font, graphics, etc.). In general, the larger the number of objects, the higher the processing load. Also, image objects generally have a higher processing load than graphic objects. Also, the larger the size, the higher the processing load generally. There are various known techniques for estimating the page processing load from such various parameters, and this known technique may be used in the present embodiment.

  Further, the current load of each RIP unit 120, the progress of the RIP processing of the page, the processing capacity (CPU speed, etc.), the RIP processing load amount of the page that each RIP unit 120 is processing, and the allocation target page Page allocation scheduling may be performed in consideration of one or more parameters such as the RIP processing load. In the scheduling, after comprehensively considering some or all of the exemplified parameters, for example, the RIP unit 120 that can complete the RIP processing of the allocation target page earliest is identified, and Assign pages. In this example, the page allocation unit 112 obtains the RIP processing load amount that has not been processed yet, for example, from the RIP processing load amount of the page that each RIP unit 120 performs RIP processing and the progress of the processing. . Then, the RIP processing load amount of the allocation target page is added to the remaining RIP processing load amount, and the addition result is divided by the processing capability (or the net processing capability that is the result of subtracting the current load from the processing capability). By doing this, an index value is obtained that represents the time required to complete the amount of processing that combines the remaining RIP processing load amount from the current time point and the RIP processing load amount of the allocation target page. Then, by obtaining and comparing such index values for each RIP unit 120, the RIP unit 120 that can complete the RIP processing of the allocation target page earliest is specified.

  Of the parameters for such evaluation, the processing capability of each RIP unit 120 may be registered in advance in the job management unit 110 by a system administrator or the like. Further, the RIP processing load amount of the page that each RIP unit 120 performs RIP processing or the allocation target page may be obtained as described above. The load monitoring unit 114 described below monitors the current load of each RIP unit 120 and the progress of RIP processing.

  The load monitoring unit 114 is a functional module that monitors the load state of each RIP unit 120. For example, the load monitoring unit 114 periodically receives information on the current load of the RIP unit 120 from each RIP unit 120. The load information of the RIP unit 120 includes, for example, the CPU (central processing unit) usage rate of the computer executing the program of the RIP unit 120, the number of active processes being executed by the CPU, and the virtual memory in the computer. There is information on whether it is in use (for example, whether memory swapping is occurring) or the like. The current load of the RIP unit 120 may be represented by the CPU usage rate, or other parameters such as the number of active processes and the usage status of the virtual memory may be obtained in addition to the CPU usage rate to obtain the load evaluation value. Good.

  In addition, the load monitoring unit 114 receives information on the progress degree of the RIP process for each assigned page (for example, up to what percentage of the page the RIP process is completed) from each RIP unit 120, for example, regularly or constantly. It may be acquired as the rate progresses. Further, the temporal change rate of the progress degree periodically acquired from the RIP unit 120, or the temporal change rate of the progress degree obtained from the time interval for each progress by a certain rate, and the RIP unit 120 performs the RIP The current net processing capacity of the RIP unit 120 is estimated from the data size of the middle page, and by using this estimated value, the amount of the remaining data size and the data size of the allocation target page is combined from the current time point. An index value representing the time required to complete the process may be calculated.

  Note that the load status information of each RIP unit 120 obtained by the load monitoring unit 114 can be used for cache management in addition to being used as the determination material of the page allocation unit 112 described above. That is, for example, when the RIP unit 120 performs RIP processing on an object and another RIP unit 120 has already started the RIP processing of the object, the former waits for the latter processing to be completed and the processing result (cache) (Data) can be used to determine whether or not it should be diverted (details will be described later).

  Each RIP unit 120 uses the cache memory 140 when performing RIP processing on the PDL data of the page allocated from the page allocation unit 112, so that the RIP of the past (that is, the page group before the allocated page) is used. Reuse processing results. That is, when each RIP unit 120 performs RIP processing on the PDL data of an object included in the allocated page and the RIP processing result of the object is already in the cache memory 140, the RIP processing is not performed and the cache memory 140 is not performed. By reusing the RIP processing result, the image of the object is added to the print image data of the page. An object is a group of drawing units, and is represented by, for example, one PDL drawing command. On the other hand, when the RIP processing result of the object is not in the cache memory 140, the RIP unit 120 performs the RIP process on the PDL data of the object, thereby adding the image of the object to the print image data of the page. The RIP processing result of the object generated at this time is registered (cached) in the cache memory 140 for later reuse.

  The RIP processing result (cache data) cached in the cache memory 140 may be in the bitmap format or may be in the intermediate language format. If the cache data read from the cache memory 140 is in bitmap format, the RIP unit 120 may simply add (synthesize) the cache data to the print image data of the page. If the cache data read from the cache memory 140 is in the intermediate language format, the RIP unit 120 generates a bitmap format image by further performing RIP processing on the cache data, and generates the generated bitmap. The format image may be simply added (synthesized) to the print image data of the page.

  The cache management unit 130 manages such read / write of each RIP unit 120 with respect to the cache memory 140 (reading of cached RIP processing results and registration of RIP processing results in the cache memory). The cache management unit 130 includes an entry management table 132, and controls cache usage by each RIP unit 120 based on the entry management table 132.

  FIG. 2 shows an example of the data contents of the entry management table 132. In the example of FIG. 2, each line other than the top heading line indicates a management record for one cache entry. A cache entry is prepared for each object in the print document data. The management record of each cache entry includes items such as ID, stage, RIP during data creation, cumulative query count, usage counter, cache area address, cache data size, score, and object attribute.

  “ID” is unique identification information for identifying a cache entry. In PDL, IDs are assigned to objects that can be reused many times in a job, such as images (bitmaps), fonts, forms (standard images), etc. Some describe the ID of each object. For example, PDF (Portable Document Format) is an example. When processing such print document data described in PDL, the object ID in the print document data may be used as the ID of the cache entry. As another example, the cache management unit 130 may identify each object and assign an ID. In this case, the object may be identified based on, for example, a combination of a PDL drawing command and a parameter group attached to the command. For example, a method may be used in which objects having the same combination of a drawing command and a parameter group attached to the command are determined to be the same object.

  “Stage” is information indicating the current state of the cache entry (in other words, the cache state of the RIP result of the object). A stage is also called a status. In the present embodiment, five stages of “New Cache”, “Caching”, “Cached”, “Cached Out”, and “Never Cache” are used. Details of each stage will be described later.

  “RIP during data creation” is an item for a cache entry in the “New Cache” (“now creating”) stage, and indicates identification information of the RIP unit 120 that is creating the cache data of the entry. This item is used when a certain first RIP unit 120 attempts to RIP an object and another second RIP unit 120 is creating cache data for the object. Is used as one material for determining whether the second RIP unit 120 waits for completion of creation of cache data or not (a specific example will be described later). There is an example in which the information of “RIP during data creation” is not used for this determination. In such an example, this item is unnecessary. Note that the identification information of the RIP unit 120 registered in this item is the cache data from the RIP unit 120 when a reservation failure is returned in response to a cache area allocation request (described later) from the RIP unit 120. When the registration request (described later) is received, the cache management unit 130 resets the value to a value indicating that no RIP is being created.

  The “total number of inquiries” is a cumulative value of the number of inquiries about the cache entry from the RIP unit 120 group until the current time after starting the RIP process from the top part of the top page of the print document data. is there.

  The “use counter” is a number indicating how many RIP units 120 currently use the cache data of the cache entry. Here, “use” of the cache data of the object by the RIP unit 120 means that the RIP unit 120 caches the object in order to synthesize the bitmap image of the object with the print image data of the currently created page. It means that data is read and used.

  The “cache area address” indicates an address (for example, a head address) in the cache memory 140 in which the cache data of the cache entry (that is, data in the bitmap format or intermediate language format of the RIP processing result) is stored. Of course, for the cache entry in which the cache data of the RIP result is stored in the cache memory 140 (that is, in the “Cached” stage described later), the address is registered in this item. The cache entry of an object that has been determined to be cached but for which cache data has not yet been completed (that is, in the “New Cache” stage described later) may be assigned an address and registered in this item. Conversely, the cache entry and cache data of an object (“Never Cache” stage described later) determined to have no value for caching the RIP result are evicted (that is, deleted) from the cache memory 140. For cache entries (in the “Cached Out” stage), the value of this item is null.

  The “cache data size” is information indicating the data size of the cache data of the cache entry, and is expressed in units of, for example, bytes, kilobytes, or megabytes. When the cache data is in the bitmap format, the data size corresponds to the area of the image represented by the cache data. For example, many PDLs explicitly specify the image size of an object such as an image, form, or font as an argument of a drawing command (for example, PDF). In such a PDL, for example, the RIP unit 120 determines the image size from the PDL description for the object. May be obtained, and the data size of the cache data may be obtained from the image size and registered in the entry management table 132.

  “Score” is an evaluation value indicating the priority of the cache entry. Since the capacity of the cache memory 140 is limited, if the cache data of all cache entries cannot be cached (stored) in the cache memory 140, the cache entry having a relatively high score (priority) The cache entry is cached in preference to the cache entry having a relatively low score. The score value is calculated based on a predetermined function or the like so that the higher the effect of using the cache, the higher the value. For example, an object with a larger RIP processing load has a greater effect when the cache data is reused.

  Further, in order to read the cache data from the cache memory 140, a certain overhead time is required regardless of the data size of the cache data due to the protocol such as preparation and completion of the read. The smaller the size, the greater the relative proportion. The smaller the cache data size, the shorter the time required to RIP the cache data from the original PDL data. In some cases, drawing from the original PDL data may be faster than reading the cache data. Accordingly, the smaller the size of the cache data, the smaller the significance / effect of caching the RIP result of the object.

  In addition, the greater the number of times the same object is drawn, the greater the significance and effect of caching the RIP result of that object.

  In the present embodiment, the score of each cache entry is calculated according to such a concept.

  As a parameter for determining the score, for example, there are types of objects. Since an image (bitmap) generally has a larger load of RIP processing than fonts and graphics (graphics), the effect of caching is increased, and thus the score is also increased.

  Further, the complexity of the object may be used as a parameter for determining the score. The complexity of an object is defined by, for example, the number and types of child objects included in the object. That is, an object such as a form (an object representing a fixed form) is represented by a combination of child objects such as an image and text. For such an object, for example, the complexity of the object is obtained by a method of summing the product of the number of child objects belonging to each object type and the weight specific to the type over all types. The weight specific to the object type is a positive number that becomes a larger value as the type having a higher cache effect. For example, image weights are greater than graphics and text weights. The higher the complexity obtained in this way (for example, the larger the numerical value), the higher the score. Of course, the method of obtaining complexity is not limited to this example. Complexity may be defined in consideration of the length (data size) of PDL data describing an object, the number of commands included in the data, and the like.

  Also, instead of uniformly handling child objects of the same object type, the attributes of the child object (for example, the size of the image when expanded to a bitmap, the type of processing necessary to draw the object, etc.) Accordingly, the weight of each child object may be obtained.

  For example, in the case of an image, an object that requires color space conversion (for example, conversion from RGB to CMYK) has a heavier processing load than an object that does not, so weight is given depending on whether or not there is color space conversion. It may be different. Moreover, since the processing load is heavier when the enlargement ratio for drawing is smaller, different weights may be used depending on the enlargement ratio. Further, the weight may be varied according to the rotation angle for rotating the original image data for drawing. In comparison of processing load between different rotation angles, for example, when the rotation angle is 0 degree, the processing load is lightest, and the rotation angle frequently used in increments of 90 degrees such as 90 degrees, 180 degrees, 270 degrees, etc. The processing load in this case is heavier than in the case of 0 degree, but lighter than in the case of other free rotation angles. Further, the larger the size of the original image data to be processed such as color space conversion, enlargement, and rotation, the greater the processing load because the number of pixels to be processed increases. Therefore, the weight may be varied according to the number of pixels of the original image data.

  Regardless of the type of object, whether the shape of the object (the shape in the case of a font is not the shape of a font but the shape of a bitmap representing a font) is a rectangle, that is, whether it is clipped Depending on whether or not (the clip is not necessary if the object is rectangular, the clip is necessary if the object is not rectangular), the weight may be varied. In general, the processing load is lighter in the case of a rectangle (not clipped).

  Also, in the case of a font object, the weight may be varied depending on whether or not the object is rotated or the rotation angle, as in the case of the image object.

  Further, when smooth shade processing (gradation processing) is performed for drawing an object, the processing load becomes heavy. In general, the processing load of a smooth shade is lighter than that of drawing an image, but heavier than that of drawing graphics or text. Accordingly, different weights may be used depending on whether smooth shade processing is performed. In addition, since the processing load also changes depending on the amount of color change by smooth shade processing within one object, when performing smooth shade processing, the weight may be varied according to the amount of color change in that processing. Good.

  The total weight of one object is the various attribute parameters exemplified above (for example, whether or not color space conversion is performed, enlargement ratio, rotation angle, necessity of clip processing, necessity of smooth shade processing, etc.). What is necessary is just to obtain | require weight by totaling (for example, multiplication, addition, etc. The method of calculation for specific total depends on implementation, such as how to determine the value of a weight).

  Further, the size of PDL data describing the object may be used as a parameter for determining the score. It can be said that the larger the size of the PDL data, the more time is required for interpretation processing for RIP processing, for example. Therefore, in one example, the score increases as the size of the PDL data increases.

  Further, based on the same concept, the time required for RIP processing of PDL data describing an object may be used as a parameter for determining a score. This required time may be obtained, for example, when the PDL data of the object is RIP processed.

  Further, the size of the cache data of the object (that is, the bitmap of the RIP processing result or the data in the intermediate language format) may be used as a parameter for determining the score. For example, for the reasons described above, the smaller the cache data size is, the smaller the significance and effect of caching the RIP result of the object is. Therefore, in one example, the larger the cache data size, the higher the score. In addition, in order to draw bitmap format cache data on a page (that is, to combine with print image data of the page), it takes time corresponding to the data size (proportional to the number of pixels). The size of is a parameter that determines the score.

  In the same way, the time required to transfer the cache data of the object between the cache memory 140 and the RIP unit 120 may be used as a parameter for determining the score of the object. For example, the longer the time required for transfer, the lower the overhead time ratio for cache transfer, so the penalty for caching is relatively low (in other words, the cache effect is high) I can say that.

  When intermediate language data is cached, the time required for the RIP unit 120 to RIP the intermediate language data to the bitmap format may be used as a parameter for determining the score. The shorter the required time, the greater the effect of reusing cache data, and the higher the score.

  Further, when the RIP unit 120 tries to start the RIP processing of an object, the RIP unit 120 inquires whether there is cache data of the object in the cache management unit 130, and if so, uses the cache data. It can be said that the greater the number of times, the higher the value of caching. Therefore, the above-mentioned cumulative inquiry count may be used as one of the parameters for determining the score. The higher the total number of inquiries, the higher the score.

  Note that the cumulative number of queries for each cache entry increases monotonously (that is, does not decrease) over time. Therefore, in one example, the scores of all the cache entries may be monotonically increased in accordance with the monotonous increase in the cumulative number of queries over time.

  In addition, because the essence of the score is a ranking of the superiority and inferiority of the value cached between the cache entries, in another example, the score of all the cache entries is a value within a predetermined range at each judgment time point. You may normalize to. In this case, even if the cumulative inquiry count increases, the score value itself does not always increase. However, if the cumulative query count increases more than other cache entries, the ranking of all scores in the cache entries increases.

  The parameters exemplified above are merely examples, and other parameters are also conceivable. The score may be calculated based on one or more of those parameters. For this calculation, the cache management unit 130 uses, for example, a function having one or more parameters as an argument, or a table in which scores are registered in association with one or more parameters. It should be noted that functions and tables for calculating scores representing cache priorities are also used in the past, and such known functions or tables may be used.

  Returning to the description of FIG. 2 again, the item “object attribute” registered in the entry management table 132 includes information of one or more attribute items of the object. Object attribute items include, for example, the type of object, the size of PDL data representing the object, the complexity of the object, and the like. Information on these attribute items may be obtained, for example, by analyzing the PDL description of the object using the method already described in the description of the score. Each of these attribute items is used to calculate the above-mentioned score, for example.

  The cache management unit 130 refers to the entry management table 132 and responds to read / write access to the cache memory 140 from each RIP unit 120. Further, the cache management unit 130 updates the data contents of the entry management table 132 in accordance with such access. The processing operation of the cache management unit 130 will be described in more detail later with a specific example.

  Each RIP unit 120 performs RIP processing on the PDL data of the page in charge, and among the objects in the page, those having cache data in the cache memory 140 are allocated by using the cache data. Print image data for the page is generated. Then, the RIP unit 120 sends the generated print image data of the page to the print control apparatus 210.

  Of the components of the print document conversion system 100 described above, the cache memory 140 is configured by a memory that can be read and written at high speed, such as a random access memory. In addition, each component other than the cache memory 140, that is, the job management unit 110, each RIP unit 120, and the cache management unit 130 is realized by causing a computer to execute a program that describes the functions of the respective components described above and later. Is done. In one example, the job management unit 110, each RIP unit 120, and the cache management unit 130 may be realized on a common computer. In this case, information exchange between the constituent elements may be performed by, for example, interprocess communication or interthread communication. In another example, the job management unit 110, each RIP unit 120, and the cache management unit 130 may be realized on separate computers. In this case, information exchange between the constituent elements may be performed using, for example, a network protocol that connects the constituent elements. In yet another example, by implementing two or more of these components on one computer, the entire print document conversion system 100 can be realized by a smaller number of computers than the number of components. Also good. For example, the job management unit 110 and the cache management unit 130 are mounted on the same computer and are managed by a database (for example, information on each object managed by the job management unit 110 and an entry management table managed by the cache management unit 130). It is also possible to unify 132). In either case, the cache memory 140 is constructed in a memory on a computer on which the cache management unit 130 is mounted. In either case, the computer to be used may be a single processor or a multiprocessor. When a multiprocessor computer is used, an operation such as assigning individual processor cores constituting the multiprocessor to different RIP units 120 is also conceivable.

  The printer engine 220 is printing hardware that prints an image represented by print image data on paper using a color material such as ink or toner. The print control apparatus 210 is an apparatus that controls the printer engine 220, and the print document conversion system 100 via a communication cable or a network (which may be the same as or different from the network between the host computer 200 and the print document conversion system 100). Communicate with. Through the communication, print image data of each page is received and various control information necessary for print control is exchanged. Also, the print control apparatus 210 accumulates the received print image data of each page and supplies it to the printer engine 220 according to the print order. The printer engine 220 prints the received print image data on paper.

  In a system in which a plurality of RIP units 120 operate in parallel while reusing RIP processing results via the cache memory 140, RIP processing of an object having a certain RIP unit 120 (referred to as a first RIP unit) There is a case where another RIP unit 120 (referred to as a second RIP unit) is executing the RIP process for the object at the time of starting the operation. In such a situation, the cache memory 140 does not yet have cache data as a result of the RIP processing of the object. In such a situation, there are two possible options for the first RIP unit. The first option is a process of waiting for the completion of the RIP process of the second RIP unit and using the cache data registered in the cache memory 140 upon completion. The second option is to perform RIP processing of PDL data of the same object without waiting for completion of RIP processing of the second RIP unit. However, in the conventional cache management method for the RIP result, whether there is cache data on the cache memory side (that is, “Hit” or “Miss”) in response to the query from the first RIP unit. Returns only a response. In other words, in the above situation, the cache memory side responds to the first RIP unit that there is no cache data of the inquired object (“Miss”). In this case, since it is not known that the cache data of the object is being created by another RIP unit, the first RIP unit has no choice in the first place, and only has to perform RIP processing on the PDL data of the object. In addition, the first RIP unit tries to register the RIP processing result in the cache memory so as to overlap the RIP processing result of the second RIP unit.

  On the other hand, in the present embodiment, when the cache data of the object inquired from the first RIP unit is already created in the second RIP unit but is not completed, the cache management unit 130 changes the status to the first. By making a notification to one RIP unit, the first RIP unit can take measures according to the situation. The configuration for this will be described in detail below.

<< State transition of the first example >>
First, the cache entry stage (Stage) used for management by the cache management unit 130 will be described with reference to FIG. In the present embodiment, five cache entry stages (that is, statuses) of “New Cache”, “Caching”, “Cached”, “Cached Out”, and “Never Cache” are used. For convenience of explanation, serial numbers are assigned to the five stages in order from 1. FIG. 3 is a state transition diagram between these five stages.

  <Stage 1> “New Cache” (new cache, that is, cache data “being created”) is a stage indicating that the cache data of the cache entry is being created and does not yet exist in the cache memory. When the cache management unit 130 receives an initial inquiry about cache data for an object from any RIP unit 120, the cache management unit 130 generates a cache entry corresponding to the object, and sets the stage of the cache entry to “New Cache”. In response to the response “Miss” (“no cache data”) to the RIP unit 120, the RIP unit 120 is caused to start RIP processing (that is, creation of cache data) of the object.

  In addition, when any RIP unit 120 makes an inquiry to a cache entry in a “Caching” stage, which will be described later, the stage of that entry transitions to “New Cache”.

  When there is a query from the RIP unit 120 to the cache entry in the “New Cache” stage, the cache management unit 130 responds with a code “Creating”. The response “Creating” indicates that the data of the cache entry is being created by another RIP unit 120.

  <Stage 2> “Caching” (waiting for free) is a state of waiting for the cache area to become free. The “Caching” stage is the same as the “New Cache” stage described above in that the cache data of the cache entry is not in the cache memory 140, but no RIP unit 120 is currently creating the cache data. Thus, it is different from the above “New Cache”.

  One example of transition to the “Caching” stage is the transition from the stage 1 “New Cache” described above. That is, when any RIP unit 120 attempts to create cache data for a cache entry in the “New Cache” stage, if the cache memory 140 does not have an area for storing the cache data, the cache entry Transitions to the “Caching” stage. The cache entry in the “Caching” stage waits for the cache data of another cache entry to be evicted (that is, deleted) from the cache memory 140 as time elapses, and a free area is created for the cache entry.

  However, the transition from “New Cache” to “Caching” is a case where it is determined that the cache entry is worth caching. If it is determined that the cache entry is not worth caching (for example, if the score is lower than any other cache entry that currently has cache data in the cache memory 140), the stage will be "Caching" Instead, the process shifts to “Cached Out” (outside the cache) described below.

  Further, as another example of transition to the “Caching” stage, there is transition from a stage 4 described later, that is, a “Cached Out” stage (details will be described in the description of the “Cached Out” stage).

  When there is an inquiry from any RIP unit 120 for a cache entry in the “Caching” stage, the cache management unit 130 changes the stage of the cache entry to 1, that is, “New Cache”, and “Miss” (“ "There is no cache data"). In response to this, the RIP unit 120 starts processing for creating cache data of the cache entry.

  <Stage 3> “Cached” (cached) is a state in which the cache data of the cache entry is stored in the cache memory 140. The “Cached” stage transitions from the above-described stage 1 “New Cache”. That is, when any RIP unit 120 tries to create cache data for a cache entry in the “New Cache” stage, an area for storing the cache data can be secured in the cache memory 140, and the cache data is stored in the area. Is created and registered, the cache entry transitions to the “Cached” stage.

  If there is an inquiry from any RIP unit 120 to a cache entry in the “Cached” stage, the cache management unit 130 responds with “Hit” (“cache data is present”). Upon receiving this response, the RIP unit 120 receives the cache data from the cache memory 140 and generates an image of the object instead of performing RIP processing on the PDL data.

  A cache entry in the “Cached” stage may transition to “Cached Out” (out of cache) when the free capacity of the cache memory 140 decreases. In general, when the cache data of a new cache entry is requested to be registered in the cache memory 140 and the amount of the cache data exceeds the free capacity of the cache memory 140, the cache data is in the “Cached” stage. The cache data of the entry selected from the cache entries is evicted (deleted) from the cache memory 140. What is evicted from the cache memory 140 is, for example, a cache entry whose cache data is not used (referenced) by any RIP unit 120 at that time. The stage of the cache entry from which the cache data is evicted in this way transitions from “Cached” to “Cached Out”. In addition, with respect to a cache entry in which cache data is being used (referenced) from any RIP unit 120 at that time, in order to avoid adverse effects on the RIP unit 120 being used, the cache data is not used until the use ends. It can't be kicked out, but it can be kicked out after use. Therefore, it is possible to make a reservation for eviction by changing the stage of the cache entry to “Cached Out” when the cache data is in use. Cache entries to be evicted may be narrowed down by score.

  <Stage 4> “Cached Out” (out of cache, that is, evicted) is a state in which the cache data of the cache entry has been evicted from the cache memory 140 or a eviction reservation has been made. The cache data of the cache entry at this stage is not in the cache memory 140 (in the case of “eject reservation”, the cache data is currently in the cache memory 140 but will be lost in the very near future), and The cache data is not created by any RIP unit 120. In the case of “evicting reservation”, the cache data is currently in the cache memory 140 but will be lost in the very near future.

  As described above, the stage shifts from the stage 3 “Cached” to the “Cached Out” stage.

  Further, in a specific case, “Create” indicates that the image data of the document element in the stage 1 is not in the cache memory and is in the “creating” state in which any data processing apparatus is creating the image data. When the response is “medium”, the cache entry in “New Cache” may be shifted to the “Cached Out” stage. For example, when the score of the cache entry is so low that the cache value is not considered to be sufficiently worthy, the cache entry is shifted to “Cached Out” without registering the cache data in the cache entry. An example of a case where the value of the cache entry is so low that it is not considered sufficient is that the score of the cache entry is lower than the score of any cache entry in the cache (“Cached”). Further, a threshold value of a score indicating whether or not it is worth the cache may be determined, and when the score of the cache entry is lower than the threshold value, the stage may be shifted to “Cached Out” without performing the cache. (See the example of FIG. 24 described later).

  If any RIP unit 120 makes an inquiry to a cache entry in the “Cached Out” stage, the cache management unit 130 responds with “Deleted” (“cache data has been deleted”). Upon receiving this response, the RIP unit 120 performs RIP processing on the original PDL data of the object, thereby generating print image data of the object. However, at this time, the RIP unit 120 does not register the generated print image data in the cache memory 140. This is because the cache entry is determined not to be worth caching cache data.

  The “Cached Out” stage is the same as the “New Cache” stage and the “Caching” stage described above in that there is no data of the cache entry in the cache memory 140. However, in the case of “New Cache” and “Caching”, “Miss” is responded to the query, whereas in the case of “Cached Out”, “Delete” is responded to the query. The difference between these two responses is that in the case of “Miss”, the inquiring RIP unit 120 attempts to register the image data created by the RIP process in cache, whereas in the case of “Delete”, the inquiring RIP unit 120. Is that the image data created by the RIP process is not registered in the cache. That is, the cache data in the “Cached Out” stage is not registered in the cache memory 140 even if an inquiry is made from the RIP unit 120. This is because, for example, the cache data of the cache entry once considered to be of low value to be cached avoids processing overhead due to repeated entry and exit of the cache memory 140. As described above, in this embodiment, the cache data of the cache entry once regarded as having a low value to be cached is not returned to the cache memory 140 for a while.

  However, as the number of queries for such cache entries increases, the score of the cache entry increases accordingly, and the value of caching relative to other cache entries that have data in the cache memory 140 is higher. Sometimes it gets higher. Therefore, in this embodiment, the cache entry in the “Cached Out” stage is periodically reviewed (for example, see the example of FIG. 15 described later). That is, the score of a cache entry in the “Cached Out” stage is periodically checked, for example, and if the score is high enough to deserve re-registration, the entry is transitioned to the “Caching” stage. When the stage of a cache entry transitions from “Cached Out” to “Caching”, when a query arrives after that entry, the stage of that entry transitions to “New Cache” and the cache data for that entry is created.・ The process for registration is started.

  <Stage 5> “Never Cache” (non-cache target) is a state in which an object corresponding to the cache entry violates a predetermined restriction condition and is determined to be out of cache target. This restriction condition is, for example, a condition that defines an object that is determined not to be cached without comparing the score with other cache entries.

  For example, an object that is too small or an object that is too simple (for example, only one line) is faster when the RIP 120 draws the cache data than it reads the cache data from the cache memory 140. Therefore, when a condition (for example, a threshold value) is set for each object size or complexity, and the size or complexity of the object to be cached satisfies a corresponding condition (for example, smaller or lower than the threshold value). Shifts the cache entry corresponding to the object to the “Never Cache” stage. It should be noted that the conditions for size or complexity can be combined in the AND condition (the transition to “Never Cache” only when all conditions are satisfied) or the OR condition (“Never Cache” only when one of the conditions is satisfied). Or may be either.

  In addition, from the viewpoint of efficiency, etc., if the shape of the object to be cached is limited to a rectangle parallel to the vertical and horizontal sides of the page, the object with a shape other than the rectangle will violate the restriction condition, and the object This cache entry is in the “Never Cache” stage.

  If any RIP unit 120 makes an inquiry to a cache entry in the “Never Cache” stage, the cache management unit 130 responds with “Invalid” (“cache invalid”). In response to this, the RIP unit 120 performs RIP processing on the PDL data of the object (and does not cache the RIP processing result).

  The cache entry stage has been described above. The above are the stages that the cache entry registered in the entry management table 132 of the cache management unit 130 can take. By the way, when the cache management unit 130 first receives a query about an object, the cache management unit 130 generates a cache entry for the object on the entry management table 132. Therefore, there is no cache entry before the first query. Therefore, there is no cache entry stage. However, since a cache entry corresponds to an object, it may be considered that a cache entry corresponding to the object exists potentially before receiving such an initial query. In view of this, before the first query is received, the state of the potential cache entry is “the cache data does not exist in the cache memory 140, and any RIP unit 120 does not store the cache data. It is considered to belong to the state of “not created”.

<< RIP section 120 processing procedure >>
Next, an example of a processing procedure of each RIP unit 120 will be described with reference to FIGS.

  The RIP unit 120 interprets the PDL data of the pages allocated from the page allocation unit 112 in order from the top. Among these, whenever the PDL data of a new object is found, the procedure illustrated in FIGS. 4 to 6 is executed.

  In this procedure, as shown in FIG. 4, it is first determined from the PDL data of the object whether the object is a cache target (S10). For example, in PDF, which is a kind of PDL, objects are roughly divided into types that can be used repeatedly, such as forms, images (bitmaps), and fonts, and types that are not. An ID is assigned to the object corresponding to the former, and when the object is used at a plurality of positions in the print document data, the object is called with the same ID. Conversely, an object without an ID is a type of object that is not supposed to be used repeatedly. Therefore, it is determined that an object to which an ID is assigned is a cache target, and an object to which no ID is assigned is not a cache target. In addition to PDF, there are PDLs that make the same distinction for objects, and the same determination as in S10 is applied to such PDLs.

  If it is determined in S10 that the object is not a cache target, the RIP unit 120 interprets and renders the PDL data of the object, generates a bitmap image of the object, and generates a bitmap image of the generated page. The print image data is combined (S12), and the RIP process for the object is completed. The bitmap image of the object generated at this time is not cached.

  On the other hand, when it is determined in S10 that the object is a cache target, the RIP unit 120 inquires of the cache management unit 130 about the state of the cache entry of the object (S14). This inquiry includes identification information for specifying the object. For example, when an ID is attached to an object to be cached such as PDF, the ID may be used for the inquiry. Since the cache entry has a one-to-one correspondence with the object, the cache entry corresponding to the object is specified from the object ID. For example, when the object ID is used as the cache entry ID, the corresponding cache entry is identified from the object ID. In the case of PDL in which no ID is attached to an object, for example, a pair of a PDL command describing the object and an argument may be included in the inquiry as identification information. In this case, a pair of a PDL command and an argument (or identification information linked thereto) may be registered in each cache entry as identification information of an object corresponding to the entry.

  After this inquiry, when receiving a response to the inquiry from the cache management unit 130, the RIP unit 120 determines a value indicated by the response (S16). As described above, there are five types of responses of the cache management unit 130 to the query: “Hit”, “Miss”, “Creating”, “Deleted”, and “Invalid”.

  If the response from the cache management unit 130 is “Hit”, the RIP unit 120 sends a cache data use start request corresponding to the object to the cache management unit 130 (S18). This use start request includes an ID for specifying a cache entry corresponding to the object. In response to this request, the cache management unit 130 provides the cache data of the requested cache entry from the cache memory 140 to the RIP unit 120. The RIP unit 120 generates a bitmap image of the object using the cache data, and synthesizes the generated bitmap image with the print image data of the page being created (S20). That is, if the cache data is in a bitmap format, the cache data may be directly combined with the page print image data. If the cache data is in an intermediate language format such as a display list, the object is drawn based on the cache data to generate a bitmap image, and the bitmap image is combined with the print image data of the page. When the cache data RIP processing is completed in this way, the RIP unit 120 sends a notification to the cache management unit 130 that the use of the cache entry is completed (“use completion notification”) to the cache management unit 130 ( S22). This notification includes an ID that identifies the cache entry used. With this notification, the RIP process for the object ends.

  Next, a processing procedure when the response from the cache management unit 130 is “Miss” in the determination of S16 will be described. As described above, the response “Miss” corresponds to an instruction “There is no cache data, so create and register”. Therefore, the RIP unit 120 executes the processing procedure illustrated in FIG.

  That is, the RIP unit 120 first issues an area reservation request to the cache management unit 130 (S30). The area securing request is a request for instructing to secure an area for caching the RIP processing result of the object in the cache memory 140. The area securing request includes, for example, information on the ID of the cache entry corresponding to the object and the size (data capacity) of the area to be secured. The size of the area to be secured is the data size of the RIP processing result (bitmap format or intermediate language format) of the object, and is obtained from the analysis of the PDL data. For example, many PDLs explicitly specify the image size of an object such as an image, a form, or a font as an argument of a drawing command. In such a PDL, the data size of cache data is obtained from information on the image size. This is because in the bitmap format, the size of the cache data is proportional to the image size. The amount of intermediate language cache data can also be estimated from the analysis of PDL data. In addition, since the amount of intermediate language cache data can be obtained accurately once RIP processing is actually performed, it is necessary to store it in, for example, the entry management table 132, and again to secure an area for the cache. You may make it refer in case.

  In response to this area securing request, the cache management unit 130 determines whether an area having the requested size can be secured (details will be described later), and returns the determination result to the requesting RIP unit 120. The RIP unit 120 determines whether the response indicates success or failure of the reservation (S31).

  When it is determined in S31 that the response is “successful reservation”, the RIP unit 120 performs RIP processing on the PDL data of the object, and the reserved area (the address of this area is included in the response). On the other hand, the result of the RIP process is written (S32). Thereby, cache data of the object is generated in the area. Next, the RIP unit 120 generates a bitmap image of the object using the generated cache data, and combines the generated bitmap image with the print image data of the page being created (S34). When the generation of the cache data and the synthesis of the page with the print image data are completed in this way, the RIP unit 120 issues a registration request for the cache data to the cache management unit 130 (S36). The registration request includes the ID of the cache entry. With this registration request, the RIP unit 120 ends the process for the object.

  If it is determined in S31 that the response is “reservation failure”, this means that there is no room in the cache memory 140 for caching the RIP processing result of the object. In this case, the RIP unit 120 generates a bitmap image by interpreting and rendering the PDL data of the object, and synthesizes the bitmap image with the print image data of the page (S38). Then, the process for the object is terminated without caching the generated bitmap image.

  Returning to the description of FIG. When the response from the cache management unit 130 is “Creating” in S16, the RIP unit 120 executes the processing procedure illustrated in FIG. As described above, the response “Creating” means that “the cache data to be inquired is currently being created by another RIP unit”. In this embodiment, such a characteristic response “Creating” causes the RIP unit 120 to know that another RIP unit is creating cache data of a target object. As a result, there is room for the RIP unit 120 to select whether to wait for the completion of the cache data by another RIP unit or to perform RIP processing by itself without waiting. Which one to select may be fixedly set, or may be dynamically selected according to the situation. FIG. 5 shows an example in which waiting / not waiting is dynamically selected.

In this procedure, the RIP unit 120 (referred to as a first RIP unit) waits for completion of the cache data by another RIP unit 120 (referred to as a second RIP unit) and uses the cache data. It is evaluated which process is completed earlier when the PDL data of the object is subjected to the RIP process (S42). This evaluation is performed using, for example, one or more of parameters (a1) to (f) exemplified below.
(A1) RIP processing load when RIP processing is performed on PDL data of the object to create a bitmap image (a2) RIP processing when RDL processing is performed on the PDL data of the object to generate intermediate language format data Load (this is used when caching intermediate language data)
(B) Data size of cache data (or time required to transfer the cache data to the first RIP unit 120)
(C) Processing load when converting the cache data of the object into a bitmap format image and compositing it to the print image data of the page (d1) Processing capacity of the first RIP unit (ideal CPU capacity, etc.) (The value in the condition. However, it is not necessary to consider when all RIP parts have the same processing capacity)
(D2) Processing capacity of second RIP unit (same as above)
(E1) Current processing load of the first RIP unit (e2) Current processing load of the second RIP unit (f) Progress of creation of cache data in the second RIP unit

  Of these parameters, (a1) and (a2) may be calculated by the same method as the calculation of the RIP processing load amount of the page by the load monitoring unit 114 described above. Further, for example, when the load monitoring unit 114 calculates the RIP processing load amount of the page, the RIP processing load amount of each object in the page is calculated and registered in the database using the object ID as a key. May be referred to.

  The parameter (b) may be acquired from the entry management table 132 of the cache management unit 130.

  Further, if the cache data is in the bitmap format, the parameter (c) does not need to be converted into the bitmap format, and is only the processing load required for combining the page with the print image data. The amount of processing load required for the synthesis is in accordance with the size of the data in the bitmap format, and can be obtained by multiplying the size by a coefficient obtained through experiments or simulations. Further, when the cache data is in an intermediate language format, it may be obtained by a known method from the number of pixels to be drawn from the cache data, the type of image processing to be performed in drawing, and the like. For example, regarding the number of pixels, if there is an instruction in the intermediate language format that draws a diagonal line of a rectangle having the same shape and size as the instruction to fill a rectangle, both instructions have substantially the same data size. Since the former has a much larger number of pixels to be drawn, the former has a much larger processing load. Therefore, it is indicated by the type of the instruction in the intermediate language format (this defines the shape of the pixel group to be drawn for the object. For example, the instruction is different between the rectangular fill and the diagonal drawing of the rectangle) and the argument of the instruction The parameter (c) may be calculated from the size of the object. As types of image processing to be performed in drawing, there are color space conversion, enlargement, rotation, smooth shade (gradation), and the like exemplified in the description of the score calculation method. Therefore, for example, for each type of image processing, a function for calculating the processing load amount is determined, and the size of the image represented by the cache data and the image processing parameters (for example, whether or not the image processing is performed) The amount of processing load related to the parameter (c) may be calculated by inputting a zoom rate, a rotation angle, a color change amount in gradation, and the like. When a plurality of image processes are performed on one object, the total processing load amount for each image processing may be obtained as the processing load amount related to the parameter (c) of the object. (The above score may be calculated in the same way.)

  Parameters (d1) and (d2) are known values and may be registered in advance by the system administrator. Since the same parameter is also used by the load monitoring unit 114, it may be shared.

  The parameters (e1), (e2), and (f) may be acquired from the load monitoring unit 114.

  In the evaluation process of S42 using these parameters, first, for example, from the parameters (a1), (d1), and (e1), the first RIP unit itself performs RIP processing on the PDL data of the object, and the object bitmap is obtained. The required time T1 for generating an image is estimated. For example, the current actual processing capability is obtained by reducing (d1) according to (e1), and the above-mentioned required time is reduced by, for example, dividing (a1) by this actual processing capability. What is necessary is just to obtain | require the evaluation value to represent.

  Similarly, from the parameters (b), (c), (d2), (e2) and (f), and (a1) or (a2), the completion of the cache data in the second RIP unit is awaited. The required time T2 for generating the bitmap image of the object is estimated by diverting it. The required time T2 is, for example, the time T21 until the cache data being created by the second RIP unit is completed, and the overhead of the protocol processing required to read the cache data from the cache memory 140 to the first RIP unit This is the sum of time T22 (this should be registered as a fixed value) and time T23 required for the first RIP unit to convert the read cache data into a bitmap image.

  Of these, the evaluation value representing T21 (time until the cache data being created by the second RIP unit is completed) is the parameter (a1) or (a2) (the former, intermediate language when the cache data is in bitmap format) In the case of the format, the remaining processing load R until completion of the cache data is obtained from (f) and (f), and the current processing capacity P2 of the second RIP unit is obtained from (d2) and (e2). What is necessary is just to obtain | require by calculation, such as remove | dividing the remaining processing load amount R by P2.

  T22 may be a fixed value and may be registered in advance.

  T23 is the sum of the time T231 required to read the cache data from the cache memory 140 to the first RIP unit and the time required for the first RIP unit to convert the read cache data into the bitmap format. It is. Of these, the time T231 may be obtained from the parameter (b). The time T232 may be obtained by calculation such as obtaining the current processing capability of the first RIP unit from the parameters (d1) and (e1) and dividing the parameter (c).

  The method of estimating the required times T1 and T2 described above is merely an example, and other methods may of course be used. For example, estimation may be performed using some of the parameters exemplified above, or other parameters may be used.

  A method corresponding to the shorter one of the required times T1 and T2 estimated as described above may be adopted. That is, if T1 is shorter than T2, it is determined that the method in which the first RIP unit RIPs the PDL data itself without waiting for completion of the cache data is good, and if the reverse is true, the first RIP unit stores the cache data. It is determined that it is better to use the system after waiting for completion.

  In the above, either case is advantageous by estimating and comparing the time required when the RIP unit 120 performs RIP processing of the PDL data by itself and when the other RIP unit waits for completion of cache data being created. This is just an example. For example, a simple determination based on one parameter may be performed such that the cache data is completed if the object type is an image, and the RIP process is performed without waiting if the object is other than an image (for example, form or font). Further, it may be determined that the object waits if the complexity of the object is equal to or greater than a predetermined threshold value, and does not wait if the object complexity is less than the threshold value. Of course, other determination methods are also conceivable.

  In S43, as a result of the evaluation in S42, it is determined whether to wait or not to wait. If it is determined in S43 that the first RIP unit 120 should wait for completion of the cache data, the first RIP unit 120 identifies the cache data (for example, the ID of the cache entry) to the cache management unit 130. (S44) is completed and registered (S44), and a notification is waited for (S45). When the notification is received from the cache management unit 130, a bitmap image is generated using the cache data and is combined with the print image data of the page (S46). Strictly speaking, this S46 includes three steps: a use start request, a bitmap generation, and a use end notification, as in S18 to S22.

  If it is determined in S43 that the first RIP unit 120 should not wait for completion of the cache data, the first RIP unit 120 performs RIP processing on the PDL data of the object to generate a bitmap image. Then, it is combined with the print image data of the page (S48). The bitmap format or intermediate language format data generated at this time is not registered in the cache memory 140 because it overlaps with the data generated and registered in the cache by the second RIP unit.

  By S46 or S48, the process for the object ends.

  Returning again to FIG. If the response from the cache management unit 130 is “Deleted” in S <b> 16, the RIP unit 120 that has received this response generates a bitmap image by performing RIP processing on the PDL data of the object, and converts it to the print image data of the page. Synthesize (S24). In this case, the bitmap format or intermediate language format data generated in S24 is not registered in the cache. This is because it is determined that the object is not worth caching.

  If the response from the cache management unit 130 is “Invalid” (stage is “Never Cache”) in S16, the RIP unit 120 that has received this response performs RIP processing on the PDL data of the object, and generates a bitmap image. Is generated and combined with the print image data of the page (S24). In this case, the bitmap format or intermediate language format data generated in S24 is not registered in the cache. This is because the object is not worth caching.

  In the above description, even in the case of an object corresponding to “Never Cache”, the cache management unit 130 is inquired in S14 and the RIP unit 120 performs RIP processing on the PDL data of the object in S24, but this is not essential. Instead, the RIP unit 120 itself determines whether or not the object corresponds to the stage “Never Cache” (the processing load of this determination situation is extremely small). Inquiry to 130 may be omitted. Further, when the RIP processing result of an object determined to correspond to “Never Cache” is cached locally by the RIP unit 120 and the same object is rendered again, the local cache is not inquired without inquiring of the cache management unit 130. You may draw using.

  The stage “Cashed Out” (response is “Deleted”) is similar to “Never Cache” in a certain point of view, but “Never Cache” can omit the inquiry to the cache management unit 130 as described above. In contrast, “Cashed Out” is different in that it is not omitted. That is, in the case of an object of the stage “Cashed Out”, the score may change as time passes and may be cached in the shared cache memory 140. Therefore, the RIP unit 120 encounters such an object every time it encounters such an object. In addition, it is necessary to make an inquiry to the cache management unit 130.

<< Processing Procedure of Cache Management Unit 130 >>
The example of the processing procedure of each RIP unit 120 has been described above. Next, an example of the processing procedure of the cache management unit 130 will be described with reference to FIGS.

  First, an example of a processing procedure executed by the cache management unit 130 when an inquiry (S14 in FIG. 4) arrives from any RIP unit 120 will be described with reference to FIG.

  In this procedure, the cache management unit 130 searches the entry management table 132 for the cache entry of the object to be queried (S50). In S50, information specifying a cache entry (for example, an object ID or a combination of a PDL command and an argument) included in the query may be searched for as a key. Then, it is determined whether or not a cache entry corresponding to the object is found (S52). If not found, a cache entry corresponding to the object is newly created (S54). In the case of using the object ID as the cache entry ID, the object ID included in the query may be set to the cache entry ID. If there is no object ID, a unique ID is assigned to the newly generated cache entry, and information specifying the cache entry such as a PDL command and argument pair included in the query is registered in the cache entry. (Omitted in FIG. 2). In addition, the cache management unit 130 sets each item of the cache entry in the entry management table 132 (see FIG. 2) such as setting the cache entry stage to 1, that is, “New Cache”, and setting the cumulative inquiry count to 1. Is initialized or set (S56). Object attribute information such as the type of object may be included in a query from the RIP unit 120, for example, and the cache management unit 130 may set it in the entry management table 132, and the size of the cache data is the same. Also, the value of the usage counter is initialized to zero. The address of the cache area may be left undecided. Further, the score of the cache entry (evaluation value indicating the value to be cached) is calculated as described above from the object attribute (object type and the like) and the cumulative value matching count, and registered in the entry management table 132. In addition, the identification information of the RIP unit 120 as the inquiry source is registered in the column of “RIP being created”. Then, the cache management unit 130 returns a response “Miss” to the RIP unit 120 that is the inquiry source (S58), and ends the processing for this inquiry. In response, the RIP unit 120 starts creating cache data for the object (see FIGS. 4 and 5).

  In S52, when a cache entry to be queried is found from the entry management table 132, the cache management unit 130 increments the cumulative inquiry count of the entry by 1 (S60). Since the score increases as the cumulative inquiry count increases as described above, the score of the entry is updated as the cumulative query count increases (S60).

  Next, the cache management unit 130 determines the stage of the cache entry (S62). As a result of the determination, if the stage is 1 “New Cache”, a response “Creating” is returned to the RIP unit 120 as the inquiry source (S64). If the stage is 2 “Caching”, “Miss” is returned to the RIP unit 120 as the inquiry source, and the stage of the cache entry is changed to 1 (S66). At this time, the identification information of the RIP unit 120 that is the inquiry source is registered in the “RIP during data creation” column of the entry in the entry management table 132. If the stage is 3 "Cached", "Hit" is set (S68). If it is 4 "Cached Out", "Deleted" is set (S70). If it is 5 "Never Cache", "Invalid" is set (S72). , Respectively, to the RIP unit 120 as the inquiry source. With this response, the cache management unit 130 ends the processing for the inquiry. In response to this response, the RIP unit 120 that is the inquiry source executes the processing shown in FIGS.

  Next, an example of a processing procedure of the cache management unit 130 when a cache data use start request is received from any of the RIP units 120 will be described with reference to FIG.

  In this procedure, the cache management unit 130 obtains the cache data that is the target of the received use start request from the cache memory 140 and returns it to the requesting RIP unit 120 (S80). For this, for example, the entry management table 132 is searched using information specifying a cache entry such as an ID included in the received use start request as a key, and the “cache area address” in the searched cache entry is searched. ”And“ cache data size ”(see FIG. 2), the cache data may be read from the cache memory 140. Then, the cache management unit 130 adds 1 to the value of the usage counter (see FIG. 2) of the cache entry (S82). Thereby, it is recorded that the number of RIP units 120 currently using the cache data of the cache entry is increased by one.

  Next, an example of a processing procedure of the cache management unit 130 when a cache data use end notification is received from any of the RIP units 120 will be described with reference to FIG.

  In this procedure, the cache management unit 130 identifies the cache entry that is the subject of notification from entry identification information such as an ID included in the received use end notification, and subtracts 1 from the value of the usage counter of the cache entry. (S85). As a result, it is recorded that the number of RIP units 120 currently using the cache data of the cache entry has decreased by one.

  Next, an example of a processing procedure of the cache management unit 130 when a cache area securing request is received from any of the RIP units 120 will be described with reference to FIGS. 10 and 11.

  In this procedure, as shown in FIG. 10, first, the cache management unit 130 specifies a cache entry to be notified from entry specifying information such as an ID included in the area reservation request. Then, based on the cache entry information and the like, whether or not the cache data of the cache entry can be cached in the cache memory 140 is evaluated from both the limiting condition and the free capacity of the cache memory 140 (S90). As described above, the restriction condition is a condition for identifying an object that is not suitable for caching in terms of complexity, size, shape, or the like of the object, or that is clearly less efficient when cached. This determination is performed using registration information in the entry management table 132 for the identified cache entry, for example, information such as cache data size and object attributes (for example, object type and complexity). Further, the cache data size and object attribute information may be received from the RIP unit 120 together with the area reservation request and registered in the entry management table 132.

  As a result of the evaluation in S90, when it is determined that the restriction condition is not cleared (that is, not suitable for the cache) (the determination result in S92 is NO), the cache management unit 130 calls the identified cache entry (hereinafter referred to as “target entry”). ) Is changed to 5 “Never Cache” (S94), and a code indicating failure to secure the area is returned to the requesting RIP unit 120 (S96). In response to this response, as shown in S38 of FIG. 5, the RIP unit 120 generates a bitmap image using the PDL data of the object, combines it with the page image, and caches the bitmap image. The process ends without

  As a result of the evaluation in S90, when it is determined that the restriction condition has been cleared (the determination result in S92 is YES), the cache management unit 130 further determines the capacity of the free area of the cache memory 140 (area where cache data is not stored). Is sufficient to store the cache data of the object (S98). In this determination, information on the size of the cache data is referred to. If it is determined in this determination process that the capacity of the free area is sufficient, the cache management unit 130 secures an area according to the size of the cache data from the free areas in the cache memory 140, and The address of the secured area is returned to the requesting RIP unit 120 together with a code indicating the success of the securing (S100). In response to this response, the RIP unit 120 executes the processes of S32 to S36 in FIG. At this time, the cache management unit 130 may register the address of the reserved area (for example, the head address) in the “cache area address” of the target entry of the entry management table 132.

  If it is determined in S98 that the capacity of the free area is not sufficient, the cache management unit 130 acquires the score of the target entry from the entry management table 132 (S102), and further, among the cache entries registered in the entry management table 132, Then, all the cache entries whose stage is 3 “Cached” and whose score is lower than that of the target entry (hereinafter referred to as “inferior entry”) are extracted (S104).

  As described above, the score of the cache entry is an evaluation value representing the value (effect) of caching the cache data of the cache entry in the cache memory 140. Accordingly, it can be said that the inferior entry is a cache entry that is relatively less cached than the target entry. In the present embodiment, if there is not enough free space in the cache memory 140 to accommodate the cache data of the target entry, the cache data of such inferior entries is sufficiently removed from the cache memory 140 (deleted). Try to make free space.

  In S106, it is determined whether or not even one inferior entry is found in S104. If no subordinate entry is found, the target entry satisfies the limit condition that is the minimum condition (in other words, an absolute condition) that defines the cache target, but the data is already cached in the cache memory 140. This means that the cache value is relatively low when viewed from any other entry. Therefore, in such a case, the data of the target entry is not cached. That is, when the determination result in S106 is NO, the cache management unit 130 changes the target entry from the current stage 1 “New Cache” to the stage 4 “Cached Out” (S108), and sends it to the requesting RIP unit 120. A code indicating a failure to secure the area is returned (S110). In response to this response, as shown in S38 of FIG. 5, the RIP unit 120 generates a bitmap image using the PDL data of the object, combines it with the page image, and caches the bitmap image. The process ends without

  If it is determined in S106 that one or more inferior entries have been found (the determination result is YES), the cache management unit 130 executes the procedure illustrated in FIG.

  That is, it is first determined whether or not there is a usage counter value of 0 in the found inferior entries (S111). A usage counter value of 0 means that no RIP unit 120 uses the cache data of the cache entry (inferior entry), so even if the cache data is deleted, other RIP units are adversely affected. Absent. Therefore, if there is an inferior entry with a usage counter value of 0, one with the lowest score is extracted (S112). Then, the address and size of the cache data of the extracted inferior entry are obtained from the entry management table 132, the cache data is evicted (deleted) from the cache memory 140 from the address, and the inferior entry is removed from the current stage 3 “Cached”. Stage 4 is changed to “Cached Out” (S114). By this eviction, the area in the cache memory 140 that was previously occupied by the cache data is released. The cache management unit 130 joins the released area to the free area, and the capacity of the free area that is expanded (that is, the free capacity of the cache memory 140) is necessary to accommodate the size of the cache data of the target entry. It is determined whether the size has been reached (S116). In this determination, if it is determined that the free space has not reached the required capacity, the cache management unit 130 returns to S111, and while there is an inferior entry with a usage counter value of 0, the score is low among them. The cache data is evicted and the stage is changed one by one (S112, S114).

  In this way, when only the inferior entry whose usage counter value is 0 is expelled, the free space necessary for caching the target entry can be secured (the determination result in S116 is YES), the cache management unit 130 Then, an area corresponding to the size of the cache data is secured from the free area of the cache memory 140, and the address of the secured area is returned to the requesting RIP unit 120 together with a code indicating the successful area securing (S118). ). Thereby, the processing of the cache management unit 130 in response to the area securing request is completed. In response to the response, the RIP unit 120 executes the processes of S32 to S36 in FIG. At this time, the cache management unit 130 may register the address of the reserved area in the “cache area address” of the target entry in the entry management table 132.

  If all the subordinate entries whose usage counter value is 0 are evicted from the cache memory 140 and the free space required to cache the target entry cannot be secured, the determination result in S111 is NO. This means that the cache data of the target entry cannot be cached unless the cache data currently used by any RIP unit 120 is evicted. In this case, in this embodiment, in order to avoid adversely affecting the processing of the RIP unit 120 that currently uses the cache data of the inferior entry, for example, the cache data of the target entry is cached for the current request. Give up. However, since the target entry has a higher value (score) to be cached than the inferior entry currently used, preparation is made for caching instead of leaving it up.

  That is, in the procedure of FIG. 11, when the determination result in S111 is NO, the target entry is changed from the current stage 1 “New Cache” to the stage 2 “Caching” (S122). As described above, in stage 2 “Caching”, the entry is worth caching as compared to other cases, but the entry is cached unless the cache data of the inferior entry in use is evicted. This indicates a “waiting for free” state where a sufficient area cannot be secured. Then, a response indicating a failure to secure the area is returned to the requesting RIP unit 120 (S124). In response to this response, as shown in S38 of FIG. 5, the RIP unit 120 performs RIP processing on the PDL data of the object to generate a bitmap image, combines it with the page image, and caches the RIP processing result. The process is terminated without doing so.

  This case is the same as when the target entry violates the restriction condition (S92 to S94 in FIG. 10) or when there is no inferior entry (S106 to S110) in that the RIP processing result generated at the corner is not cached. is there. However, both of these cases are cases where the value of caching the target entry is absolutely or relatively low compared to others, and the stage will be set to “Never Cache” or “Cached Out”. It is controlled not to be cached for a long time or for a while. On the other hand, in this case where the determination result in S111 is NO, the target entry is relatively high in cache value, so the target entry is placed in a special stage called “Caching” and either When the same target entry is inquired from the RIP unit 120, the cache registration of the target entry is retried.

  Further, the cache management unit 130 extracts inferior entries whose use counter value is not 0 one by one in ascending order of score (S126), and the extracted inferior entries are changed from stage 3 “Cached” to stage 4 “Cached Out”. Change (S128).

  Then, it calculates the free capacity of the cache memory 140 when it is assumed that the cache data of the inferior entry changed to “Cached Out” is evicted, and the free capacity reaches an amount sufficient to cache the cache data of the target entry. Steps S126 and S128 are repeated until (S130). When the necessary free space is secured, the processing of the cache management unit 130 in response to the area securing request ends.

  In S114 described above, the cache data of the extracted inferior entry is evicted from the cache memory 140. However, in S128, the cache data of the extracted inferior entry is still used by another RIP unit 120. Do not do. However, since the extracted subordinate entry is changed to stage 4 “Cached Out”, the cache data is evicted from the cache memory 140 after it is not used by any RIP unit 120 (details). Will be described later). S128 may be regarded as an eviction reservation for an inferior entry in use. In this way, by making an eviction reservation for an entry having a lower score than the target entry (an inferior entry), for example, the possibility that the target entry is cached at the next opportunity increases.

  In addition, when the inferior entry in use is shifted to stage 4 “Cached Out”, if there is an inquiry about the inferior entry from any RIP unit 120 thereafter, the cache management unit 130 responds with “Deleted” (see FIG. 7 S70). In this case, the inquiring RIP unit 120 performs RIP processing on the PDL data without using the cache data (S24 in FIG. 4). In other words, since the inferior entry is in use, even though the cache data is still in the cache memory 140, this embodiment prevents the RIP unit 120 that is the inquiry source from using it. In such a case, if the RIP unit 120 that is the inquiry source permits the use of cache data, under the basic policy that the cache data used by any RIP unit is not evicted, how long the entry that is reserved for evicting There is a possibility that it will not be able to be driven out even after a while. On the other hand, in this embodiment, since the use of the inferior entry reserved for eviction is not permitted, such a problem is avoided.

  Note that either the processing of S122 and S124 or the processing of S126 to S130 may be executed first.

  Although omitted in FIG. 11 to avoid complication, even if all the subordinate entries whose usage counters are not 0 are set to stage 4 “Cached Out”, there is a case where sufficient capacity cannot be secured to cache the target entry. Can happen. In such a case, the cache management unit 130 may end the process for the area securing request. Even if the necessary area could not be secured here, there was no problem because the allocation failure was responded to the current area allocation request, and eviction reservation was made for future queries for the same target entry. The possibility of successful cache registration is increased by the number of subordinate entries.

  Next, an example of a processing procedure of the cache management unit 130 when a cache data registration request is received from any of the RIP units 120 will be described with reference to FIG.

  In this procedure, as shown in FIG. 12, first, the cache management unit 130 specifies a cache entry to be notified from entry specifying information such as an ID included in the registration request, and stores cache data in the cache entry. Register (S142). For example, an item (not shown in FIG. 2) indicating whether or not cache data is registered is provided in the entry management table 132, and the value of the item may be changed from “not registered” to “registered” in S142.

  In addition, the cache management unit 130 resets the value of the usage counter of the cache entry to 0 (S144), and changes the entry from stage 1 “New Cache” to stage 3 “Cached” (S146). At this time, the item “RIP during data creation” of the entry is cleared.

  The example of the processing procedure of the cache management unit 130 for various accesses (inquiries, use start requests, use end notifications, area allocation requests, registration requests) from the RIP unit 120 has been described above. In the processing procedure described above, there is a case where one or more RIP units 120 make a eviction reservation (cache data is not deleted, but the stage is changed to 4 “Cached Out”) for a cache entry in use. is there. In the present embodiment, the cache entry for which the eviction reservation has been made is actually evicted as soon as possible when the cache data is no longer used by any of the RIP units 120, and the data until then is occupied. Free up space. There are several methods for releasing a cache entry that has been reserved for such eviction.

  One is a method of periodically checking the entry management table 132 to search for a cache entry to be released. An example of the procedure according to this method is shown in FIG.

  In the procedure of FIG. 13, the cache management unit 130 has a stage of 4 “Cached Out” and the value of the usage counter is 0 in the entry management table 132 when a periodic check timing (referred to as a cleaning timing) arrives. An entry is searched (S150). If there is such an entry (YES in S152), it is determined whether or not the cache data of the entry is still in the cache memory 140 (S154). If the determination result in S154 is YES, the cache data is evicted from the cache memory 140, the area occupied by the data so far is released (S156), and the process returns to S150 again. If the determination result in S154 is NO, S156 is skipped and the process returns to S150. In this way, the cache data of all cache entries whose stage is 4 “Cached Out” and whose usage counter value is 0 are evicted.

  Another possible method is to release the cache entry for which the eviction reservation has been made in the processing for the cache data use end notification. An example of the procedure according to this method is shown in FIG.

  The procedure in FIG. 14 is executed when a cache data use end notification is received from any of the RIP units 120. In this procedure, the cache management unit 130 identifies a cache entry corresponding to the ID or the like included in the use end notification from the entry management table 132, and decrements the use counter value of the entry by one (S160). Then, it is determined whether or not the entry is stage 4 “Cached Out” (S162) and the value of the usage counter is 0 (S164). If both the determination results are YES, The cache data of the entry is evicted from the cache memory 140 (S166). As a result, the area previously occupied by the cache data is released. If the identified entry is not stage 4 or the value of the usage counter is not 0, the cache data is not deleted and the process is terminated. The cache data of the cache entry for which the eviction reservation has been made is immediately evicted as soon as the RIP unit 120 being used disappears by this procedure.

  Next, with reference to FIG. 15, an example of a cache entry review process once shifted to the “Cached Out” stage will be described. This process is executed each time a review timing determined according to a predetermined rule is reached, for example, at regular intervals.

  In this procedure, when the review timing comes, the cache management unit 130 searches the entry management table 132 for a cache entry whose stage is 4 “Cached Out” (that is, in an evicted state) (S170), and so on. If there is a correct entry (the determination result in S172 is YES), it is determined whether or not the score of the entry is equal to or greater than a predetermined threshold (S174). As long as the cache entry score is normalized so as to fall within a certain range, the threshold value used for this determination may be a fixed value. Otherwise, the threshold value is adaptively changed according to the score distribution of the cache entries registered in the entry management table 132. There are various methods for determining the threshold value. For example, an average value of scores over all cache entries may be set as a threshold value, or a score corresponding to a predetermined rank as viewed from the top among scores of all cache entries may be set as a threshold value. Further, the threshold value may be set based on the lowest score among the cache entries whose current stage is 3 “Cached” (for example, a value obtained by adding a predetermined value to the lowest score is used as the threshold value).

  If it is determined in S174 that the score of the cache entry is greater than or equal to the threshold value, it means that the cache entry is worth returning to the cache memory 140. In this case, the cache management unit 130 transitions the cache entry from the stage 4 “Cached Out” to the stage 2 “Caching” (S176). As a result, when an inquiry is subsequently received from any RIP unit 120 for the cache entry, the entry transits to stage 1 “New Cache”, and the RIP unit 120 attempts to register the cache data of the entry. (See S66 in FIG. 7 and FIGS. 4 and 5).

<< Modification 1-1 >>
Thus, one embodiment has been described. In the embodiment of the first example described above, when cache data of a certain cache entry is being created (that is, stage 1 “New Cache”), the RIP unit 120 that inquired the entry completes the cache data. Whether or not to wait is determined (see S42 and S43 in FIG. 6). However, the determination may be made by the cache management unit 130 that has received the inquiry. An example in which the cache management unit 130 makes the determination will be described below.

  In this example, the processing procedure of the cache management unit 130 when receiving a cache entry inquiry from the RIP unit 120 is shown in FIGS. 16 and 17. The procedure shown in FIG. 16 is the same as the procedure shown in FIG. 7 except for the processing when the stage of the cache entry to be queried is 1 “New Cache”.

  When the stage of the cache entry to be queried is 1 “New Cache”, the cache management unit 130 executes the procedure illustrated in FIG. In this procedure, the cache management unit 130 determines whether the RIP unit 120 (referred to as the first RIP unit) as the inquiry source performs RIP processing on the PDL data of the object itself, and the RIP unit that is creating the cache data of the object. 120 (referred to as the second RIP unit) waits for completion of the cache data and evaluates which one completes the processing earlier when the first RIP unit uses the cache data (S182). . The evaluation method may be the same as the evaluation in S42 of FIG.

  In S183, it is determined based on the result of the evaluation in S182 whether it is better to wait for completion of the cache data being created or not. If it is determined in S183 that the first RIP unit should wait for completion of the cache data, the cache management unit 130 waits for a registration request for the cache data from the second RIP unit (S184). When the registration request comes, you can see that the cache data is complete. Then, the cache management unit 130 responds “Hit” to the first RIP unit (S186). At this time, since the cache data is in the cache memory 140, the first RIP unit that has received “Hit” advances the RIP processing of the page using the cache data.

  In other words, in this example, if it is determined that it is better to wait, the cache management unit 130 does not immediately respond to the inquiry RIP unit 120 (first RIP unit), but waits for completion of the cache data. Responds with “Hit”. Meanwhile, the RIP unit 120 as the inquiry source is waiting for a response from the cache management unit 130. However, such processing is only an example. Instead, when it is determined that it is better to wait, the RIP unit 120 as the inquiry source is instructed to wait immediately, and then “Hit” is returned when the cache data is completed. Good.

  On the other hand, if it is determined in S183 that it is better not to wait (that is, the process is completed faster if the first RIP unit performs the RIP process on the PDL data), the cache management unit 130 immediately determines the first A “Creating” response is returned to the RIP unit (S188). Upon receiving “Creating”, the first RIP unit RIP-processes the PDL data of the object without waiting for completion of the cache data, and advances the page drawing. In this case, the RIP processing result is not cached.

<< Modification 1-2 >>
Still another modification is also conceivable. That is, in the embodiment of the first example, when cache data of a certain cache entry is being created, it is determined whether the RIP unit 120 that inquired about the entry waits for the cache data to be completed or not. This determination has been made (see S42 and S43 in FIG. 6), but in this modification, this determination itself is canceled when the processing load of the RIP unit 120 is high. An example of the procedure of this modification is shown in FIG. FIG. 18 shows a flow of processing after it is determined that the response is “Creating” in S16 of the procedure of FIG.

  In this modification, in this case, first, the RIP unit 120 determines whether or not its own load is equal to or greater than a predetermined threshold (S41). The load here is the same as the load of the RIP unit 120 monitored by the load monitoring unit 114. When the load exceeds the threshold, executing the processing of S42 and S43 similar to the procedure of FIG. 6 itself becomes a burden that cannot be ignored for the RIP unit 120 (that is, preventing other processing being executed). Becomes). In other words, the threshold value is a lower limit load that causes such a situation, and this may be obtained in advance by an experiment or the like.

  If it is determined in S41 that the load on the RIP unit 120 is greater than or equal to the threshold, the RIP unit 120 skips S42 and S43, automatically determines that cache data is complete, and proceeds to S44. That is, in this case, since the RIP unit 120 has a high load, the PIP data is not subjected to RIP processing by itself. Other steps may be the same as in FIG.

<< Modification 1-3 >>
In the first example, when the RIP unit 120 receives a “Creating” response from the cache management unit 130, the RIP unit 120 determines whether to wait for completion of the cache data being created. If “Creating” is received without performing the process, the completion of the cache data being created may always be waited. Conversely, the RIP process may be performed without waiting and the processing result may not be cached. Even in such a case, the RIP unit 120 takes a different response from the conventional response of “Miss” or “Hit” by receiving a response of “Creating” indicating that the cache data is being created. In other words, if there are only responses “Miss” and “Hit”, when the response “Miss” is received, the RIP unit 120 does not know even if the target cache data is being created. It is impossible to wait, and the PDL data is subjected to RIP processing by itself, and the processing result is registered in the cache memory 140 in duplicate.

[Second example]
A second example of the embodiment of the present invention will be described. The configuration of the print document conversion system of the second example may be the same as that shown in FIG.

<< Outline of the second example >>
In the first example described above, when there is an area reservation request, the cache entry in use whose score is lower than that of the target entry of the request is shifted to stage 4 “Cached Out” to explicitly show the cache entry. The eviction reservation is made (S126 and S128 in FIG. 11). On the other hand, in this second example, the cache entry in use that is determined to be evicted remains in stage 3 “Cached” and is implicitly set in the evicting reservation state. In the first example, stage 4 “Cached Out” is also applied to a cache entry (an entry reserved for eviction) in which the cache data is in the cache memory 140, whereas in this second example, stage 4 “Cached Out”. Does not apply to such an entry, but applies only to entries whose cache data is actually evicted from the cache memory 140.

<< State transition of the second example >>
FIG. 19 shows the state transition of the cache entry in the second example. As shown in FIG. 19, in the second example, when any RIP unit 120 inquires about a cache entry of stage 3 “Cached”, the cache management unit 130 responds according to the score of the cache entry. Switch between “Hit” and “Deleted”. That is, if the score of the cache entry is equal to or greater than the threshold value, “Hit” is returned, otherwise “Deleted” is returned. The meaning of the response “Deleted” and the behavior of the RIP unit 120 when it is received are the same as in the first example. That is, in this second example, if the score is less than the threshold value, the use of the cache data of the cache entry is not permitted by responding “Deleted”. As a result, the cache data of the cache entry is not newly used, so even if the cache data is currently in use, the value of the usage counter decreases with time and can be evicted. (That is, the use counter value is 0). As described above, in the second example, when the score is lower than the threshold value, “Deleted” is responded to implicitly (that is, without explicitly shifting to “Cached Out”), and the cache entry is evicted. It is said. The threshold used here has a different purpose of use from the threshold used in S174 in the procedure of FIG. 15 of the first example, and basically may be determined independently of the threshold used in S174, but the process is simplified. For example, the same value as the threshold used in S174 may be used.

  The cache entry of stage 3 “Cached” transitions to stage 4 “Cached Out” when the cache data of that cache entry is actually evicted from the cache memory 140.

  The state transition of FIG. 19 is the same as the example of FIG. 3 except for the points described above.

<< Processing Procedure of Second Example >>
The processing procedure of each RIP unit 120 in the second example may be the same as that in the first example (see FIGS. 4 to 6).

  The operation of the cache management unit 130 in the second example is partially different from that in the first example. In the following, the description will be focused on parts different from the first example.

  First, an example of the processing procedure of the cache management unit 130 when an inquiry is received from any of the RIP units 120 is shown in FIG. In this example, when the stage of the cache entry to be queried is 3 “Cached”, the cache management unit 130 checks the score of the entry, and if the score is equal to or greater than the threshold, responds “Hit” and is less than the threshold If so, "Deleted" is returned (S68A). The other steps in FIG. 20 may be the same as the steps in FIG. 7 of the first example.

  The processing procedure of the cache management unit 130 when receiving a use start request from any of the RIP units 120 may be the same as the procedure of FIG. 8 of the first example.

  The processing procedure of the cache management unit 130 when receiving a use end notification from any of the RIP units 120 is, for example, as shown in FIG.

  In this procedure, the cache management unit 130 specifies the target cache entry from the ID included in the use end notification and decreases the use counter value of the entry by one (S160). Then, it is determined whether or not the value of the usage counter of the cache entry is 0 (S163) and whether or not the score of the cache entry is less than the threshold (S165). If both the determination results are YES, the cache data of the cache entry is evicted from the cache memory 140, and the stage of the entry is changed from the current value 3 “Cached” to 4 “Cached Out” (S167). As a result, the area previously occupied by the cache data is released. When the value of the usage counter of the target cache entry is not 0 or the score is equal to or greater than the threshold value, the cache data is not deleted and the stage is changed (S167), and the process is terminated. A cache entry whose score is less than the threshold value and is implicitly in the eviction reservation state (that is, a state in which new use is not accepted) is immediately evicted by this procedure as soon as the RIP unit 120 in use is lost, Transition to stage 4 “Cached Out” will occur.

  In the example of FIG. 21, the cache data of the cache entry whose score is lower than the threshold value is evicted at the timing when the use end notification is received, but this is only an example. Instead, for example, in the same manner as the procedure of FIG. 13 in the first example, the score and the use counter of each cache entry whose stage is “3” “Cached” are checked at each review timing, and the score is less than the threshold and the use counter The cache data of the entry whose value is 0 may be evicted.

  The processing procedure of the cache management unit 130 when an area reservation request is received from any of the RIP units 120 is as shown in FIGS. 22 and 23, for example.

  First, the procedure shown in FIG. 22 is almost the same as the processing procedure in the first example shown in FIG. 10, and only the point including S99 is different from that in the first example. That is, in this second example, when it is determined in S96 and S98 that the cache entry (object) for which the area is to be secured clears the restriction condition and can be accommodated in the free area of the cache memory 140 (the free capacity is sufficient). The threshold value of the score used in S165 of the procedure of FIG. 21 is reset to the lowest value (for example, 0) in the possible range (S99). The fact that the free space is determined to be sufficient in S98 means that at the time of this determination, the free area of the cache memory 140 is large enough not to require eviction. Therefore, in this example, in such a case, the selection criterion for the cache entry to be left in the cache memory 140 is lowered so that the cache data currently in the cache memory 140 is not expelled as much as possible.

  In the above example, the threshold is reset to the minimum value in S99, but this is only an example. Instead, in S99, the threshold value may be decreased by a predetermined width or ratio.

  Further, the processing procedure performed by the cache management unit 130 when it is determined that there is an inferior entry in S106 of the procedure of FIG. 22 is as shown in FIG. This processing procedure is substantially the same as the procedure of FIG. 11 in the first example, but the step group of S126 to S130 in the procedure of FIG. 11 is replaced with S125 in the procedure of FIG.

  If the necessary free space cannot be secured even after all the cache data of inferior entries that have not been used is evicted from the cache memory 140, the procedure of FIG. 23 is similar to the procedure of FIG. The stage of “target entry”) is changed to 2 “Caching” (S122), and a response indicating a failure to secure the area is sent to the RIP unit 120 of the request source (S124). That is, also in the second example, by returning a failure for the current area securing request, the RIP unit 120 of the request source generates a bitmap image by RIP processing the PDL data of the target entry, and does not cache it To. Then, when an inquiry is made for the same target entry after that, a reservation is made for the eviction of the cache data of the other entries that are currently in use so that there is enough free space to accommodate the cache data of the target entry. . The procedure in FIG. 23 is different from the procedure in FIG. 11 in the reservation method.

  In other words, in the procedure of FIG. 11, the eviction reservation is made by explicitly shifting the cache entry stage still used by any RIP unit 120 to 4 “Cached Out”. On the other hand, in the procedure of FIG. 23, such an explicit stage change is not performed, and the threshold value for determination of the eviction target in the eviction process (S165 and S167 in FIG. 21) is adjusted (S125), and the adjusted threshold value is adjusted. By preventing cache entries with lower scores from being used in the future (see S68A in FIG. 20), cache entries with lower scores can be easily evicted.

  The threshold adjustment in S125 is based on the score of the target entry. That is, here, in order to be able to cache the target entry, a score slightly lower than the score of the target entry is set as the threshold value. That is, in S125, for example, a result obtained by subtracting a predetermined value or ratio from the score of the target entry is set as the threshold value. More precisely, the threshold is set to a value lower than the score of the target entry and higher than the score of at least one subordinate entry. If the threshold value is set in this way, cache data of a cache entry having a score lower than the threshold value is evicted from the cache memory 140 by the procedure of FIG. 21 when it is no longer used by any RIP unit 120. Thereafter, when any RIP unit 120 makes an inquiry to the cache management unit 130 about the target entry, if the capacity larger than the cache data of the target entry can be secured by eviction based on the threshold value, the cache data of the target entry is stored in the cache memory. 140 will be registered. Since the score of the target entry is higher than the threshold value set in S125, the cache data of the target entry remains in the cache memory 140 without being evicted if no change is made to further increase the threshold value after the setting. Become.

  The processing procedure of the cache management unit 130 when a cache data registration request is received from any of the RIP units 120 may be the same as the procedure of FIG. 12 of the first example.

  Also in the second example, as in the first example, the cache entry review process (FIG. 15) once shifted to the “Cached Out” stage may be performed.

<< Modification 2-1) >>
The second example of the embodiment has been described above. In the above, the threshold value is used to determine the cache entry to be evicted (see S68A in FIG. 20 and S165 and S167 in FIG. 21). On the other hand, the same threshold value may be used to determine whether the cache entry requested to be cached deserves cache registration. An example of performing such a determination is shown in FIG.

  The procedure shown in FIG. 24 is a processing procedure of the cache management unit 130 when there is an area securing request from any RIP unit 120, and S103 is added to the procedure of FIG. In the procedure of FIG. 24, if the free space of the cache memory 140 is not sufficient, before extracting a subordinate entry for the target entry (S104), it is determined whether the score of the target entry is equal to or greater than a threshold (S103). If the score of the target entry is less than the threshold value, the target entry is regarded as not worthy to be cached in the first place, and the target entry is moved to stage 4 “Cached” without performing the steps such as extraction of the inferior entry (S104). Transition to “Out” (S108), a failure to secure the area is returned to the requesting RIP unit 120 (S110).

<< Modification 2-2 >>
In the above-described modified example 2-1, a threshold for determining a cache entry to be evicted (see S68A in FIG. 20 and S165 and S167 in FIG. 21) and a threshold used for determining whether or not to register a cache entry (FIG. 24). And S103). On the other hand, in this modified example 2-2, an example in which these two types of threshold values are used properly will be described. Hereinafter, the former threshold is referred to as an eviction threshold, and the latter threshold is referred to as a registration threshold.

  In this modified example 2-2, when an inquiry about a cache entry is received from any RIP unit 120, as shown in FIG. 25, if the score of the cache entry is equal to or greater than the “for eviction” threshold, “ “Hit” is returned, and if it is less than the “for eviction” threshold, “Deleted” is returned (S68B). The other steps of FIG. 25 may be the same as the procedure of the second example shown in FIG. That is, in this example, the threshold value for eviction is used among the two types of threshold values to determine the cache entry to be set in the implicit eviction reservation state in S68B.

  Further, the processing procedure of the cache management unit 130 when receiving a use end notification from any of the RIP units 120 is as shown in FIG. This procedure is the same as the procedure of FIG. 21 except that the score of the target entry is compared with the “for eviction” threshold (S165A). That is, in this example, the threshold for eviction is used among the two types of thresholds for determining whether or not the cache entry is evicted.

  Further, the processing procedure of the cache management unit 130 when an area securing request is received from any of the RIP units 120 is as shown in FIGS.

  The procedure of FIG. 27 corresponds to the procedure of FIG. 24 of the above-described modification 2-1. In the modified example 2-1 (FIG. 24), when determining whether or not the target entry is worth caching, the score of the target entry is compared with a common threshold (not divided for eviction and registration). On the other hand, in the modified example 2-2, the score of the target entry is compared with the “registration” threshold value (S103A). In the modified example 2-1 (FIG. 24), when it is determined in S98 that the free space is sufficient, the common threshold value is reset (S99). In the modified example 2-2, in this case, for registration. The threshold value is reset (S99A).

  The procedure of FIG. 28 corresponds to the procedure of FIG. 23 of the second example. In the second example (FIG. 23), in step S125, a common threshold value (not divided for eviction and registration) was calculated based on the score of the target entry, whereas in this modification 2-2 The threshold values for eviction and registration are calculated based on the score of the target entry (S125A). For example, the threshold values for eviction and registration are both set to a value lower than the score of the target entry and higher than the score of at least one inferior entry. Here, for example, each threshold value may be determined such that the registration threshold value is equal to or greater than the eviction threshold value. For example, a value obtained by subtracting a first positive value from the score of the target entry is used as a registration threshold, and a value obtained by subtracting a second positive value from the registration threshold is used as an eviction threshold.

  Also in this modified example 2-2, as with the procedure of FIG. 15 of the first example, the cache entry that has been once shifted to the “Cached Out” stage may be reviewed. In this modified example 2-2, as shown in FIG. 29, a registration threshold is used as a threshold for determining whether or not to return a cache entry in the “Cached Out” stage to a cache registration target ( S174A). Other steps of the procedure in FIG. 29 are the same as those in FIG.

[Third example]
"Overview"
A third example of the embodiment of the present invention will be described. The configuration of the print document conversion system of the third example may be the same as that shown in FIG.

  In the first and second examples described above, when an area reservation request is received from the RIP unit 120, if there is not enough free space, the inferior entry that is cached (stage 3 “Cached”) is evicted (deleted). As a result, the free space is increased (see S104 and S106 in FIG. 10 and FIG. 11). In contrast, in this third example, in addition to expelling the inferior entry whose stage is 3 “Cached”, the area secured by the inferior entry during the creation of the cache data (stage 1 “New Cache”) is also released. Increase free space.

  That is, as shown in FIG. 5, the RIP unit 120 that creates cache data first makes an area reservation request (S30) to secure an area for storing cache data, and then generates cache data in that area (S32). And registration (S36). Therefore, the cache entry of stage 1 “New Cache” has a period in which the area is reserved but the cache data is not yet completed. In the third example, the cache entry area of stage 1 “New Cache” corresponding to such a period is released. It should be noted that the cache entry of stage 1 “New Cache” is not subject to the processing of the third example because there is no area to be released for the area before securing the area.

  In the third example, the cache management unit 130 manages an item indicating whether or not the area of the cache entry has been secured in the entry management table 132 as shown in FIG. ing. This item only needs to be registered for the cache entry of stage 1 “New Cache”, and this item is a null value for cache entries of other stages.

<< State transition of the third example >>
The state transition of the cache entry in the third example is as shown in FIG. The state transition of FIG. 31 is almost the same as the state transition of the first example shown in FIG. 3 except that the condition for transition from stage 1 “New Cache” to stage 4 “Cached Out” is increased. It is. That is, in the third example, even if the area is forcibly released to secure the area of another cache entry after the area has been successfully secured, stage 1 “New Cache” to stage 4 “Cached Out” A transition to occurs.

  The processing procedure of each RIP unit 120 in the third example may be the same as that in the first example (see FIGS. 4 to 6).

  The operation of the cache management unit 130 in the third example has many parts similar to those in the first example. To explain the same point, the processing procedure of the cache management unit 130 when receiving an inquiry from any of the RIP units 120 may be the same as that in the first example (see FIG. 7). Further, the processing procedure of the cache management unit 130 when receiving a use start request from any of the RIP units 120 may be the same as in the first example (see FIG. 8). Further, the processing procedure of the cache management unit 130 when receiving a use end notification from any of the RIP units 120 may be the same as in the first example (see FIG. 9 or FIG. 14). Also, the processing when the cleaning timing comes and the review timing comes may be the same as in the first example (see FIGS. 13 and 15).

  In the following, the description will be focused on parts different from the first example.

  In the third example, the processing procedure of the cache management unit 130 when an area securing request arrives from any of the RIP units 120 is, for example, as shown in FIGS.

  First, the procedure shown in FIG. 32 is almost the same as the processing procedure in the first example shown in FIG. 10, and only S100A and S104A are different. In S100A, in addition to responding to the requesting RIP unit 120 to secure the area, the fact that the area of the cache entry has been secured is registered in the “area secured” column of the entry management table 132. Note that it is assumed that the “area reservation” column of the cache entry is initialized to “unsecured” when the entry transitions to stage 1 “New Cache”. In S104A, not only the stage 3 but also the inferior entry (entry having a lower score than the target entry) is searched from the cache entries in the stage 1 and area reserved. At this time, whether or not the area has been secured may be determined by referring to the “area secured” column of the entry management table 132.

  If it is determined in S106 of the procedure of FIG. 32 that there is an inferior entry, the cache management unit 130 proceeds to the procedure of FIG.

  In the procedure of FIG. 33, it is first determined whether or not there are any remaining subordinate entries that have a stage of 3 “Cached” and a usage counter value of 0 (S111A). If they remain, they are extracted one by one in descending order of score (S112A), the cache data of the extracted inferior entries is evicted (deleted) from the cache memory 140, and the inferior entries are extracted from the current stage 3 “Cached”. To stage 4 “Cached Out” (S114). While the inferior entry whose stage is 3 “Cached” and the usage counter value is 0 remains, the free capacity of the cache memory 140 is increased by repeating S112A and S114 (S116). When the free space necessary for caching the target entry can be secured by such a procedure (the determination result in S116 is YES), the cache management unit 130 stores the cache data from the free area of the cache memory 140. An area corresponding to the size is secured, the success of the area securing is responded to the requesting RIP unit 120, and the fact that the area of the cache entry has been secured is indicated in the "area securing" column of the entry management table 132 sign up. (S118A).

  Stage 3 If the inferior entry whose value is “Cached” and the usage counter is 0 is evicted from the cache memory 140 and the free space required to cache the target entry cannot be secured, the determination result in S111A is NO.

  When the determination result in S111A is NO, the cache management unit 130 attempts to secure sufficient free space by releasing the area of the cache entry in which the stage 1 “New Cache” and the area has been secured. There are three types of processing procedures for this purpose, for example. Hereinafter, examples of processing procedures of these three methods will be described in order.

<< First method: Temporary overflow method >>
In this method, the cache management unit 130 executes the processing procedure illustrated in FIG. In this procedure, first, it is determined whether or not an inferior entry whose area has been secured remains in stage 1 “New Cache” (S200). If it remains, one with the lowest score is extracted from the list (S202). Then, the extracted inferior entry is shifted from stage 1 “New Cache” to stage 4 “Cached Out” (S204), and the free capacity of the cache memory 140 is calculated when it is assumed that this area has been released. It is determined whether or not this free capacity is large enough to accommodate the cache data of the target entry (S206). If it is still insufficient, the process returns to S200 and another area that has been secured in stage 1 “New Cache”. While the inferior entry remains, the processes of S202 to S204 are repeated. If the free area necessary for the target entry can be secured by releasing the area of the subordinate entry group whose area has already been secured in stage 1 “New Cache” (the determination result in S206 is YES), the cache management unit 130 An area is reserved in the cache memory 140, and a code indicating the area reservation and information on the address of the reserved area are returned to the requesting RIP unit 120 (S208). At this time, the fact that the area has been secured is recorded in the “area secured” column of the record of the target entry in the entry management table 132. After S208, the cache management unit 130 ends the process for the area reservation request.

  Note that in S204, the inferior entry is only transitioned from stage 1 “New Cache” to stage 4 “Cached Out”, and the cache area reserved for the inferior entry is actually released. It is a point not to do. This release is performed when the creation of the cache data of the inferior entry currently created is completed and the created RIP unit 120 finishes using the data. In this way, in S204, only an area release reservation is made. For this reason, when S208 is executed, the free capacity of the cache memory 140 is actually smaller than the amount necessary for the target entry. For this reason, the area reservation in S208 is performed in a form that temporarily exceeds (overflows) the limit of the total capacity allocated to the cache memory 140 from the operating system. That is, the operating system of the computer that implements the cache management unit 130 and the cache memory 140 allocates a part of the memory space managed by the cache management unit 130 and the cache memory 140. In S208, the operating system of the target entry extends beyond the allocated area. A cache area is secured for this purpose. By such a securing process, the size of the area actually used as the cache memory 140 exceeds the originally allocated cache memory size, but this overflow state is the cache of the inferior entry group reserved for releasing the area. It is only a temporary state until data creation is completed.

  Note that even if the areas of all subordinate entries whose areas have been secured in stage 1 “New Cache” are released, if the area necessary for the target entry cannot be secured, the determination result in S200 is NO. In this case, the cache management unit 130 performs the same processing as S122 to S130 of FIG. 11 of the first example. That is, the stage of the target entry is changed to 2 “Caching” (S122), a failure to secure the area is responded to the requesting RIP unit 120 (S124), and the score is lower than the inferior entry in use in stage 3 In order, the eviction reservation (change to stage 4) is performed until the necessary free space is secured (S126 to S130).

  In this first method, the RIP unit 120 that requested the area reservation for the target entry releases the cache entry whose use counter value is 0 in stage 3 and the area of the cache entry whose area is already reserved in stage 1, When the necessary capacity is obtained, the cache area is immediately allocated.

  In the first method, the area release for the cache entry for which the area release reservation has been made in S204 is executed, for example, using a cache data registration request as a trigger. This procedure is shown in FIG.

  In the procedure of FIG. 35, when the cache management unit 130 receives a cache data registration request from any of the RIP units 120 (this request includes specific information such as an ID for specifying a cache entry), the cache management unit 130 The stage of the cache entry (target entry) that is the target of the request is obtained from the entry management table 132 (S140). In this case, the stage of the target entry is either 1 “New Cache” or 4 “Cached Out”. When the stage is 1 “New Cache”, the steps S142 to S146 are executed as in the procedure of FIG. 12 in the first example. When the stage is 4 “Cached Out”, the cache area of the target entry is reserved to be released in order to secure the cache area of other cache entries (see S204 in FIG. 34). Cache data is deleted, and the area occupied by the cache data until then is released (S148). Note that the RIP unit 120 that has created the cache data of the target entry before this release has already completed the process of combining the cache data with the print image data of the page (the registration request is performed in the procedure of FIG. 5). Since bitmap creation (S34 is executed before (S36)), even if the cache data is deleted at this time, the RIP unit 120 is not adversely affected.

  By such an area release process, the overflow state of the cache memory 140 is eliminated.

<< Second method: Immediate forced release method >>
In this method, the cache management unit 130 executes the processing procedure illustrated in FIG. In this procedure, first, it is determined whether or not an inferior entry whose area has been secured remains in stage 1 “New Cache” (S200). If it remains, one with the lowest score is extracted from the list (S202). Then, for the RIP unit 120 that is creating the cache data of the extracted inferior entry (this is known from the entry management table 132), the creation of the object is canceled using the normal memory area instead of the cache area. An instruction is given to RIP process the original PDL data (S203). In response to this, the RIP processing unit 120 performs RIP processing of the object, combines the RIP processing result with the image of the page, and discards it without caching. Further, the extracted inferior entry is shifted from stage 1 “New Cache” to stage 4 “Cached Out”, and the cache area reserved for this inferior entry is released (S204A).

  That is, in S204 of the first method described above, the cache area is not released only by changing the stage of the inferior entry to 4, whereas in the second method, the cache area is released. After the release, it is determined whether or not the free capacity of the cache memory 140 is large enough to accommodate the cache data of the target entry (S206). While another subordinate entry whose area has been secured remains in “New Cache”, the processing of S202 to S204A is repeated. If the free area necessary for the target entry can be secured by releasing the area of the inferior entry group in which the stage 1 “New Cache” and the area has already been secured (the determination result in S206 is YES), the cache management unit 130 An area to be reserved is secured in the cache memory 140, and a code indicating that the area is secured and information on the address of the secured area are returned to the RIP unit 120 that requested the request (S208). At this time, the fact that the area has been secured is recorded in the “area secured” column of the record of the target entry in the entry management table 132. After S208, the cache management unit 130 ends the process for the area reservation request.

  Note that even if the areas of all subordinate entries whose areas have been secured in stage 1 “New Cache” are released, if the area necessary for the target entry cannot be secured, the determination result in S200 is NO. In this case, the cache management unit 130 performs the same processing as S122 to S130 of FIG. 11 of the first example.

  In this second method, the RIP unit 120 whose cache area has been forcibly released needs to perform RIP processing of the PDL data again, but the capacity allocated to the cache memory 140 does not exceed, and the cache required for the target entry Space is reserved immediately. Note that the inferior entry in stage 1 whose cache area has been forcibly released has a lower score than the target entry, so it is better to release the cache data area in which the inferior entry is being created than to give up the cache of the target entry. This is advantageous in terms of utilization efficiency.

<< Third method: Release method after completion of creation >>
In this method, the cache management unit 130 executes the processing procedure illustrated in FIG. This procedure is the same as that of the first method shown in FIG. 34 from S200 to S206. That is, in step S204, the stage of the inferior entry whose area has been secured in stage 1 “New Cache” is only changed to 4, and the cache area is not released.

  In the first method described above, the area for the target entry is secured (S208) without overflowing the capacity allocated to the cache memory 140 without waiting for the area release of the inferior entry of stage 1 that has been transitioned to stage 4 went. On the other hand, in this third method, without overflowing, wait for the release of the areas of all the subordinate entries in stage 1 that have been transitioned to stage 4 (S207), and after the areas of all the subordinate entries are released, An area for the target entry is reserved (S208). Here, in S207, for example, for each subordinate entry in stage 1 that has been transitioned to stage 4, the “area reservation” column (see FIG. 30) of the entry management table 132 is monitored, and all the subordinate entries to be monitored are monitored. When the “area reservation” column has a value indicating “released”, the determination result in S207 is YES. That is, in this method, the values that can be taken in the “area reservation” column are three values of “unsecured”, “secured”, and “released”. It should be noted that the value of “released” is written in the “area reservation” column of each inferior entry, as will be described later, during the processing for the registration request. Other steps may be the same as the procedure in FIG. 34 of the first method.

  In this third method, when there is a cache data registration request from any of the RIP units 120, the cache management unit 130 executes the processing shown in FIG. This procedure is obtained by adding S149 to the procedure of the first method shown in FIG. That is, in the third example, after the cache management unit 130 deletes the cache data targeted for registration from the cache memory 140 and releases the area in S148, for example, “reserve area” of the entry in the entry management table 132, for example. Is changed to a value “reserved” indicating that the area of the entry has been released.

  In the third method, the cache area necessary for the target entry is not immediately secured, but the capacity allocated to the cache memory 140 is not exceeded.

《Mixing method》
The three methods have been described above. The cache management unit 130 may execute the process according to one of these three methods, but may advance the process while dynamically switching these three methods. An example of such a dynamic switching procedure is illustrated in FIGS.

  The procedure in FIG. 39 is a procedure when the determination result of S106 in the procedure in FIG. 32 is YES, and S300 and S302 are added to the procedure in FIG. Therefore, description of the steps already described with reference to FIG. 33 is omitted.

  In the procedure of FIG. 39, if the determination result in S111A is NO, that is, if the cache area necessary for the target entry cannot be secured even if all unused cache data is deleted from the cache memory, first the cache management unit 130 inquires of the operating system (OS) about the free capacity (or usage rate, etc.) of the memory managed by the OS (S300). The cache management unit 130 has a part of the memory managed by the OS allocated as the cache memory 140, but in S300, the cache memory 140 temporarily exceeds the allocated memory space (overflows). In order to determine whether or not it can be expanded, the free space of the memory is inquired. Then, the cache management unit 130 determines whether the memory managed by the OS has a room for increasing the allocation for the cache memory 140 by the amount of the cache area for the target entry based on the free memory capacity obtained as a result. Is determined (S302). In this determination, for example, if the cache data size of the target entry is subtracted from the free space of the memory managed by the OS, and the subtraction result exceeds the threshold of the free space of the memory for the OS to operate properly, For example, it may be determined that “there is room”.

  If it is determined in S302 that there is a margin in the memory, the cache management unit 130 executes processing of the first (overflow) method shown in FIG.

  If it is determined in S302 that there is no room in the memory, the cache management unit 130 selects the second (immediate forced release) method or the third (release after completion of creation) method. An example of the procedure for this is shown in FIG. The procedure in FIG. 40 is similar to the procedure in FIG. 37 (third method). In the procedure of FIG. 40, it is first determined whether or not an inferior entry whose area has been secured remains in stage 1 “New Cache” (S200). When the determination result in S200 is NO, the processes of S122 to S130 are executed as in the procedure of FIG.

  If the determination result in S200 is YES, the one with the lowest score is extracted from the inferior entries whose areas have been secured in the unprocessed stage 1 “New Cache” (S202). Then, the extracted inferior entry is shifted from stage 1 “New Cache” to stage 4 “Cached Out” (S204), and the free capacity of the cache memory 140 is calculated when it is assumed that this area has been released. Also, a penalty is assumed when it is assumed that the cache area of the extracted inferior entry is forcibly released immediately (S304). The “penalty” here is an evaluation value that represents the damage suffered by each RIP unit 120 that has previously created cache data of the subordinate entries when the cache area of each subordinate entry extracted is immediately released. .

  Various methods can be considered for the penalty evaluation in S304. For example, as one method, the size of cache data of an inferior entry to be evaluated (this cache data is being created and not yet completed, but the size of the cache data is calculated or estimated from the PDL description, and the entry management table There is a method of evaluating based on (registered in No. 132). In this method, the larger the cache data size, the higher the penalty when the cache data area is immediately released. This is because large cache data is expensive to create again.

  In another method, evaluation is performed based on the completion rate of creation of cache data of an inferior entry to be evaluated (that is, the ratio of the portion of the cache data that has been created). In this method, the higher the cache data creation completion rate, the higher the penalty when the cache data area is immediately released. This is because the processing for cache data creation up to that point is wasted due to the area release, but the waste is increased as the creation completion rate is higher. For this method, the cache data creation completion rate of each cache entry is notified, for example, periodically to the cache management unit 130 from the RIP unit 120 that creates the cache data. The cache management unit 130 may register and manage the notified creation completion rate in the entry management table 132 or the like.

  In yet another method, evaluation is performed based on a predicted value (creation completion predicted time) of completion of creation of cache data of an inferior entry to be evaluated. In this method, the closer the cache data creation completion prediction time is to the current time, the higher the penalty is when the cache data area is immediately released. This is because the closer the creation completion prediction time is to the current time, the shorter the time for waiting for the creation completion for the target entry, so there is less loss if the area is not released immediately (that is, the penalty is lower). For this method, the RIP unit 120 that creates the cache data of each cache entry, for example, periodically notifies the cache management unit 130 of the creation completion rate, and the cache management unit 130 is notified each time. What is necessary is just to obtain | require creation completion estimated time from the time change of a creation completion rate. The calculated creation completion prediction time may be registered and managed in the entry management table 132 or the like.

  Further, it is determined whether or not the free space calculated in S204 is large enough to accommodate the cache data of the target entry (S206). If it is still insufficient, the process returns to S200, and in stage 1 “New Cache” While another subordinate entry whose area has been secured remains, the processes of S202, S204, and S304 are repeated. When the determination result in S206 is YES due to such repetition, the cache management unit 130 determines that the penalty value calculated in S304 is smaller than a predetermined threshold value for all the inferior entries to be subject to area release extracted in S202. Is determined (S306). If the penalty for a subordinate entry is smaller than the threshold value, even if the cache area reserved for the subordinate entry is immediately released, the disadvantage caused by the release (for example, RIP that created cache data in the released area) There is a high possibility that the cost due to the RIP processing of the PDL data again by the unit 120 is more than the benefit of immediately securing the cache entry for the target entry. In other words, such a threshold value is obtained in advance through experiments and simulations.

  If the penalty of the inferior entry subject to area release is smaller than the threshold value in the determination in S306, it means that there is little disadvantage even if the cache area is immediately released for any inferior entry. In this case, the cache management unit 130 notifies the RIP unit 120 that is creating the cache data of each subordinate entry to be released, that the area is to be released, and releases the cache area reserved for each subordinate entry. (S308). This is the second (immediate release) method. In step S208, an area for the target entry is secured in the cache memory 140, and a code indicating that the area is secured and information on the address of the secured area are returned to the requesting RIP unit 120.

  On the other hand, if it is determined in S306 that there is at least one inferior entry subject to area release with a penalty equal to or greater than the threshold, there is an inferior entry that has many disadvantages if the cache area is immediately released. In this case, the third (release after completion of creation) method is adopted. In other words, the cache management unit 130 waits for the cache areas of all subordinate entries to be area released to be actually released (S207), and after all the subordinate entry areas are released, An area is secured (S208).

  In the example of FIGS. 39 and 40, the three methods are dynamically switched and used, but it is not necessary to use all three methods. As another example, two of the three schemes may be dynamically switched and used. For example, when the first (overflow) method and the second (immediate forced release) method are dynamically switched, it is only necessary to determine which method is used according to the steps S300 and S302 in FIG. The same applies when the first method and the third (release after completion of creation) method are dynamically switched. In addition, when dynamically switching between the second method and the third method, it is only necessary to determine which method is to be used in steps S304 and S306 in FIG.

  In the above example, in order to switch between the second method and the third method, the criterion of whether or not all penalty values are smaller than the threshold value in S306 is used, but this is only an example. Instead, if the ratio of inferior entries whose penalty value is equal to or greater than a threshold is equal to or less than a predetermined ratio threshold, another determination criterion such as adopting the second (immediate forced release) method may be considered.

  In addition, although the 3rd example demonstrated above was demonstrated in the form based on the above-mentioned 1st example, the example based on the above-mentioned 2nd example is also considered. In other words, the difference between the first example and the second example is that the reservation for eviction of cache entries (deletion of cache data) whose stage is 3 “Cached” and whose use counter value is not 0 is explicitly (first example). This is a difference between whether to perform or implicitly (second example), and this difference is dependent on the process of releasing the area of the cache entry in the stage 1 “New Cache” in the third example. Not. When the area securing method of the third example is applied to the second example, for example, the step group of S122 to S130 in the procedure of FIG. 34 may be replaced with S122, S124, and S125 of FIG.

[Fourth example]
"Overview"
A fourth example of the embodiment of the present invention will be described. The configuration of the print document conversion system of the fourth example may be the same as that shown in FIG. The data structure of the entry management table 132 may be the same as that of the first example illustrated in FIG. Further, the state transition of the cache entry may be the same as the state transition of the first example illustrated in FIG.

  In the above-described first example, in the area securing process in response to the area securing request, the necessary free capacity cannot be secured even if all the cache data of the subordinate entries whose stage is 3 “Cached” and whose usage counter value is 0 are evicted. In this case, the allocation of the area at that time is given up, and a failure to allocate the area is returned to the requesting RIP unit 120 (for example, S124 in FIG. 11).

  On the other hand, in this fourth example, in such a case, if there are inferior entries whose use counter value is not 0 (that is, being used by any RIP unit 120) in stage 3, the cache of these inferior entries is cached. Secure the area by expelling (deleting) data.

  The processing procedure of each RIP unit 120 in the fourth example may be the same as that in the first example (see FIGS. 4 to 6).

  Regarding the operation of the cache management unit 130 in the fourth example, the processing procedure of the cache management unit 130 when receiving an inquiry from any of the RIP units 120 may be the same as in the first example (see FIG. 7). Further, the processing procedure of the cache management unit 130 when receiving a use start request from any of the RIP units 120 may be the same as in the first example (see FIG. 8). Also, the processing when the cleaning timing comes and the review timing comes may be the same as in the first example (see FIGS. 13 and 15).

  On the other hand, the operation of the cache management unit 130 when receiving an area reservation request from any of the RIP units 120 is the case of the first example (see FIGS. 10 and 11) and the second example (see FIGS. 22 and 23). And different. In particular, in the area allocation process in response to the area allocation request, the necessary free capacity cannot be secured even if all the cache data of the inferior entry whose stage is 3 “Cached” and the usage counter value is 0 is removed (in the first example) When the determination result of S111 in FIG. 11 is NO, in the second example, the processing in the case of the determination result of S111 in FIG. 23 being NO) is different from the case of the first example and the second example.

  This fourth example in the case where the necessary free capacity cannot be secured even if all the cache data of the inferior entry whose stage is 3 “Cached” and the value of the use counter is 0 is expelled during the area securing process in response to the area securing request For example, there are the following three methods. Hereinafter, examples of processing procedures of these three methods will be described in order.

<< First method: Temporary overflow method >>
In the first method of the fourth example, when an area reservation request is received from one of the RIP units 120, the cache management unit 130 performs the same operation as the procedure of the first example shown in FIG. If it is determined in S106 that there is an inferior entry during the execution of this procedure, in this first method, the cache management unit 130 executes the procedure shown in FIG. 41, steps S111 to S118 and S126 to S130 are the same as those in the first example shown in FIG.

  The procedure shown in FIG. 41 is different from the procedure shown in FIG. 11. First, in S111, after it is determined that there are no inferior entries whose stage is 3 “Cached” and whose use counter value is 0, the processing in S122 and S124 (target entry) Are not transferred to stage 2 “Caching” and a response to failure of securing). Another difference is that if it is determined in S130 that the necessary capacity has been secured, the cache management unit 130 secures an area for the target entry in the cache memory 140, and secures the area in the requesting RIP unit 120. This is a point that returns a code indicating this and information on the address of the secured area (S132).

  That is, in S128, only the stage of the inferior entry being used is changed to 4 “Cached Out”, and the cache data of the inferior entry being used is not purged from the cache memory 140. Even if it is determined in S130 that the necessary free space has been secured in the cache memory 140, it is a case where it is assumed that the cache data of the inferior entries being used is evicted. For this reason, when S132 is executed, the free capacity of the cache memory 140 is actually smaller than the amount necessary for the target entry. Therefore, what is performed in S132 is that the capacity of inferior entries that have not yet been expelled (subordinate entries reserved for expulsion) is expected to be expelled in the near future when the use ends. It is to secure an area for entry. For this reason, the area reservation in S132 is performed in a manner that temporarily exceeds (overflows) the limit of the total capacity allocated to the cache memory 140 from the operating system. That is, the operating system of the computer that implements the cache management unit 130 and the cache memory 140 allocates a part of the memory space managed by the cache management unit 130 and the cache memory 140. A cache area is secured for this purpose. By such a securing process, the size of the area actually used as the cache memory 140 exceeds the originally allocated cache memory size, but this overflow state is caused by the cache data of the inferior entry group reserved for eviction. It is only a temporary state until creation is completed.

  Although omitted in FIG. 41 to avoid complication, even if all of the inferior entries whose usage counters are not 0 are set to stage 4 “Cached Out”, there may be a case where sufficient capacity cannot be secured to cache the target entry. . In such a case, for example, the cache management unit 130 may respond with an allocation failure to the area allocation request.

  In this first method, the RIP unit 120 that requested the area reservation for the target entry immediately receives the cache area allocation if the necessary capacity can be obtained by releasing the area of the inferior entry in stage 3. It will be.

  In the first method, the area for the cache entry reserved for eviction in S128 may be released by the procedure illustrated in FIG. 13 or FIG. 14 as in the first example. The cache data of the inferior entry reserved for eviction is actually evicted from the cache memory 140 according to the procedure of FIG. 13 or FIG.

<< Second method: Immediate forced eviction method >>
The second method of the fourth example is the same as the first method described above in that an area for the target entry is immediately secured. However, in the first method, the area for the target entry is secured so as to overflow the capacity allocated to the cache memory 140, so that the inferior entry in use is not actually expelled (reservation for eviction is made). On the other hand, in this second method, in order to avoid such an overflow, an inferior entry in use is actually evicted and an area for the target entry is secured. Here, when the inferior entry in use is actually evicted, the RIP unit 120 that uses the cache data of the inferior entry cannot continue the RIP process. The RIP unit 120 that is using the cache data is notified that the cache data is to be deleted. In response to this, the RIP unit 120 performs RIP processing on the original PDL data of the object corresponding to the cache data to be deleted.

  For this process, in the second method, for example, “RIP in use” is recorded in the entry management table 132 as illustrated in FIG. Identification of the RIP unit 120 that uses the cache data of the cache entry (that is, in a period from when the use start request is issued to the cache data until the use end notification is issued) for “RIP in use” Information is registered. If there are a plurality of RIP units 120 using the cache data of one cache entry, the identification information of all the plurality of RIP units 120 used is registered. For cache entries in which cache data is not in the cache memory 140 (entries at stages other than stage 3 “Cached”), “RIP in use” is blank. In the period when the cache entry whose stage is 3 “Cached” is not used by any RIP unit 120, “RIP in use” is blank.

  For the maintenance of the “RIP in use” column, when a cache data use start request arrives from any RIP unit 120, the cache management unit 130 registers the cache data as illustrated in FIG. In S80, the value of the usage counter is incremented by 1 (S82), and the identification information of the requesting RIP unit 120 is added to the "RIP in use" column of the cache entry (S83). Further, when a cache data use end notification arrives from any of the RIP units 120, the cache management unit 130 decreases the value of the use counter by 1 as illustrated in FIG. 44 (S85), and the RIP of the notification source The identification information of the section 120 is deleted from the “in-use RIP” column of the cache entry (S87).

  In the second method of the fourth example, when an area reservation request is received from any of the RIP units 120, the cache management unit 130 performs the same operation as the procedure of the first example shown in FIG. If it is determined in S106 that there is an inferior entry during execution of this procedure, in this second method, the cache management unit 130 executes the procedure shown in FIG. The procedure of FIG. 45 is the same as the procedure (FIG. 41) in the case of the first method described above, except for S127, S128A, and S132A.

  That is, in the procedure of FIG. 45, after it is determined in S111 that there are no inferior entries whose stage is 3 “Cached” and whose usage counter value is 0, the processing in S122 and S124 (the target entry is changed to stage 2 “Caching”). In addition, the procedure is the same as that of the first method (FIG. 41) in that the reservation failure is not performed. The procedure of FIG. 45 differs from the procedure of FIG. 41 in that each SIP unit 120 registered in the “RIP in use” column of the cache entry extracted in S127 after S126 has cache data for that entry. In addition to changing the stage of the entry to 4 in S128A, the cache data of the entry is actually deleted from the cache memory 140, instructing to stop using (and re-RIP processing from the original PDL data) However, the area of the data is released.

  As described above, in this second method, the cache data is actually deleted and the area is released, so in S132A in the procedure of FIG. 45, the cache area for the target entry may be secured from those free areas. . Therefore, it is not necessary to secure a cache area beyond the range allocated as the cache memory 140 by the operating system. In the second method, a cache area necessary for the target entry is immediately secured. Since the inferior entry in stage 1 whose cache area has been forcibly released has a lower score than the target entry, it is better to release the cache data area in use of the inferior entry than to give up the cache of the target entry. In many cases, it is advantageous in terms of utilization efficiency.

《Third method: Ejection method after completion of use》
In the first method described above, the area for the target entry is overflowed from the capacity allocated to the cache memory 140 without waiting for the data to be deleted after the use of the subordinate entry shifted to the stage 4 is finished. Securement (S132) was performed. On the other hand, in the third method, as shown in FIG. 46, the overflow of the cache memory 140 waits for the end of use of all the subordinate entries shifted to the stage 4 (S131), and then all the subordinates. After the entry area is released, an area for the target entry is reserved (S132A). Here, in S131, each inferior entry whose stage is changed to 4 “Cached Out” in the repetition of S126 to S130 is stored, and the usage counter of each inferior entry is periodically monitored, for example. When the values of all the usage counters of the entries become 0, it is determined that the use of these inferior entries has been completed.

  46, the subordinate entry stage is changed from 3 “Cached” to 4 “Cached Out” in S128, but the cache data is not deleted unlike S128A of the second method.

  In this third method, the cache area required for the target entry is not secured until the use of the subordinate entry group changed to stage 4 is completed, but does not exceed the capacity allocated to the cache memory 140.

《Mixing method》
The three methods have been described above. The cache management unit 130 may execute the process according to one of these three methods, but may advance the process while dynamically switching these three methods. An example of such a dynamic switching procedure is illustrated in FIGS. 47 and 48.

  The procedure of FIG. 47 is a procedure when the determination result of S106 of the procedure of FIG. 10 is YES, and S320 and S322 are added to the procedure of FIG. Therefore, the description of the steps already described with reference to FIG. 41 is omitted.

  In the procedure of FIG. 41, if the determination result in S111 is NO, that is, if the cache area required for the target entry cannot be secured even if all unused cache data is deleted from the cache memory, first the cache management unit 130 inquires of the operating system (OS) about the free capacity (or usage rate, etc.) of the memory managed by the OS (S320). The cache management unit 130 allocates a part of the memory managed by the OS as the cache memory 140. In S320, the cache management unit 130 temporarily exceeds the allocated memory space (overflows) and temporarily allocates the area of the cache memory 140. In order to determine whether or not it can be expanded, the free space of the memory is inquired. Then, the cache management unit 130 determines whether the memory managed by the OS has a room for increasing the allocation for the cache memory 140 by the amount of the cache area for the target entry based on the free memory capacity obtained as a result. Is determined (S322). In this determination, for example, if the cache data size of the target entry is subtracted from the free space of the memory managed by the OS, and the subtraction result exceeds the threshold of the free space of the memory for the OS to operate properly, For example, it may be determined that “there is room”. If it is determined in S322 that there is a margin in the memory, the cache management unit 130 executes the first (overflow) method processing shown in S126 to S132.

  If it is determined in S322 that there is no room in the memory, the cache management unit 130 selects the second (immediate forced eviction) method and the third (ejection after use completion) method. An example of the procedure for this is shown in FIG. In the procedure of FIG. 48, one with the lowest score is extracted from the subordinate entries (stage 3) (S126). Then, the extracted inferior entry is transitioned to stage 4 “Cached Out” (S128), and the free capacity of the cache memory 140 is calculated when it is assumed that the cache data of the inferior entry is expelled and the cache area is released. Also, a penalty is assumed when it is assumed that the cache data of the inferior entry is immediately forced out (S324). The “penalty” here is an evaluation value representing the damage suffered by each RIP unit 120 that has been using the cache data of the inferior entries until the extracted cache data of the inferior entries immediately. .

  Various methods can be considered for the penalty evaluation in S324. For example, as one method, there is a method of evaluating based on the size of cache data of an inferior entry to be evaluated. In this method, the larger the cache data size, the higher the penalty for eviction of the cache data. In the case of the immediate forced eviction method, the cache data being used is evicted, so that the RIP unit 120 that is using the cache data needs to perform RIP processing on the PDL data corresponding to the cache data. . In general, the larger the cache data, the longer the RIP processing takes, so the penalty is increased.

  In another method, for each RIP unit 120 that uses the cache data of the inferior entry to be evaluated, the read completion rate of the cache data (that is, the ratio of the portion of the cache data that has been read) Based on the evaluation. In this method, the higher the cache data read completion rate, the higher the penalty for immediately expelling the cache data. This is because the cache data read processing up to that point is wasted due to the eviction, but the higher the read completion rate, the greater the waste. When there are a plurality of RIP units 120 using the cache data of one inferior entry, a penalty is obtained for each RIP unit 120. Note that for this method, the cache management unit 130, for each RIP unit 120, the cache entry that currently provides cache data to the RIP unit 120 and the ratio of the cache data that has been provided (ie, the RIP unit). For example, the read completion rate from 120 is periodically recorded.

  In yet another method, evaluation is performed based on a predicted value (use completion prediction time) of a time at which use of cache data of an inferior entry to be evaluated is completed. When there are a plurality of RIP units 120 that use cache data of one subordinate entry, the use completion prediction time is obtained for each RIP unit 120, and the penalty value is obtained from the use completion prediction time. In this method, the closer the cache data usage completion prediction time is to the current time, the higher the penalty for immediately evicting the cache data. This is because, as the predicted use completion time is closer to the current time, the time for waiting for completion of the use for the target entry is shorter, so there is less loss if not immediately expelled (ie, the penalty is lower). Note that for this method, the cache management unit 130, for each RIP unit 120, the cache entry that currently provides cache data to the RIP unit 120, and the proportion of the cache data that has been provided (ie, RIP For example, the read completion rate) may be periodically recorded, and the read completion prediction time may be obtained from the temporal change in the read completion rate.

  Further, it is determined whether or not the free space calculated in S128 is large enough to accommodate the cache data of the target entry (S130). If it is still insufficient, the process returns to S126, and another subordinate entry with a low score is obtained. While S remains, the processes of S126, S128, and S324 are repeated. If the determination result in S130 is YES due to such repetition, the cache management unit 130 determines whether or not the penalty value calculated in S324 is smaller than a predetermined threshold value for all the inferior entries to be evicted extracted in S126. Is determined (S326). When the penalty for a certain subordinate entry is smaller than the threshold value, even if the cache data of the subordinate entry is immediately evicted, the RIP unit 120 that has used the evicted cache data again converts the PDL data to RIP. The cost of processing) is likely to be less than the benefit of immediately securing a cache entry for the target entry. In other words, such a threshold value is obtained in advance through experiments and simulations.

  If the penalty of the inferior entry subject to area release is smaller than the threshold value in the determination of S326, it means that there is little disadvantage even if the cache data is immediately evicted for any inferior entry. In this case, the immediate forced eviction method is adopted. That is, the cache management unit 130 notifies each RIP unit 120 that is using the cache data of each inferior entry to be evicted to stop using the data, evicts the cache data of each inferior entry, The cache area previously occupied by the cache data is released (S328). In step S132, an area for the target entry is secured in the cache memory 140, and a code indicating that the area is secured and information on the address of the secured area are returned to the requesting RIP unit 120.

  On the other hand, if it is determined in S326 that there is at least one inferior entry to be evicted that has a penalty equal to or greater than the threshold value, there is an inferior entry with many disadvantages if the cache data is immediately evicted. In this case, the third (purge after use completion) method is adopted. That is, the cache management unit 130 waits for all the RIP units 120 using the cache data of all the subordinate entries to be evicted to complete the use (S131), and the areas of all the subordinate entries are released. After that, an area for the target entry is secured (S132).

  In the above example, in order to switch between the second method and the third method, the determination criterion whether or not all penalty values are smaller than the threshold value in S326 is used, but this is only an example. Instead, if the ratio of inferior entries whose penalty value is equal to or greater than the threshold is equal to or less than a predetermined ratio threshold, other determination criteria such as adopting the second (immediate forced eviction) method may be considered.

  In the example of FIGS. 47 and 48, the three methods are dynamically switched and used, but it is not necessary to use all three methods. As another example, two of the three schemes may be dynamically switched and used. For example, when the first (overflow) method and the second (immediate forced eviction) method are dynamically switched, it is only necessary to determine which method is used in steps S320 and S322 in FIG. The same applies to the case where the first method and the third (purge after use completion) method are dynamically switched. In addition, when the second method and the third method are dynamically switched, it is only necessary to determine which method is to be used in steps S324 and S326 in FIG.

  The print document conversion system 100 exemplified above, or the job management unit 110, the RIP unit 120, and the cache management unit 130 constituting the print document conversion system 100, for example, causes a general-purpose computer to execute a program representing the processing of each functional module described above. It is realized by. Here, the computer includes, as hardware, a microprocessor such as a CPU, a memory (primary storage) such as a random access memory (RAM) and a read only memory (ROM), an HDD controller that controls an HDD (hard disk drive), Various I / O (input / output) interfaces, network interfaces that perform control for connection to a network such as a local area network, and the like have a circuit configuration connected via a bus, for example. Also, portable non-volatile recording of various standards such as a disk drive and a flash memory for reading and / or writing to a portable disk recording medium such as a CD or a DVD via the I / O interface, for example. A memory reader / writer for reading from and / or writing to a medium may be connected. A program in which the processing contents of each functional module exemplified above are described is stored in a fixed storage device such as a hard disk drive via a recording medium such as a CD or DVD, or via a communication means such as a network, and stored in a computer. Installed. The program stored in the fixed storage device is read into the RAM and executed by a microprocessor such as a CPU, thereby realizing the functional module group exemplified above.

  100 print document conversion system, 110 job management unit, 112 page allocation unit, 114 load monitoring unit, 120 RIP unit, 130 cache management unit, 132 entry management table, 140 cache memory, 200 host computer, 210 print control device, 220 printer engine.

Claims (16)

A cache device and a plurality of data processing devices;
The cache device is:
A cache memory for storing image data in bitmap format or intermediate language format created by each of the data processing devices;
A state storage unit that stores, for each document element of the print document data, one of a plurality of predefined states of the cache state of the image data of the document element in the cache memory, A plurality of states include a first state indicating that no image data is stored in the cache memory even though it is worth storing the image data in the cache memory, and storing the image data in the cache memory. State storage means comprising: a second state indicating that image data is not in the cache memory because of no value; and
When an inquiry about image data of a document element is received from the data processing device, a first response is obtained when the cache state of the image data of the document element stored in the state storage means is the first state. Response providing means for providing a second response to the data processing device when the cache state of the image data of the document element is the second state;
With
Each of the plurality of data processing devices
Image data creating means for processing page description language data of each document element in the print document data to create image data in a bitmap format or an intermediate language format;
For each document element in the print document data, before causing the image data creation means to create the image data of the document element, an inquiry means that inquires whether the image data is in the cache device;
If the response from the cache device to the inquiry by the inquiry means is the first response, the image data creation means executes creation of image data of the document element, and the image data created as a result When the image for printing is generated using, processing for registering the image data in the cache device is executed, and the response from the cache device to the inquiry by the inquiry means is the second response The image data creation means executes creation of image data of the document element, generates a print image using the image data created as a result, and the image data is not registered in the cache device. Control means for performing control of discarding;
Comprising
A printed document processing system. When the response providing unit receives an inquiry about the image data of the document element from the data processing device, the document element stores the image data in the cache memory even if the image data exists in the cache memory. Providing the second response to the data processing device if it is determined that it is not worth storing;
The print document processing system according to claim 1. The plurality of states include a third state indicating that the image data of the document element is stored in the cache memory,
The cache device is:
When the reservation request for securing the storage area for the image data of the document element is received from the data processing device, the reservation is made when the free space of the cache memory is insufficient to store the image data. A process for expelling at least one image data of the document element in the third state whose cache priority is lower than that of the requested document element from the cache memory and stored in the state storage means Eviction processing means for changing the cache state of the document element to be processed for eviction to the second state;
Further comprising
When the response providing unit receives an inquiry about image data of a document element from the data processing device, and the cache state of the document element stored in the state storage unit is the second state, the response providing unit Determining that the element is not worth storing image data in the cache memory;
The print document processing system according to claim 2. When the eviction processing unit receives a securing request for securing a storage area for image data of a document element from the data processing device, the eviction processing unit has a cache priority lower than that of the document element and the image data If the image data of all the document elements in the third state that are not in use from the data processing apparatus is expelled and the free space of the cache memory is insufficient to store the image data, At least one of the cache priority is low and image data is being used by any data processing device is changed to the second state, and the image data of the document element changed to the second state is Expelling the image data from the cache memory after it is no longer used by the data processing device;
The print document processing system according to claim 3. The plurality of states include a third state indicating that the image data of the document element is stored in the cache memory,
The response providing means, when receiving an inquiry about image data of a document element from the data processing apparatus, has a cache priority of the document element for determining a document element for evicting image data from the cache memory. Is lower than the threshold value, the document element is worth storing image data in the cache memory even if the cache state of the document element stored in the state storage means is the third state. Determining that the second response is provided to the data processing device;
The print document processing system according to claim 2. When the response providing unit determines that the document element to be queried is worth storing image data in the cache memory, the cache priority is lower than the document element and the image data is any data. Even if the image data of all the document elements that are not in use is evicted from the processing device, if the free space of the cache memory is insufficient to store the image data, the first threshold value is set to the document element. Threshold changing means for changing to a value lower than the cache priority and higher than at least one cache priority of the document element in which image data is stored in the cache memory;
The print document processing system according to claim 5, further comprising: The cache device is:
When a reservation request for securing a storage area for document element image data is received from the data processing apparatus, the cache priority of the document element subject to the securing request registers the image data in the cache memory. When the document element is less than a second threshold value for determining the document element to be changed, the cache state of the document element that is the target of the reservation request stored in the state storage unit is changed to the second state, and the reservation request First state changing means for responding to the data processing apparatus that has performed a failure of securing;
Further comprising
The data processing device includes:
In response to a request for securing failure from the cache device in response to the securing request, the image data creating unit executes creation of image data of the document element, and printing is performed using the image data created as a result. Means for generating an image and discarding the image data without registering it in the cache device;
Further comprising
The print document processing system according to claim 5 or 6. 8. The printed document according to claim 7, wherein the second threshold value used by the first state changing unit is a minimum value among cache priorities of each document element in the third state. Processing system. When the cache priority of the document element to be secured is equal to or higher than the second threshold, the cache priority is lower than the document element and the image data is not being used by any data processing apparatus. Even if the image data of all document elements in the state 3 is evicted, if the free space of the cache memory is insufficient to store the image data, the second threshold value is set as the cache priority of the document element. Second threshold value changing means for changing to a value lower than the cache priority of at least one of the document elements in which image data is stored in the cache memory,
The print document processing system according to claim 7 or 8, further comprising: For image data of a document element that is in the third state and whose cache priority is lower than the first threshold, after the image data is no longer used by any data processing device, the image data is stored in the cache Evicting means for evicting from the memory and changing the cache state of the document element stored in the status storage means to the second status;
The print document processing system according to claim 5, further comprising: The cache device is:
When a request for securing a storage area for image data of a document element is received from the data processing apparatus, the cache priority is lower than that of the document element, and the image data is used by any data processing apparatus When the image data of all the document elements in the third state that are not in the middle is evicted, the free space of the cache memory is insufficient to store the image data. In contrast, a means of responding to the failure of securing,
With
The data processing device includes:
In response to a request for securing failure from the cache device in response to the securing request, the image data creating unit executes creation of image data of the document element, and printing is performed using the image data created as a result. Means for generating an image and discarding the image data without registering it in the cache device;
Comprising
The print document processing system according to claim 3, wherein Cache priority update means for updating the cache priority of each document element so as to increase as the number of times the document element has received the inquiry increases;
When the cache priority stored in the state storage means exceeds the predetermined state change threshold for the document element in the second state, the cache state of the document element is changed to the first state. Second state changing means for changing to
The print document processing system according to claim 1, further comprising: A cache memory for storing image data in bitmap format or intermediate language format created by each data processing device;
A state storage unit that stores, for each document element of the print document data, one of a plurality of predefined states of the cache state of the image data of the document element in the cache memory, In a plurality of states, a first state indicating that there is no image data in the cache memory even though the cache priority is higher than each document element in which the image data is stored in the cache memory; And a second state indicating that there is no image data in the cache memory because the cache priority is lower than each document element in which the image data is stored in the cache memory. Means,
When an inquiry about image data of a document element is received from the data processing device, a first response is obtained when the cache state of the image data of the document element stored in the state storage means is the first state. Response providing means for providing a second response to the data processing device when the cache state of the image data of the document element is the second state;
A cache device comprising: Image data creating means for processing page description language data of each document element in the print document data to create image data in a bitmap format or an intermediate language format;
For each document element in the print document data, before causing the image data creation means to create image data of the document element, an inquiry means for making an inquiry as to whether or not the image data is in the cache device;
The response from the cache device to the inquiry by the inquiry means is that the image data is the object even though the document element to be queried has a higher cache priority than each document element in which image data is stored in the cache device. If the response is a first response indicating that the cache device is not in the first state, the image data generation unit executes the generation of image data of the document element, and is generated as a result. A print image is generated using the image data, and a process for registering the image data in the cache device is executed. A response from the cache device to the query by the query unit is a target of the query. is low cache priority than the document element document element is image data stored in said cache device If the image data is second was a response indicating a second state indicating that is not in the cache device, the image data generating means to execute the creation of image data of the document element is created as a result Control means for generating a print image using the obtained image data and for discarding the image data without registering it in the cache device;
A data processing apparatus comprising: Computer
A cache memory for storing image data in bitmap format or intermediate language format created by each data processing device;
A state storage unit that stores, for each document element of the print document data, one of a plurality of predefined states of the cache state of the image data of the document element in the cache memory, The plurality of states includes a first state indicating that no image data is stored in the cache memory, although the cache priority is higher than each document element in which the image data is stored in the cache memory, And a second state indicating that there is no image data in the cache memory because the cache priority is lower than each document element in which the image data is stored in the cache memory. ,
When an inquiry about image data of a document element is received from the data processing device, a first response is obtained when the cache state of the image data of the document element stored in the state storage means is the first state. Response providing means for providing a second response to the data processing device when the cache state of the image data of the document element is the second state;
Program to function as. Computer
Image data creation means for processing page description language data of each document element in the print document data to create image data in a bitmap format or an intermediate language format;
Inquiry means for making an inquiry as to whether or not the image data is in the cache device before causing the image data creation means to create image data of the document element for each document element in the print document data;
The response from the cache device to the inquiry by the inquiry means is that the image data is the object even though the document element to be queried has a higher cache priority than each document element in which image data is stored in the cache device. If the response is a first response indicating that the cache device is not in the first state, the image data generation unit executes the generation of image data of the document element, and is generated as a result. A print image is generated using the image data, and a process for registering the image data in the cache device is executed. A response from the cache device to the query by the query unit is a target of the query. is low cache priority than the document element document element is image data stored in said cache device If the image data is second was a response indicating a second state indicating that is not in the cache device, the image data generating means to execute the creation of image data of the document element is created as a result Control means for generating a print image using the obtained image data and discarding the image data without registering it in the cache device;
Program to function as.

Images