JP2016210175A - Image formation apparatus, control method of the same, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform control in a page order in a memory saving manner in an image formation apparatus which generates a display list from PDL data with a plurality of processor cores.SOLUTION: When generation of intermediate data of a subsequent page is completed prior to a preceding page, rasterization of the intermediate data of the subsequent page is performed after completion of generation and rasterization of the intermediate data of the preceding page.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a technique for generating intermediate data by a plurality of processing means.

  In rasterization based on print data (PDL data) described in a page description language, intermediate data described in an intermediate language is generated from PDL data, and the intermediate data is rasterized into a raster image. Patent Document 1 discloses a technique in which a printer equipped with a multi-core CPU having a plurality of processor cores executes rasterization of different pages in parallel. That is, each processor core is assigned pages to be rasterized in order, generates intermediate data from the PDL data of the assigned pages, and rasterizes the intermediate data, thereby generating a raster image of the page. The raster image is transmitted to the printer engine and printed.

JP 2012-16842 A

  In the printer of Patent Document 1, since a plurality of processor cores rasterize different pages, the following control is performed to make the order of printed pages coincide with the order of pages of PDL data. Even if the rasterization of the subsequent page by one processor core is completed first, it waits for the completion of the rasterization of the previous page and transmission of the raster image to the printer engine by the other processor core. Thereafter, the raster image of the subsequent page is transmitted to the printer engine.

  That is, in Patent Document 1, a raster image of a page after rasterization is completed first is kept in a memory until a raster image of the previous page is generated. For this reason, an amount of memory required to hold the raster image is required.

  In view of the above circumstances, the image forming apparatus of the present invention includes a receiving unit that receives PDL data including a plurality of pages, and a plurality of processing units that generate intermediate data of the plurality of pages for each page from the PDL data. Rasterizing means for rasterizing the intermediate data for each page generated by the plurality of processing means for each page, and the rasterizing means includes intermediate data for the succeeding page prior to generation of the intermediate data for the preceding page. When the data generation is completed, the rasterization of the intermediate data of the subsequent page is delayed until the generation and rasterization of the intermediate data of the previous page are completed.

  In an image forming apparatus having a plurality of processing units, the page order can be controlled with a smaller amount of memory.

1 is a hardware configuration diagram of an image forming apparatus. FIG. 3 is a software configuration diagram of the image forming apparatus. Schematic diagram of pipeline parallel processing and page parallel processing. 6 is a selection control flow of pipeline parallel processing and page parallel processing according to the first embodiment. Processing flow of pipeline PDL thread. Pipeline DL thread processing flow. Page DL thread processing flow. 7 is a control flow for switching between pipeline parallel processing and page parallel processing according to the second embodiment.

Example 1
Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

<Hardware block diagram of image forming apparatus>
FIG. 1 is a hardware block diagram illustrating the configuration of an image forming apparatus (printing apparatus) according to this embodiment. 1 includes a CPU 101, a RAM 102, a ROM (not shown), an HDD 103, a network I / F (interface) 104, a RIP (Raster Image Processor) 106, and a device I / F 107. The controller unit 100 receives print data (PDL data) described in a page description language (PDL) from an external device (for example, a host PC) connected via the network I / F 104 via a network 105 such as a LAN. Receive. The controller 100 is connected to the printer engine 108 via the device I / F 107 and controls input / output of image signals and device information.

  The CPU 101 reads out the program stored in the HDD 103 to the RAM 102 and executes it. Further, the CPU 101 comprehensively controls each device connected to the system bus 109 according to the program. The CPU 101 is a multi-core CPU having a plurality of processor cores, and generates (converts) intermediate data (display list: DL) from PDL data in parallel by multi-threads. The reason why the PDL data is converted into the dill playlist is that even if there are various types of PDL data, the subsequent processing is made common by converting the PDL data into DL of a common format. The multi-thread described below means that each of a plurality of cores of the CPU 101 functions as a separate thread. That is, it is equivalent to that different threads execute processing, and that different cores execute different processing. It is also possible to use a technique such as hyperthreading that can use a technique in which one processor core appears as if it is a plurality of processor cores by effectively utilizing the registers and pipeline idle time in the CPU.

  The RAM 102 functions as a main memory, work memory, and DL storage memory for the CPU 101. The DL storage memory is a memory area reserved in advance in the RAM 102, and is a memory area having a predetermined data size for storing DL for a plurality of pages. Note that the maximum DL data size for one page is determined in advance. As will be described later, if the generated DL data size of a page exceeds the maximum value during DL generation of a page, the DL thread (DL thread generation unit) temporarily stops generating the DL. A raster image is generated from the generated DL using the RIP 106. The DL thread then deletes the generated DL to create a storage memory space, compresses the generated raster image, and stores the compressed raster image as a DL in the storage memory where the space is created. Store. Then, the DL thread resumes generating the DL for the page. Such a series of processes is called a fallback process, and the DL thread prevents the DL data size for one page from exceeding a predetermined maximum value by the fallback process. A part of the RAM 102 also functions as a buffer 204 described later. The HDD 103 is used for the purpose of holding large-capacity data temporarily or for a long time.

  The RIP 106 generates a raster image from the DL generated by the CPU 101 (rasterization into a bitmap). As will be described later, the RIP 106 sequentially receives DLs transmitted by the CPU 101 in the page order for each page, and rasterizes the DLs in the page order. The rasterization by the RIP 106 is executed in parallel with the generation of DL by the CPU 101 and in units of pages. That is, the CPU 101 rasterizes the previous page (previous page) by the RIP 106 in parallel with the DL generation of the subsequent page (subsequent page). A raster image generated by rasterization is sent to the printer engine 108.

  The printer engine 108 has an engine I / F (not shown), and the engine I / F sequentially converts raster images sent from the controller unit 100 into image signals. The printer engine 108 is a printing unit that prints an image on a sheet such as paper based on the image signal. The printing method of the printer engine 108 may be an electrophotographic method, an inkjet method, or the like, and the method is not limited.

<Software module diagram of image forming apparatus>
FIG. 2 is a software configuration diagram of the image forming apparatus according to the present exemplary embodiment. Each software module illustrated in FIG. 2 is realized by loading the above-described program stored in the HDD 103 into the RAM 102 and executing it by the CPU 101.

  The parallel control unit 200 divides the image forming process of PDL data into predetermined units, and assigns the image forming process of each unit to a plurality of threads. Processing in a predetermined unit includes processing for generating one page of DL from PDL data, processing for interpreting PDL data for one object, processing for generating DL from the interpretation result of PDL data for one object, and the like. . The parallel control unit 200 according to this embodiment may initialize the buffer 204 and assign the thread 1 to interpret the PDL data for one object and store the interpretation result in the buffer 204. That is, the thread 1 is a pipeline PDL thread that functions as the interpretation unit 201. In addition, the parallel control unit 200 may assign a process of generating a DL from the interpretation result of the PDL data for one object stored in the buffer 204 to the thread 2. That is, the thread 2 is a pipeline DL thread that functions as the generation unit 203. Further, the parallel control unit 200 may assign to the thread 3 an interpretation process of PDL data of all objects included in a certain page and a DL generation process from the interpretation result. That is, the thread 3 is a page DL thread that functions as the interpretation unit 211 and the generation unit 213. The parallel control unit 200 may assign to the thread 4 interpretation processing of PDL data of all objects included in another page and generation processing of DL from the interpretation result. That is, the thread 4 is also a page DL thread that functions as the interpretation unit 211 and the generation unit 213. The difference is that the pages to be processed for DL generation are different.

  Note that the object here refers to, for example, a text object composed of a plurality of characters, a line segment object determined by specifying the coordinates of both ends of the line segment, a bitmap object, and the like.

  In this embodiment, the interpretation unit 201 performs processing for interpreting the PDL data of the object. The interpreter 201 interprets the PDL data one object at a time and writes the interpretation result to the work memory. In this work memory, the interpretation result for one object included in the PDL data is temporarily stored, and is overwritten by the interpretation result of the PDL data for one object that is subsequently interpreted.

  On the other hand, the interpretation unit 211 interprets the PDL data and writes the interpretation result to the work memory like the interpretation unit 201, but the interpretation unit 211 subsequently transfers the interpretation result to the buffer 204.

  In the present embodiment, the generation unit 203 performs processing for generating the DL of the object from the interpretation result. The generation unit 203 generates a DL from the interpretation result of the PDL data. The generation unit 203 acquires the interpretation result from the work memory of the interpretation unit 201. On the other hand, unlike the generation unit 203, the generation unit 213 acquires from the buffer 204, not from the work memory.

  In the present embodiment, the software module is realized as described above by the CPU 101 executing the program. However, the present invention is not limited to this, and a function equivalent to the above-described software module may be realized by hardware. In FIG. 2, four threads 1 to 4 are shown, but one thread may perform the operation of another thread. For example, the thread 1 becomes a pipeline DL thread in the pipeline parallel processing described later, but may be a page DL thread like the thread 3 in the page parallel processing. The same applies to thread 2 and thread 4.

  In the present embodiment, switching between two types of parallel processing realizes speedup by parallel processing not only for a plurality of pages but also for print data for only one page.

  One parallel process is a pipeline process in which the respective processes of the interpretation unit 201 and the generation unit 203 are executed by the pipeline control unit 202 using separate threads. That is, for the first page, when generating a DL for one object, a thread for processing to interpret the PDL data of the object and a thread for generating DL from the interpretation result are executed in parallel by pipeline processing. Is done. That is, while a thread that functions as the generation unit 203 generates a DL from the interpretation result of PDL data of a certain object, another thread that functions as the interpretation unit 201 interprets PDL data of the next object. Then, the thread that functions as the generation unit 203 generates a DL from the interpretation result of the other thread. The above-described series of processing flow is repeated for one page. This is pipeline parallel.

  In the other parallel processing, for each page after the second page, the number of threads that function as the interpreter 211 and the generator 213 is limited to one per page. Then, a plurality of threads that generate respective DLs of a plurality of pages are executed in parallel in units of pages. That is, a certain thread generates DL from PDL data of a certain page, and another thread generates DL from PDL data of another page. This is page parallel.

<Overview of page parallel and pipeline parallel switching>
In page parallel using a multi-core CPU, separate threads (processor cores) generate DLs for different pages in parallel for multiple pages of PDL data, so the effect of speeding up by multi-thread parallel processing is obtained. be able to. On the other hand, in page parallel performed by a multi-core CPU for single-page PDL data, a certain thread (processor core) generates a DL from PDL data for the first page. However, the remaining threads (processor core) are in an idle state in which neither PDL data is interpreted nor DL is generated, and the parallel effect of multithread (multicore) is obtained for DL generation from single-page PDL data. I can't.

  Therefore, in this embodiment, DL is generated by pipeline parallel processing for the first page of PDL data, and DL generation for the second and subsequent pages is performed by page parallel processing. As a result, not only a plurality of pages of PDL data but also a single page of PDL data has a multithread parallel effect. FIG. 3 is a conceptual diagram of this process.

  In the processing of the first page, pipeline parallel processing is performed. For example, when the objects 1 and 2 are included in the PDL data of the first page, the thread 1 (processor core 1) interprets the PDL data of the object 1, and then the thread 2 (processor core 2) The DL of object 1 is generated from the interpretation result. During the generation, the thread 1 interprets the PDL data of the object 2, and the thread 2 that has generated the DL of the object 1 generates the DL of the object 2 from the interpretation result. In this way, since the processing of each thread can be executed in parallel, the generation of the DL for the first page is accelerated.

  In the subsequent second and third page processes, page parallel processing is performed. For example, when the objects 3 and 4 are included in the PDL data on the second page, the thread 1 continues to interpret the PDL data of the object 3 and generate the DL. The same applies to the object 4. Thereby, the DL of the second page is generated. On the other hand, for example, when the objects 5 and 6 are included in the PDL data of the third page, the thread 2 continues to interpret the PDL data of the object 5 and generate the DL. The same applies to the object 6. As a result, the DL of the third page is generated. In this way, the processing of each thread can be executed in parallel in units of pages, so that the generation of DL for the second page and the third page is speeded up. The pipeline parallel processing may be applied to the second page and the third page as in the first page. However, the reason for not doing so in the present embodiment is as follows. In other words, multi-thread pipeline parallel processing requires communication between threads, so communication costs are higher than page parallel processing.

<Parallel processing switching control flow>
A control flow for switching between page parallel processing and pipeline parallel processing will be described with reference to the flowchart of FIG. This control flow is executed in the thread A in response to the parallel control unit 200 receiving PDL data from the external device via the network I / F 104.

  In S401, the parallel control unit 200 performs an initialization process. Specifically, the parallel control unit 200 manages the DL generation target page number and the DL transmission target page number, and sets both to 1. The parallel control unit 200 activates a plurality of threads for the number of processor cores of the multi-core CPU. The parallel control unit 200 reserves the buffer 204. For example, in the case of a 2-core CPU, two threads are activated. In the subsequent processing, the parallel control unit 200 sets the affinity mask so that each of the activated multiple threads is executed by a separate core.

  In S403, the parallel control unit 200 determines whether the generation target page number is 1, that is, whether the DL generation target is the first page. If the generation target page number is 1, in order to generate and transmit DL for the first page by pipeline parallel processing, the processing after S404 is performed. The pipeline parallel processing is as described with reference to FIG. Further, if the generation target page number is not 1, the processing from S410 onward is performed in order to generate and transmit DL by page parallel processing for the second and subsequent pages. Here, the first page is a predetermined page that is handled separately from the second and subsequent pages, the second and subsequent pages are pages other than the predetermined page, and the parallel control unit 200 determines whether the DL generation target page is a predetermined page. It is a generation control means for controlling the parallel processing method depending on whether or not.

  Processing subsequent to S404 for pipeline parallel processing will be described.

  In S <b> 404, the parallel control unit 200 instructs the plurality of activated threads to perform pipeline parallel processing of the PDL data of the generation target page number in cooperation with each other. For example, the parallel control unit 200 instructs the thread 1 to perform interpretation processing of PDL data of an object in pipeline parallel processing. At this time, the thread 1 functions as the interpretation unit 201. On the other hand, the parallel control unit 200 instructs the thread 2 to perform object DL generation processing from the interpretation result of the object PDL data in the pipeline parallel processing. At this time, the thread 2 functions as the generation unit 203. As will be described later, the parallel control unit 200 causes the generation unit 203 to secure a memory area for the maximum value of the DL data size of one page on the DL storage memory.

  In S405, the parallel control unit 200 increments the generation target page number by 1, thereby setting the DL generation target page to the second page.

  In step S406, the parallel control unit 200 determines whether a DL generation completion notification for the first page has been received from a thread that is performing pipeline processing. If the notification has been received, the process advances to step S407. This completion notification includes the memory address of the DL storage destination of the first page. The completion notification includes the page number of the generated DL, and the page number here is “first page”. In other cases, it waits until a DL generation completion notification for the first page is received. The reason for waiting only for the first page is to prevent the next page from being processed during the pipeline parallel processing of the first page, thereby increasing the processing efficiency of the first page.

  In S407, the parallel control unit 200 transmits the DL of the transmission target page number (that is, the DL of the first page) to the RIP 106. The transmitted DL is rasterized by the RIP 106. Specifically, the parallel control unit 200 notifies the RIP 106 of the memory address included in the completion notification, and the RIP 106 acquires the DL stored in the memory address and performs rasterization. The parallel control unit 200 causes the DL thread generation unit 203 to release the memory area secured on the DL storage memory in response to detecting the completion of the DL rasterization of the first page. The generated raster image is sent to the printer engine 108 and printed on paper.

  In S408, the parallel control unit 200 increments the transmission target page number by 1, thereby setting the DL transmission target page to the second page. Then, the process proceeds to S409.

  On the other hand, the page parallel processing after S410 will be described.

  In S410, the parallel control unit 200 repeats S411 to S413 until there is no thread to which no process is assigned (empty thread).

  In step S411, the parallel control unit 200 determines whether an instruction to perform DL generation processing of all pages included in the PDL data is performed in page parallel. If the DL generation process is instructed for all pages, the process exits the loop process of S410 and shifts to S414. Otherwise, the process proceeds to S412. In addition to the processing of S411, the parallel control unit 200 calculates the total size of DLs not transmitted (unrasterized) to the RIP 106 among the DLs of pages generated so far, and this total size is a threshold value. If smaller, the process may be shifted to S412. Conversely, if the total size is greater than or equal to the threshold, the process proceeds to S414. In this way, by suppressing the generation of DL with reference to the total size of untransmitted DL, DL generation of a page with a young page number is significantly delayed, and other pages (page numbers larger than the young page number) Even if the DL of the page is generated first, it is possible to prevent a shortage of DL storage memory and buffer overflow. Note that this control may be performed based on the number of generated and unrasterized pages, not the total size of generated and unrasterized DLs. For example, if the number of generated and unrasterized pages is smaller than the threshold, the process proceeds to S412; otherwise, the process proceeds to S414. Alternatively, the parallel control unit 200 determines whether the free capacity of the DL storage memory is high enough to secure a memory area for the maximum DL data size of one page. That is, it is determined whether the free space is equal to or greater than a predetermined threshold value (maximum value). If so, the process proceeds to S412; otherwise, the process proceeds to S414. The free space is a data size of a memory area that can be used or released in the DL storage memory, and is obtained as follows. That is, the free space is a size obtained by subtracting “the total size of the memory area reserved for each page for the generated and unrasterized DL” from the predetermined data size reserved in advance as the DL storage memory. Instead of the “total size of memory area reserved for each page for generated and unrasterized DL”, “total size of generated and unrasterized DL” may be used.

  In S <b> 412, the parallel control unit 200 assigns the DL generation process for the page with the generation target page number to the empty thread so as to execute the page parallel process. That is, a thread to which a DL generation process is assigned interprets PDL data of each object included in the page and generates a DL from the interpretation result. By doing so, DL generation processing is performed in parallel on a page-by-page basis, so a plurality of pages of DL included in the PDL data can be generated at high speed. For example, if the thread 1 is an empty thread, the thread 1 functions as the interpretation unit 211 and the generation unit 213. The same applies when thread 2 is an empty thread. As will be described later, the parallel control unit 200 causes the generation unit 213 to secure a memory area for the maximum value of the DL data size of one page on the DL storage memory.

  In S413, the parallel control unit 200 increments the generation target page number by one to set the DL generation target page as the next page.

  The above is the content of the loop processing.

  In step S <b> 414, the parallel control unit 200 determines whether a DL generation completion notification has been received from a thread that is performing page parallel processing. If received, the process proceeds to S415, and the generated and untransmitted DL is transmitted to the RIP 106. Otherwise, the process proceeds to S409. The completion notification includes the storage address and page number of the generated DL (for example, the second page and the third page), and is managed by the parallel control unit 200 as the generated DL page number. The

  In S415, the parallel control unit 200 determines whether the generation of the DL of the transmission target page number has already been completed based on the generated DL page number and the transmission target page number. That is, it is determined here whether the generated DL can be transmitted (delivered) to the RIP 106 in order from the first page. This transmission target page number is nothing but the page number to be rasterized this time by the RIP 106. If the generation of the DL for the transmission target page number has already been completed, the process proceeds to S416, and if not, the process proceeds to S409. Here, when proceeding to S409, the DL of the page number included in the completion notice received in S414 is held in the storage destination memory. However, a memory saving effect can be obtained rather than being held in a state converted into a raster image.

  In S416, the parallel control unit 200 transmits the DL of the transmission target page number to the RIP 106. The transmitted DL is rasterized by the RIP 106. Specifically, the parallel control unit 200 notifies the RIP 106 of the DL storage destination memory address of the transmission target page number included in the completion notification acquired so far, and the RIP 106 is stored in the memory address. Rasterize by acquiring the DL. Then, the parallel control unit 200 causes the DL thread generation unit 213 to release the memory area of the page secured in the DL storage memory in response to detecting the completion of the DL rasterization for one page. .

  In step S417, the parallel control unit 200 increments the transmission target page number by one so that the DL generated in the page order can be transmitted to the RIP 106.

  In S415 to S417 described above, the parallel control unit 200 corresponds to transmission control that performs control of the page order of DL sent to the RIP 106 for rasterization. In this order control, even if the DL generation of the subsequent page (for example, the third page) is completed prior to the DL generation of the page in the order to be transmitted (the preceding page: for example, the second page), the third page The transmission of the DL to the RIP 106 is waited until the transmission of the DL of the second page, which is the previous page, is completed. That is, DL rasterization of the subsequent page is delayed. As a result, the DL is rasterized in an appropriate page order to generate a raster image. The raster images are sent to the printer engine 108 in the order in which they are generated, and the pages are printed in the correct order.

  In S409, the parallel control unit 200 determines whether DL generation and transmission to the RIP 106 of all pages included in the PDL data are completed. If completed, the process of this flow ends. Otherwise, the process returns to S403.

  With the above processing, control for switching between page parallel processing and pipeline parallel processing by the parallel control unit 200 is performed.

<Pipeline PDL thread>
Next, the PDL data interpretation process in the pipeline parallel processing assigned to the empty thread based on the instruction in S404 will be described using the flowchart of FIG. The processing of this flowchart is executed by a pipeline PDL thread triggered by an instruction that an empty thread functions as a pipeline PDL thread. Therefore, the subject of processing is the interpretation unit 201 of the pipeline PDL thread.

  In step S1102, the interpretation unit 201 interprets PDL data of an uninterpreted object included in the page with the generation target page number. The interpretation result is written in the work memory area.

  In step S <b> 1104, the interpretation unit 201 transfers the interpretation result written in step S <b> 1102 to the buffer 204. This transfer is performed in response to receiving a transfer permission notification from a pipeline DL thread described later. The transferred interpretation result is processed by the pipeline DL thread as described later.

  In step S1105, the interpretation unit 201 determines whether the interpretation of the PDL data of all objects included in the page with the generation target page number has been completed. If completed, the process moves to S1106. Otherwise, the process returns to S1102, and interpretation of PDL data of another uninterpreted object is performed.

  In step S1106, the interpretation unit 201 waits for a DL generation completion notification from the pipeline DL thread. Here, the interpretation unit 201 acquires a memory address (described later) included in the DL generation completion notification. When the DL generation completion notification is received, the process proceeds to S1107.

  In step S <b> 1107, the interpretation unit 201 transmits a DL generation completion notification including the acquired memory address to the parallel control unit 200 that is realized by a thread other than the pipeline PDL thread and the pipeline DL thread. Further, the DL generation completion notification includes the page number of the generated DL and is received in S406.

  The above processing is processing by the pipeline PDL thread.

<Pipeline DL thread>
Next, DL data generation processing in pipeline parallel processing allocated to an empty thread based on the instruction in S404 will be described using the flowchart of FIG. The processing of this flowchart is executed by a pipeline DL thread triggered by an instruction that an empty thread functions as a pipeline DL thread. Therefore, the main subject of the processing is the pipeline DL thread generation unit 203.

  In step S <b> 1201, the generation unit 203 acquires the interpretation result of the object PDL data from the buffer 204.

  In step S1202, the generation unit 203 generates an object DL from the interpretation result acquired in step S1201. The generated DL is stored in the DL storage memory. At this time, the generation unit 203 transmits a transfer permission notification to the pipeline PDL thread. As a result, after the interpretation result in the buffer has been processed, the next interpretation result is controlled to be overwritten in the buffer. In addition, when the generation unit 203 starts generating DL for one page, the generation unit 203 secures a memory area for the maximum value of the DL data size of one page in the DL storage memory, and is generated in the memory area. DL is sequentially stored. In addition, as described above, the generation unit 203 determines whether the generated DL data size exceeds the maximum DL data size for one page during DL generation. Process.

  In step S1203, the generation unit 203 determines whether the DL generation of all objects included in the page with the generation target page number has been completed. If it has been completed, the process moves to S1204. Otherwise, the process returns to S1201 to process the interpretation result stored in the buffer.

  In step S1204, the generation unit 203 transmits a DL generation completion notification to the pipeline PDL thread. This DL generation completion notification is received in S1106. This DL generation completion notification includes the memory address where the generated DL for one page is stored.

  The above processing is processing by the pipeline DL thread.

<Page parallel DL generation flow>
Next, DL data generation processing in page parallel processing allocated to a free thread based on the instruction in S412 will be described using the flowchart of FIG. The processing of this flowchart is executed by a page DL thread triggered by an instruction that an empty thread functions as a page DL thread. Therefore, the main subject of processing is the page DL thread interpreter 211 and the generator 213.

  In step S1301, the interpretation unit 211 interprets the PDL data of the object included in the page having the generation target page number, and writes the interpretation result to the work memory.

  In step S1302, the generation unit 213 reads the interpretation result written in the work memory and generates a DL of the object. The generated DL is stored in the DL storage memory. When the generation unit 213 starts generating one page of DL, the generation unit 213 secures a memory area for the maximum value of the DL data size of one page in the DL storage memory, and is generated in the memory area. DL is sequentially stored. In addition, as described above, the generation unit 213 determines whether the generated DL data size exceeds the maximum DL data size for one page during DL generation. Process.

  In step S1303, the generation unit 213 determines whether DL generation of all objects included in the page with the generation target page number has been completed. If completed, the process proceeds to S1304; otherwise, the process returns to S1301 to interpret the PDL data of the next object.

  In S1304, the generation unit 213 transmits a DL generation completion notification to the parallel control unit 200. This DL generation completion notification includes a memory address that is a storage destination of the generated DL for one page and the page number of the DL.

  Since the page DL thread described above is performed for each page by a plurality of threads, DL generation can be performed efficiently.

  As described above, by switching between pipeline parallel and page parallel, the processing by the multi-core CPU can be efficiently parallelized for single-page PDL data or multiple-page PDL data.

  Note that the above processing is pipeline parallel, but the point is that parallel processing is also performed for the first page. For example, for the first page, the first object included in the PDL data may be converted by the thread 1 into DL, and the second object may be converted by the thread 2 into DL. However, in the case of such object parallelism, the timing of completion of object DL generation differs from object to object, so the order of saving the generated object DLs must be controlled.

  In contrast, pipeline parallelism has the advantage that multiple threads perform pipeline processing on a single object, so there is no need to be aware of the order in which objects are stored in DL.

(Example 2)
In the first embodiment, pipeline parallel processing is executed for the first page so that the effect of parallel execution by multi-core (multi-thread) can be obtained regardless of the number of pages of PDL data received from the external device, Page parallel processing was executed for the second and subsequent pages.

  However, if there are a plurality of pages, the parallel effect may be higher when page parallel processing is performed from the first page. In particular, PDL data with an even number of pages is more effective.

  For example, in the case of 2-page PDL data, the parallel effect by multithread (multi-core) is higher when the DL is generated in page parallel between the first page and the second page.

  Therefore, in the present embodiment, a configuration for determining whether to perform pipeline parallel processing or page parallel processing according to the number of pages of PDL data will be described with reference to FIG. In the following embodiments, the description overlapping with the first embodiment will be omitted.

  FIG. 8 is obtained by changing a part of the flowchart of FIG. 4 of the first embodiment. That part is that the process of S402 is added and the process of S403 is changed to S403-1. Therefore, only the above difference will be described.

  In S402, the parallel control unit 200 determines the total number of pages of the PDL data received from the external device by interpreting the PDL data. The following judgment methods can be considered.

  For example, the parallel control unit 200 counts the total number of page break commands while sequentially reading PDL data from the beginning to the end. Since the page break command separates pages, the total number of pages of PDL data can be acquired.

  Alternatively, the parallel control unit 200 reads the header information if the PDL data includes header information, and acquires the total number of pages if the total number of pages of the PDL data is described in the header information.

  Alternatively, the parallel control unit 200 may determine the PDL type of the PDL data and determine the total number of pages based on the PDL type. For example, if the PDL type is Jpeg, TIFF, or the like, since it is a single page image, it is preferable to perform pipeline parallel instead of page parallel. Therefore, the parallel control unit 200 determines that the total number of pages is 1 when the PDL type is Jpeg, TIFF, or the like.

  In step S403-1, the parallel control unit 200 advances the process to step S404 if the total number of pages is one page based on the determination of the total number of pages in step S402, and advances the process to step S410 if the number of pages is plural.

  As described above, if it can be regarded as single-page PDL data, pipeline parallel processing is performed on the first page, and if it can be regarded as multiple-page PDL data, page parallel processing is performed in parallel with the second page. To the first page. Thereby, a parallel effect can be heightened more.

(Other examples)
In the above embodiment, pipeline parallel processing is performed only on the first page (first page) of PDL data. However, pipeline parallel processing may be performed on the last page of PDL data.

  In the above embodiment, the pipeline parallel processing and the page parallel processing are appropriately selected when generating the DL from the PDL data, but the same applies when generating the raster image from the PDL data. The idea can be applied.

  Further, in the image forming apparatus of the above embodiment, one RIP 106 rasterizes DL sequentially for each page, but a plurality of RIPs may be employed instead of the RIP 106.

  As an example, a configuration in which a plurality of RIPs perform rasterization of different regions on the same page in parallel from the generated DL may be used. For example, the first RIP uses the DL of the generated page to rasterize the first band region of the page, and the second RIP performs the page parallel to the rasterization by the first RIP. The second band is rasterized using the DL. In this way, the delay caused by waiting for DL generation can be recovered by increasing the speed of parallel rasterization of pages.

  As another example, a configuration in which a plurality of RIPs perform rasterization of DLs of different pages in parallel may be used. For example, the first RIP rasterizes the DL of the preceding page, and the second RIP rasterizes the DL of the subsequent page in parallel with the rasterization by the first RIP. Even in such a configuration, when the DL generation of the subsequent page is completed earlier than the preceding page, the rasterization of the DL of the subsequent page generated earlier is delayed as in the above embodiment. However, in this configuration, the first RIP and the second RIP perform rasterization of the DL of the preceding page and the succeeding page in response to completion of generation of the DL of the preceding page, not completion of rasterization of the preceding page. Start. In this way, it is possible to obtain the effect of speeding up rasterization by parallel rasterization while correcting the page order in DL generation. If the rasterization of the subsequent page is completed before the preceding page in this parallel rasterization, the second RIP holds the raster image for one page of the subsequent page, and further waits for the rasterization of the subsequent page. Configure as follows. In response to the completion of rasterization of the preceding page by the first RIP, the first RIP transmits the raster image of the preceding page to the printer engine 108, and then the second RIP transmits the raster image of the subsequent page to the printer engine. To 108. With this configuration, even if the page DL generation order is reversed, the effect of speeding up by parallel rasterization is obtained while keeping the raster image of multiple pages kept, and printing is performed in the correct page order. It can be carried out.

  The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in a computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

Claims (20)

Receiving means for receiving PDL data including a plurality of pages;
A plurality of processing means for generating intermediate data of the plurality of pages for each page from the PDL data;
Rasterizing means for rasterizing the intermediate data for each page generated by the plurality of processing means for each page;
Have
When the generation of the intermediate data of the subsequent page is completed prior to the generation of the intermediate data of the previous page, the rasterizing means performs the generation of the intermediate data of the subsequent page until the generation of the intermediate data of the previous page and the completion of the rasterization. An image forming apparatus characterized by delaying rasterization.   The image forming apparatus according to claim 1, wherein each of the plurality of processing units generates intermediate data of different pages. Management means for managing page numbers to be rasterized by the rasterizing means;
When the generation of the intermediate data of the subsequent page is completed prior to the generation of the intermediate data of the page of the page number managed by the management unit, the rasterizing unit starts from the page of the page number. 3. The image forming apparatus according to claim 1, wherein rasterization of intermediate data of the subsequent page is delayed until generation of intermediate data of each page up to the previous page and completion of rasterization are completed.   The rasterizing means obtains the amount of intermediate data that has not been rasterized among the intermediate data of pages that have already been generated, and suppresses the generation of intermediate data for the remaining pages when the amount is equal to or greater than a threshold value. The image forming apparatus according to claim 1, wherein the image forming apparatus is an image forming apparatus.   The rasterizing means is configured such that the amount is the total size of intermediate data that has not yet been rasterized among the intermediate data of the already generated pages. If the total size is greater than or equal to the threshold value, the intermediate data of the remaining pages The image forming apparatus according to claim 4, wherein generation of the image is suppressed.   The rasterizing means is configured such that the amount is the number of intermediate data pages that have not yet been rasterized among the intermediate data of already generated pages, and the intermediate data of the remaining pages when the number of pages is equal to or greater than a threshold value. The image forming apparatus according to claim 4, wherein generation of the image is suppressed. The plurality of processing means, when generating the intermediate data for each page from the PDL data,
For the predetermined page of the PDL data, the generation of intermediate data of objects included in the page is performed in parallel by the plurality of processing means,
7. For any page other than the predetermined page, intermediate data generation of an object included in the page is performed in parallel by the processing means different for each page. The image forming apparatus described. About the predetermined page
One processing means of the plurality of processing means interprets PDL data of an object included in the page,
The image forming apparatus according to claim 7, wherein another processing unit of the plurality of processing units generates intermediate data of the object from the interpretation result. About the predetermined page
The one processing means transfers the interpretation result from the memory in which the interpretation result is written to a buffer,
9. The image forming apparatus according to claim 8, wherein the another processing unit acquires the interpretation result transferred to the buffer and generates the intermediate data. For pages other than the predetermined page,
The one processing means interprets PDL data of an object included in a certain page and generates intermediate data of the object from the interpretation result, and
The image forming apparatus according to claim 9, wherein the another processing unit interprets PDL data of an object included in another page and generates intermediate data of the object from the interpretation result. For pages other than the predetermined page,
The said one processing means acquires the said interpretation result from the memory in which the said interpretation result was written without going through the said transfer, and produces | generates the said intermediate data. Image forming apparatus.   The plurality of processing means treats all pages included in the PDL data as the predetermined page when the total number of pages of the received PDL data is regarded as one page. Item 12. The image forming apparatus according to any one of Items 7 to 11.   The plurality of processing means treats all pages included in the PDL data as pages other than the predetermined page when the total number of pages of the received PDL data is regarded as a plurality of pages. The image forming apparatus according to any one of claims 7 to 12.   The image forming apparatus according to claim 12, wherein the total number of pages is acquired by interpreting the received PDL data.   14. The image forming apparatus according to claim 12, wherein the total number of pages is regarded as one page by determining a PDL type of the received PDL data. .   And determining means for determining whether to treat all pages included in the PDL data as the predetermined page or as pages other than the predetermined page, based on the PDL type of the received PDL data. The image forming apparatus according to any one of claims 7 to 11.   The image forming apparatus according to claim 7, wherein the predetermined page is a first page of the PDL data.   18. The image forming apparatus according to claim 1, further comprising: a printing unit that receives the data rasterized by the rasterizing unit in the order of rasterization and prints the data on a sheet. A control method of an image forming apparatus having a plurality of processing means,
A receiving step of receiving PDL data including a plurality of pages;
A control step of generating intermediate data of the plurality of pages from the PDL data by the plurality of processing means different for each page;
A rasterizing step of rasterizing the generated intermediate data for each page for each page;
Have
In the rasterizing step, when the generation of the intermediate data of the subsequent page is completed before the generation of the intermediate data of the previous page, the intermediate data of the subsequent page is generated until the generation of the intermediate data of the previous page and the completion of the rasterization. A control method for an image forming apparatus, characterized by delaying rasterization.   The program for functioning a computer as at least one of each means of any one of Claims 1 thru | or 18.

Images