Patent Application
20110199635
Printing involves converting data from a form viewable on a computer monitor or the like to a form that may be printed by a printer. For example, an image in the portable document format (PDF) may be represented by several images that are stacked on top of each other and may include a plurality of different colors. The image that should be printed is typically the image viewed from the top. In most circumstances, images underneath other images cannot be printed in two-dimensional printing. In order to covert an image for printing, a display list may be created, wherein the display list includes shapes to be printed. In addition, the display list may include the sizes, locations, and colors of the shapes.
The display list is then translated to a bit map or the like, wherein each dot to be printed is represented by a number or, in the case of color printing, a plurality of numbers. In some embodiments, bit maps are created for each colorant, such as ink in the printer. Thus, in a printer that uses six different colorants, six bit maps are created. In order to convert an image for printing, a display list may be created, wherein the display list includes shapes to be printed and/or high level constructs such as fonts and images. In addition, the display list may include the sizes, locations, orientations, and colors of the shapes and/or high level constructs.
As printers become faster, the need to speed up the data processing increases. At present, the creation of the bit maps slows the printing process.
As described in greater detail below, methods of using two processors to convert data for printing purposes are disclosed herein. Some embodiments of the printing process are referred to as raster image printing. An example of a computer system 100 having two processors associated therewith is shown in
The computer system 100 also includes a second processor 110, which is referred to herein as the graphics processing unit (GPU) 110. The GPU 110 may be a graphics accelerator as used in some computer systems. The GPU 110 may be manufactured to generate data representative of three-dimensional graphics for display. Accordingly, the GPU 110 performs repetitive processing on very large amounts of data. For example, the GPU 110 generates a number for every colorant of each dot that is to be printed on a page. If a page is printed using several colors, a number representing each color is generated for every no that is to be printed. Because the GPU 110 is adapted for graphics processing, it can process a great number of dots simultaneously, which significantly reduces the processing time. The GPU 110 is typically not a general purpose processor like the CPU 104.
A GPU memory 114 is associated with the GPU 110 by way of a bus 116. The GPU memory 114 is sometimes referred to as the second memory and may be used or accessed solely by the GPU 110. Because the GPU 110 processes such a large amount of data, the data transfer rate or memory throughput between the GPU 110 and the GPU memory 114 is fast. The increased data transfer rate is improved by the width of the bus 116, which may be three hundred to five hundred bits. The memory throughput of the GPU memory 114 may be greater than the memory throughput of the first memory 106. Both the GPU 110 and the second memory 114 may be located on the same circuit board, which may be connected to a personal computer interface (PCI) bus or the like within the computer system 100.
A bus 120 is used to transfer data between the CPU 104 and the GPU 110. The data transferred between the CPU 104 and the GPU 110 may be compressed. Therefore, the data transfer rate on the bus 120 may not need to be as great as the data transfer rate on the bus 116.
Having described the components of the computer system 100 relevant to printing, the use of the components will now be described in processing data for printing purposes.
A flow chart 150 of the printing process is shown in
At step 156, the objects generated by the interpretation are decomposed into individual elements. At this stage, the date representative of the page is sometimes referred to as a display list. The display list may represent the page in three dimensions. For example, some images may be on top of one another. In one example, the page may be created by laying one image partiality on top of another, thus, the display list may contain both images, although only a portion of the lower image is to be displayed or printed. In other embodiments, the upper image may be transparent or have a degree of transparency. In these embodiments, portions of upper and lower images may be displayed and need to be decomposed, which is sometimes referred to as blending or combining.
With additional reference to
The data representation of the page may be in the form of a his map or other similar data structure, wherein numbers represent each dot to be printed. In the case of color images, a color conversion process, as shown at step 164, may need to be performed. The color conversion may generate several bit maps each representing a different color plane. For example, the original page may have been generated using red, green, and blue. The printer for which the page is to be sent may use cyan, yellow, black, and magenta to generate all colors. Therefore, a specific color generated by the red, green, and blue color planes needs to be represented by the cyan, yellow, black, and magenta color planes or bit maps. The GPU 110 may generate one color plane for each color that is printed by the printer. Each color plane may be a bit map wherein the bit map consists of a number corresponding to the color intensity for every dot to be printed. Thus, a high number may be indicative of a lot of colorant to be placed on a dot and a low number may be indicative of a little or no colorant, to be placed on a dot. The combination of colors printed will then print the aforementioned color.
The GPU 110 is well suited for the rasterization and color conversion. The GPU 110 may be a graphics accelerator that is designed to perform repetitive tasks on large amounts of data very quickly. Because the memory throughput associated with the GPU 110 is very large, it is able to process the bit maps quicker than the CPU 104. Therefore, having the GPU 110 perform these tasks improves the printing times of pages. In addition, the GPU 110 may only perform this single function. Accordingly, the data stored in the GPU memory 114 will not be changed by other processes, which enables the GPU 110 to process the data using high performance memory caching and access mechanisms that would be unsuitable for a general purpose CPU. Thus, after the GPU 114 sets a value of a memory location, the value will likely not have to change, which further increases the efficiency of the GPU 110.
The GPU 110 may then planarize the bit maps to place them in a form that is usable by the printer at step 166. The bit maps may then be transferred to the CPU 104 or the CPU memory 106 in step 168. In some embodiments, the bit maps are compressed prior to transferring by an conventional compression tool. In some embodiments, an Indigo Compression Format is used to compress the bit maps, in step 170 data is transmitted from the CPU 104 to a printer for printing.
The printing process has been described above as being performed on a page. In some embodiments, the process may be performed on a portion of a page, which reduces the amount of memory required for the processing. For example, several lines on a page may be processed as described above, followed by processing of successive lines of the page. In other embodiments, a page may be divided into portions, such as rectangular portions, and may be processed individually.
Because the processing described above is performed on two processors, they may work simultaneously on different portions of a page. An example of this processing is shown in the chart 200 of
Printing commences at step 204 with the CPU 104 processing a first page or first portion of a page. The processing performed by the CPU 104 may include interpreting the page and decomposing the page to create a display list as described above and as shown by the flow chart 150 of
While the GPU 110 is processing the first page, the CPU 104 processes a second page at step 208. When the GPU 110 is finished processing the first page at step 206, the GPU 110 transfers the data to the CPU 104 and the CPU 104 transfers the processed second page to the GPU 110. The GPU 110 and CPU 104 may now operate simultaneously. More specifically, while the GPU 110 processes the second page at step 210, the CPU 104 sends the data representative of the first page to the printer at step 214 and processes a third page at step 216. This simultaneous processing proceeds until all pages or portions of pages are printed. By using the GPU 110 and the CPU 104 simultaneously, the time required to print a page is significantly reduced.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/US2008/053503 | 2/8/2008 | WO | 00 | 4/15/2011 |
