How to change the number of pages being processed by each processor core in ABBYY FineReader Server 14

Depending on the workflow properties below, ABBYY FineReader Server 14 creates a job either for each file or for each folder:

When processing a job, ABBYY FineReader Server splits documents into parts named slices, each of which is processed by a station's processor core.

The PagesSlice attribute of the DefaultJob element in the Configuration.xml configuration file specifies how many pages from a job will be processed by one processor core.

By default, the document is split into 25-page slices. The minimum slice size is 5 pages. 

Note: Decreasing the slice size increases the number of slices, thus increasing the time spent on transferring slices. Tests have shown that the optimal slice size is 25 pages.

Here are three examples of how ABBYY FineReader Server handles multi-page documents:

1. A station with a quad-core processor receives a document of 100 pages. The document is split into 4 slices containing 25 pages each (as per the default settings). One of the four slices is assigned to each of the four processor cores, and the processing of each slice takes the same amount of time, let's say a minute.

2. The same 100-page file is assigned to a station with a 32-core processor. It is split into four slices of 25 pages each, just like in the previous example, and each of these slices is assigned to a core, taking a minute to process while the remaining 28 cores stay idle.

3. A station with a quad-core processor receives a 20-page document. It is processed by one core since it is smaller than the default slice of 25 pages. It could be processed much faster if it was divided into four parts.

As seen from these examples, in some cases you can significantly increase the productivity of ABBYY FineReader Server by changing this attribute to take the document size and the number of processing cores into account.

If your workflow uses a document separation script, document slicing will be forbidden and the PagesSlice attribute will be automatically set to -1. Removing the document separation script from the workflow will allow document slicing and reset the PagesSlice attribute to its default value of 25.

You can forbid document slicing by setting the PagesSlice attribute either to -1 or to 0:

  • -1 forbids document slicing only if a document separation script is detected in the workflow properties. If no document separation script is detected, the setting will be changed from -1 to 25 and documents will be sliced.
  • 0 always forbids document slicing. To turn slicing back on, you will need to change the value of the PagesSlice attribute manually.

Configuration.xml file is located in the %PROGRAMDATA%\ABBYY FineReader Server 14 folder. The location of the PagesSlice attribute in Configuration.xml is: 

<OCRServer> > <QueueManager> > <Workflows> > <WorkflowSettings> > <InputSettings> > <DefaultJob>

For more information on the Configuration.xml file please refer to the Configuration File topic in the ABBYY FineReader Server 14 Help File or in the ABBYY Online Help:

Have more questions? Submit a request



Please sign in to leave a comment.