Metadata Sequencer VS PDF Splitter using a PlanetPress Suite form

When used in combination with the Create PDF plugin, tests have shown a significant speed increase performance when going from Metadata Sequencer to PDF Splitter based on metadata. These are the benchmark results from those tests:

Environment:

• Windows 10 OS
• 4 Cores
• 16 gb RAM
• Workflow 8.6.1

Using PDF Splitter, split based on metadata:

• Data file + PP7 form = 25 369 data pages (10 000 documents)
• Time of completion = 01:26:25:683 (1 hour, 26 minutes)

Using Metadata Sequencer:

• Data file + PP7 form = 25 369 data pages (10 000 documents)
• Time of completion = 07:48:40:369 (7 hour, 48 minutes)

Metadata Sequencer VS PDF Splitter using a PDF as input

When used in combination with the Create PDF plugin, tests have shown an insignificant speed increase performance when going from PDF Splitter based on metadata to Metadata Sequencer. These are the benchmark results from those tests:

Environment:

• Windows 8 OS
• 16 Cores
• 32 gb RAM
• Workflow 8.6.1

Using PDF Splitter, split based on metadata*:

• PDF = 25 369 data pages (10 000 documents)
• Time of completion = 00:39:19:390 (39 minutes, 19 seconds)

Using Metadata Sequencer:

• PDF = 25 369 data pages (10 000 documents)
• Time of completion = 00:37:14:312 (37 minutes, 14 seconds)

Note: The test had to be done on a machine with more RAM because it ran out of it while running the PDF Splitter on the previous environment (16 gb RAM).

Output Creation Setting Separation VS PDF Splitter

Using the Separation option of the Output Creation Settings has shown a significant speed increase performance in comparison to using the PDF Splitter plugin. These are the benchmark values from those tests:

Environment:

• Windows 10 OS
• 4 Cores
• 16 gb RAM
• Workflow 8.6.1, Connect 1.6.1

Separation through Connect Output Settings:

• Data file = 10 000 records
• Time of completion = 00:25:44:347 (25 minutes, 44 seconds)

PDF Splitter on a Word change:

• Data file = 10 000 records
• Time of completion = 00:43:48:818 (43 minutes, 48 seconds)

Text scripts using ID as a selector VS using a text selector

When designing a Template, drag and drop is the easiest way to add data field (data selection, data extract). However, the default Text script generated by the application is using a text selector. To achieve a better performance, use an ID as selector. Testing like the following can be achieve from the Designer->Context->Profile Scripts. These are the benchmark values from those tests:

Environment:

• Windows 10 OS
• 4 Cores
• 16 gb RAM
• Connect 1.6.1

Text script using an ID selector:

• Fields = 9
• Repetition = 1000
• Time of completion = 3449.01 milliseconds (3 seconds, 449.01 milliseconds)

Text Script using a text selector:

• Fields = 9
• Repetition = 1000
• Time of completion = 5101.65 milliseconds (5 seconds, 101.65 milliseconds)

Always use the All In One plugin when possible

The All In One Connect plugin is very useful when no Job Creation, other than a single job, is required. Because of the parallelization of Merge engine, the speed of the datamapping step is increased. Add to that the new ability of not savings the record into the Connect DB, ability only available in the All in One plugin, it should always be the plugin to use when the situation, as described previously, is encountered.

Without being an exact formula, consider the following when setting the Merge engine in the Connect Server Configuration: “1 Merge engine per 1gb of RAM per physical core”

Again, this is to use as an approximation. It doesn’t take into account, other software running on the workstation, shared resources, etc… These are the benchmark values from those tests:

Environment:

• Windows 10 OS
• 4 Cores
• 16 gb RAM
• Workflow 8.6.1, Connect 1.6.1

Using the All In One Connect plugin:

• Data file = 10 000 records
• Time of completion = 00:12:23:978 (12 minutes, 23 seconds)

Using regular Connect plugins:

• Data file = 10 000 records
• Time of completion = 00:41:04:609 (41 minutes, 04 seconds)