OSDC Performance


On a conversion test of 1300 PowerPoint documents (2.4GB of files, 29700 pages) OSDC completed in just under 4 minutes and 20 seconds (114.23 pages/second)*. A very simple Python script was written to call OSDC on each document to convert to PDF. The jobs ran in parallel with a 5-worker thread pool. The performance was measured using the ‘time' utility. Using the System Monitor utility the memory usage did not exceed about 1GB during the test.
Performance scales linearly with the number of cores and clock speed (assuming equivalent processor architectures).

Test Documents

For data, 13 PowerPoint documents were taken and copied each 99 times to produce 1300 total files.
The documents came from this site: http://www1.idc.ac.il/tecs/plan.html

1300 PowerPoint documents (2.4GB of files, 29700 pages)


Debian GNU/Linux 8 (jessie) 64-bit
CPU: Intel Xeon X3430 (Quad-core @ 2.4GHz, Turbo disabled)


It's also important to note that the os.system() method was used to invoke the command-line. Based on prior tests you can expect as much as a 5% improvement when using the C/C++ or Java JNI interface due to the overhead of creating and destroying a process for each invocation.

*Performance will vary depending on environment, hardware & other factors.