PDF to Word conversion taking time\resources

Hi,



I have developed a pdf to MSFT conversion application, while testing this discover the time to do the conversion takes a lot of time. See screenshots ~33 seconds for <10 pages.



The second concern I had was the amour of CPU resources it occupies during the 33 seconds while the process is rendering (1 core). I am concern if this solution can be scalable for dozens of concurrent active users.



I am also attaching the PDF for you to test the conversion.

Thanks

Hi Jeremy,


Thanks for contacting support.

I have tested the conversion of PDF file to DOC format using Aspose.Pdf for Java 9.3.1 in Eclipse Juno application and have managed to reproduce the same issue that time taken for conversion is around 37 seconds. For the sake of correction, I
have logged it in our issue tracking system as PDFNEWJAVA-34470. We will investigate this
issue in details and will keep you updated on the status of a correction.

We apologize for your inconvenience.

Hi Jeremy,


I have also observed that during PDF to DOC conversion, the CPU utilization hikes to 85% and also memory utilization is hiked by 600MB. .
For the sake of correction, I have separately logged it in our issue tracking system as PDFNEWJAVA-34471. We will
investigate this issue in details and will keep you updated on the status of a
correction.

We apologize for your inconvenience.

Hi,

Can you tell me what is the ETA of fixing this performance issue?

We just purchase ASPOSE.total early September. OrderID 140902002127.

Hi Jeremy,


Thanks for your patience,.

As we recently have been able to notice this issue, and until or
unless we have investigated and have figured out the actual reasons of this
problem, we might not be able to share any timelines by which this problem will
be resolved.<o:p></o:p>

However, as soon as we have made some significant progress towards the resolution of this issue, we would be more than happy to update you with the status of correction. Please be patient and spare us little time. Your patience and comprehension is greatly appreciated in this regard.

The issues you have found earlier (filed as PDFNEWJAVA-34470) have been fixed in Aspose.Pdf for Java 12.0.0.


This message was posted using Notification2Forum from Downloads module by Aspose Notifier.

Hi Jeremy,


Thanks for your patience. Our product team has investigated the issue and it seems the performance has been improved in result of other fixes in later version.

However we have noticed difference in performance of i586 and x64 JDK version. The problem can be in using i586 version of JDK in x64 environment. When we used correct JDK the performance was increased twice. So it is recommended to use the latest JDK build for x64,

Furthermore, please note for the best quality conversion in the latest version we have changed default image conversion resolution to 300 dpi. To get a fair comparison with previous versions of the conversion or further reduce the conversion time - you can set this same dpi, which was set up by default in previous versions (192) or less as following.


com.aspose.pdf.Document pdfDocument = new com.aspose.pdf.Document(myDir+“untitled+(13).pdf”);<o:p></o:p>

DocSaveOptions opt = new DocSaveOptions(); <o:p></o:p>

opt.setImageResolutionX(192);<o:p></o:p>

opt.setImageResolutionY(192);<o:p></o:p>

pdfDocument.save(myDir+“untitled+(13).doc”,opt);


Best Regards,

Hi,


Thanks for your patience.

We have further investigated the earlier reported issue PDFNEWJAVA-34471 and it does not seem to be a bug in our API. JVM uses Heap Memory when its necessary, therefore the garbage collector is not used too often.

This code works fine even if we limit the memory to 128 mb (-Xmx128m -Xms128m). Also the processor usage is not a problem since we cannot manually handle the processor usage. But the OS and JVM takes free computer resources to perform the programs calculations. In earlier shared file, we try to make internal conversion for images and we need to read the required fonts, therefore we have to use different matrices to correctly scale content on pages to get suitable result. This process takes some resources and its a normal behavior.