Page 1 of 1

Performance of doxypress versus doxygen

Posted: Wed Nov 15 2017 2:54 pm
by marlowa
Hello everyone,

Has anyone compared the performance of doxypress versus doxygen for a non-trivial project? I was wondering. Surely a from scratch rewrite using a better foundation should give something that is at least no slower.

I am trying to tune an overnight build process and one area I am considering is where we currently run doxygen. If doxypress gives similar results and runs faster this would be a good argument to switching to it.

Re: Performance of doxypress versus doxygen

Posted: Wed Nov 15 2017 7:15 pm
by barbara
We have run some benchmarks when generating the documentation for the CopperSpice API. Over the span of the project we have seen a two fold improvement in the speed as well as better HTML documentation output. The combination of faster and more accurate seems like a good direction for DoxyPress.

Another project we use as a test case is generating the docs for clang itself. We have noted about a 3x speed improvement.

A user in South America contacted us with a request for a minor change so he could run a very large project. He noted the speed improvement in running DoxyPress was massive.

We are looking at other ways to improve the speed and we know that adding UTF-8 support in CopperSpice should translate to a performance improvement in DoxyPress. There are also several other things we are considering to fine tune DoxyPress. One of which is modifying how data is marshaled to disk.

Barbara