Github user afs commented on the issue:
https://github.com/apache/jena/pull/209
As mentioned earlier, there was a change at 3.1.1 to reduce the memory foot
print. It is a tradeoff - space vs time - for the default setup. If you want
pure speed, use a `ParserProfile` with a different `RDFFactory` -
`RDFFactorStd` is the non-caching one, `RDFFactorCaching` is the caching one.
What would be useful to make any progress is to understand what you are
trying to do. Raw parser performance is not necessarily the right metric
because data is used for a purpose. The figures you have don't match what I
measured, and remeasured for this issue, and without seeing the test and data,
it is not possible to draw any insight.
If you wish to explore improvements, it would be useful to use a Java
profiler on the code. `RDFFactorCaching is the place to look. It use a cache
from Google Guava so maybe there is a better choice there are no concurrency
issues.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
