Planning for JavaFX.next
bourges.laurent at gmail.com
Fri Dec 9 09:01:09 UTC 2016
I would like to promote performance & quality improvements on the
rasterizer that I could do in jdk10 within a tiger team focused on fx
- multithreaded rendering (Marlin is thread-safe) ie the javafx could do
parallel rendering of layers or shapes
- software rasterizers: optimize inner / outer strokes, clipping in
particular for dashes, intersection of curves with scanlines (integer maths)
- rendering quality: use exact pixel coverage (agg approach), gamma
- in general: optimize array/buffer copies and memory allocations in the
complete javafx stack (jvmstat + profiling)
- use value types and more generally, enjoy vahalla, panama integration to
improve java and c code in the rendering pipeline
- gpu / shader upgrade ?
Finally all these items represent an important workload that need both
Oracle and OSS contributors = manpower.
The main issue consists in building a team with new contributors. Maybe
some Oracle or OpenJDK sponsorship is needed to fund such team ie donate
some money for some contributions... at least to organise meetings.
Let me cite the GNU philosophy:
Thus, “free software” is a matter of liberty, not price. To understand the
concept, you should think of “free” as in “free speech,” not as in “free
More information about the openjfx-dev