1) Architecture option - 1 : Based on OpenGL Rendering Pipeline implementation of Java2d uses a single threaded RenderQueue to process graphics commands.
2) Architecture option - 2 : Based X11Render Pipeline implementation of Java2d uses direct rendering without single threaded RenderQueue.
This task is to evaluate differences/pros/cons of above options.
Things to consider would be :
1) Performance sampling of render calls
2) possible limitations in implementation using Metal APIs
- blocks
-
JDK-8224046 ☂ Metal Graphics API Implementation
- Resolved
- is blocked by
-
JDK-8227664 Disable additional RQ.flushNow() that hogs the AWT-EventQueue thread
- Closed
- relates to
-
JDK-8230641 MTLRenderer_FillSpans method needs to be optimized by reducing total number of drawPrimitives calls
- Resolved
-
JDK-8230837 MTLRenderer_DrawParallelogram method - can be optimized by combining 4 drawPrimitives calls
- Resolved
-
JDK-8231614 Avoid passing of redundant vertex 'z' data to vertex shader
- Resolved
-
JDK-8231456 Check possibility of usage of shaders to generate vertices in required format
- Closed
-
JDK-8231883 Combine adjacent primitive draw calls to issue a single draw call with multiple data
- Closed