-
Bug
-
Resolution: Fixed
-
P3
-
8
There is significant performance regression in Controls.TableView-XmasTree observed on all platforms:
Winxp-Mid: 69% (31 fps)
Win7-High: 70% (44 fps)
Mac-Mid: 79% (26 fps)
Note, other Controls benchmarks like Checkbbox, RadioButton show up to 50% improvement.
8.0-graphics-scrum-1836 contains the following changes:
1.RT-33597: TriangleMeshValidationTest unit test fails on SW pipeline
Reviewed-by: kcr — Yao Wang / detail
2. RT-33562: Java_com_sun_prism_es2_GLContext_nReadPixelsByte/Int need to check null pointer
Reviewed-by: kcr, cyang — Yao Wang / detail
3.RT-30158: Investigate String Measurement in FX (cache results, call less, ...)
Reviewed-By: Steve Northover, Phil Race — Felipe Heidrich / detail
4.RT-33203: Regression: incorrect size for OPEN_HAND and CLOSED_HAND cursors on Windows
Reviewed-by: jgodinez
Contributed-by: Mo Chicharro <mo.chicharro@oracle.com> — kcr / detail
Most probably the regression is caused byRT-30158 which changed files:
modules/graphics/src/main/java/com/sun/javafx/font/PrismFontFactory.java
modules/graphics/src/main/java/com/sun/javafx/text/LayoutCache.java
modules/graphics/src/main/java/com/sun/javafx/text/PrismTextLayout.java
I did run Controls.TableView-XmasTree benchmark with com.sun.javafx.font/text classes from build 1835 and
performance returned back. So, the regression is caused byRT-30158 fix.
Steps to run Controls.TableView-XmasTree:
> cd JFX_WS/tests/performance/Controls/
> ant
> java -Xmx512m -Djavafx.animation.fullspeed=true
-cp "JFX_HOME/rt/lib/ext/jfxrt.jar;./dist/Controls.jar;../FXBenchmark/dist/FXBenchmark.jar;../../../import/benchmarks-2.1.1/benchmarks-2.1.1.jar"
jrockit.bm.Main controls.bm.XmasTreeTableViewBenchmark -i 1 -wt 0 -tr 60
-win-width 1280 -win-height 1024 -col-width 25 -row-height 10 -rows 90 -columns 50 -batch-size 10
Controls.TableView-XmasTree is part of PRD, so the issue is need to be fixed.
Winxp-Mid: 69% (31 fps)
Win7-High: 70% (44 fps)
Mac-Mid: 79% (26 fps)
Note, other Controls benchmarks like Checkbbox, RadioButton show up to 50% improvement.
8.0-graphics-scrum-1836 contains the following changes:
1.
Reviewed-by: kcr — Yao Wang / detail
2. RT-33562: Java_com_sun_prism_es2_GLContext_nReadPixelsByte/Int need to check null pointer
Reviewed-by: kcr, cyang — Yao Wang / detail
3.
Reviewed-By: Steve Northover, Phil Race — Felipe Heidrich / detail
4.
Reviewed-by: jgodinez
Contributed-by: Mo Chicharro <mo.chicharro@oracle.com> — kcr / detail
Most probably the regression is caused by
modules/graphics/src/main/java/com/sun/javafx/font/PrismFontFactory.java
modules/graphics/src/main/java/com/sun/javafx/text/LayoutCache.java
modules/graphics/src/main/java/com/sun/javafx/text/PrismTextLayout.java
I did run Controls.TableView-XmasTree benchmark with com.sun.javafx.font/text classes from build 1835 and
performance returned back. So, the regression is caused by
Steps to run Controls.TableView-XmasTree:
> cd JFX_WS/tests/performance/Controls/
> ant
> java -Xmx512m -Djavafx.animation.fullspeed=true
-cp "JFX_HOME/rt/lib/ext/jfxrt.jar;./dist/Controls.jar;../FXBenchmark/dist/FXBenchmark.jar;../../../import/benchmarks-2.1.1/benchmarks-2.1.1.jar"
jrockit.bm.Main controls.bm.XmasTreeTableViewBenchmark -i 1 -wt 0 -tr 60
-win-width 1280 -win-height 1024 -col-width 25 -row-height 10 -rows 90 -columns 50 -batch-size 10
Controls.TableView-XmasTree is part of PRD, so the issue is need to be fixed.