Community
Participate
Working Groups
ActivateJavaEditorTest#testActivateEditor() (Activate 30 Java editors) has a small regression on all platforms. See http://download.eclipse.org/eclipse/downloads/drops/I20110329-0800/performance/epwin2/Scenario1.html OpenJavaEditorStartupTest#testJavaEditorStartup() also shows a smallish regression. See http://download.eclipse.org/eclipse/downloads/drops/I20110329-0800/performance/epwin3/Scenario53.html This could possibly be same as bug 337007.
I can reproduce this one on my machine. M3 Scenario 'org.eclipse.jdt.text.tests.performance.ActivateJavaEditorTest#testActivateEditor()' (average over 5 samples): Used Java Heap: 9.78M (95% in [-16.96M, 36.52M]) Measurable effect: 38.52M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6401) Working Set: -238.4K (95% in [-15M, 14.54M]) Measurable effect: 21.28M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6400) Committed: -26.4K (95% in [-15.08M, 15.03M]) Measurable effect: 21.69M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6401) Working Set Peak: 0 (95% in [0, 0]) Elapsed Process: 6.22s (95% in [6.08s, 6.35s]) Measurable effect: 192ms (1.8 SDs) Kernel time: 2.04s (95% in [1.88s, 2.2s]) Measurable effect: 229ms (1.8 SDs) (required sample size for an effect of 5% of mean: 26) Page Faults: 3.95K (95% in [877, 7.02K]) Measurable effect: 4.42K (1.8 SDs) (required sample size for an effect of 5% of mean: 2512) CPU Time: 6.97s (95% in [6.67s, 7.26s]) Measurable effect: 425ms (1.8 SDs) (required sample size for an effect of 5% of mean: 8) GDI Objects: 0 (95% in [0, 0]) I20110406-1223 Scenario 'org.eclipse.jdt.text.tests.performance.ActivateJavaEditorTest#testActivateEditor()' (average over 5 samples): Used Java Heap: 2.38M (95% in [-1.8M, 6.56M]) Measurable effect: 6.02M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6401) Working Set: -407.2K (95% in [-4.92M, 4.13M]) Measurable effect: 6.52M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6400) Committed: 36K (95% in [-5.43M, 5.5M]) Measurable effect: 7.87M (1.8 SDs) (required sample size for an effect of 5% of stdev: 6400) Working Set Peak: 0 (95% in [0, 0]) Elapsed Process: 6.58s (95% in [6.51s, 6.65s]) Measurable effect: 102ms (1.8 SDs) Kernel time: 2.25s (95% in [2.09s, 2.41s]) Measurable effect: 225ms (1.8 SDs) (required sample size for an effect of 5% of mean: 21) Page Faults: 2.84K (95% in [2.26K, 3.42K]) Measurable effect: 832 (1.8 SDs) (required sample size for an effect of 5% of mean: 173) CPU Time: 7.23s (95% in [7.06s, 7.4s]) Measurable effect: 243ms (1.8 SDs) GDI Objects: 0 (95% in [0, 0])
I ran the test with diff versions(tags) of o.e.jdt.ui plugin and everything else same. v20110201-0800 (i.e. < v20110208-800) CPU Time: 7.00s v20110208-0800 CPU Time: 7.16s v20110215-0800 (i.e. > v20110208-800) CPU Time: 7.23s I looked at the changes from v20110201-0800 to v20110215-0800 and nothing really jumps out.
The only change code between v20110201-0800 and v20110215-0800 which is executed by the test is in org.eclipse.jdt.internal.ui.javaeditor.ASTProvider. This change was fix for bug 322914: [implementation] deadlock in ASTProvider when copy paste in CU editor. When I reverted this change in o.e.jdt.ui v20110215-0800 I see the same performance numbers as with o.e.jdt.ui v20110201-0800. I guess this is a valid fix and we cannot do anything here.
(In reply to comment #3) > The only change code between v20110201-0800 and v20110215-0800 which is > executed by the test is in org.eclipse.jdt.internal.ui.javaeditor.ASTProvider. > > This change was fix for bug 322914: [implementation] deadlock in ASTProvider > when copy paste in CU editor. When I reverted this change in o.e.jdt.ui > v20110215-0800 I see the same performance numbers as with o.e.jdt.ui > v20110201-0800. > > I guess this is a valid fix and we cannot do anything here. Please mark the test gray with a comment pointing to the bug fix.
(In reply to comment #4) > Please mark the test gray with a comment pointing to the bug fix. Done.