harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chen Yang (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HARMONY-5612) [drlvm][gc] dacapo fails intermittently on 32-bit systems
Date Mon, 07 Apr 2008 05:28:24 GMT

    [ https://issues.apache.org/jira/browse/HARMONY-5612?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12586242#action_12586242
] 

Chen Yang commented on HARMONY-5612:
------------------------------------

Would you please try the new build? Recent GC check-in should have fixed the problem.
I have tried with the new build and finished 15 runs under Windows XP  and 15 Win32 runs under
Windows 2003 x64 R2. All finished successfully.
Thanks.


> [drlvm][gc] dacapo fails intermittently on 32-bit systems
> ---------------------------------------------------------
>
>                 Key: HARMONY-5612
>                 URL: https://issues.apache.org/jira/browse/HARMONY-5612
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: x86
>            Reporter: Pavel Pervov
>            Assignee: Xiao-Feng Li
>
> According to snapshot testing results dacapo intermittently fails on 32-bit architecture
(both Windows and Linux).
> run:
>      [exec] Apache Harmony Launcher : (c) Copyright 1991, 2008 The Apache Software Foundation
or its licensors, as applicable.
>      [exec] java version "1.5.0" 
>      [exec] pre-alpha : not complete or compatible
>      [exec] svn = r636792, (Mar 14 2008), Windows/ia32/msvc 1310, release build
>      [exec] http://harmony.apache.org
>      [exec] ===== DaCapo antlr starting =====
>      [exec] Running antlr on grammar antlr/java.g
>      [exec] ANTLR Parser Generator   Version 2.7.2   1989-2003 jGuru.com
>      [exec] ===== DaCapo antlr PASSED in 1266 msec =====
>      [exec] ===== DaCapo bloat starting =====
>      [exec] Optimized with: EDU.purdue.cs.bloat.optimize.Main -only EDU.purdue.cs.bloat.trans
-pre -dce -diva -prop -stack-alloc -peel-loops all -f EDU.purdue.cs.bloat.trans.ValueNumbering
.\scratch\optimizedcode
>      [exec] ===== DaCapo bloat PASSED in 6359 msec =====
>      [exec] ===== DaCapo chart starting =====
>      [exec] Done plotting: .\scratch/compress-tgt-ptr-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-src-ptr-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-dst-ptr-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-tgt-mut-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-src-mut-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-dst-mut-prf-time.pdf
>      [exec] Done plotting: .\scratch/compress-heap-comp-b.pdf
>      [exec] Done plotting: .\scratch/compress-heap-comp-a.pdf
>      [exec] Done plotting: .\scratch/compress-tgt-ptr-inf-time.pdf
>      [exec] Done plotting: .\scratch/compress-src-ptr-inf-time.pdf
>      [exec] Done plotting: .\scratch/compress-dst-ptr-inf-time.pdf
>      [exec] Done plotting: .\scratch/compress-tgt-mut-inf-time.pdf
>      [exec] Done plotting: .\scratch/compress-src-mut-inf-time.pdf
>      [exec] Done plotting: .\scratch/compress-dst-mut-inf-time.pdf
>      [exec] ===== DaCapo chart PASSED in 6656 msec =====
>      [exec] Out of Memory!
>      [exec] 
>      [exec] 
>      [exec] This application has requested the Runtime to terminate it in an unusual
way.
>      [exec] Please contact the application's support team for more information.
>      [exec] Result: 3
>      [copy] Copying 1 file to C:\cc\cc-suites\build\results\dacapo
>      [echo] Number of tests: (...........)
>      [echo] N of runned tests(...)
> The execution FAILED with the following exception:
> The following error occurred while executing this line:
> C:\cc\cc-suites\adaptors\dacapo\adaptor.xml:112: Not all workloads completed successfully
> Noticing "Out of Memory!" I tacked the message down to GC_GEN code.
> I was able to reproduce the failure on Linux automated snapshot testing host running
dacapo by hand.
> The steps to reproduce are:
> -> login to snapshot testing host
> -> export MY_USER_NAME=<your_name>
> -> mkdir /export/users/$MY_USER_NAME
> -> cd /export/users/$MY_USER_NAME
> -> export http_proxy=<proxy_host>:<proxy_port>
> -> wget http://people.apache.org/builds/harmony/snapshots/r637385/apache-harmony-jdk-r637385-linux-x86-32-snapshot.tar.gz
> -> tar xf http://people.apache.org/builds/harmony/snapshots/r637385/apache-harmony-jdk-r637385-linux-x86-32-snapshot.tar.gz
> -> export JAVA_HOME=/export/users/$MY_USER_NAME/harmony-jdk-r637385
> -> export PATH=$JAVA_HOME/bin:$PATH
> -> mkdir dacapo
> -> cd dacapo
> -> wget http://switch.dl.sourceforge.net/sourceforge/dacapobench/dacapo-2006-10-MR2.jar
> -> jar xf dacapo-2006-10-MR2.jar xalan.jar
> -> jar xf dacapo-2006-10-MR2.jar "antlr/*"
> -> mkdir test
> -> cd test
> -> export DACAPO_DIR=/export/users/$MY_USER_NAME/dacapo
> -> java -Xbootclasspath/p:$DACAPO_DIR/xalan.jar:$DACAPO_DIR -jar $DACAPO_DIR/dacapo-2006-10-MR2.jar
-s small antlr bloat chart eclipse fop hsqldb lusearch luindex pmd xalan jython
> The run fails after dacapo.chart (major collection fails). For me it fails 2 out of 3
runs.
> I tried to statically analize the code but didn't quite succeed. What I guessed is that
some of GC collector threads fail to set their Collector::result field to TRUE (or set it
to FALSE for some reason).

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message