Re: dacapo bloat benchmark fails on openjdk7
Tom Rodriguez wrote: I built a copy of hotspot from b33 on ubuntu with gcc 4.2 and I'm seeing the same failure you are. I believe this is the same issue covered by 6741642 which was found and fixed by the IcedTea folks. Applying the fix to a copy of b33 allowed dacapo to complete normally. Tom, thank for your help. I confirm that now all benchmarks run without errors. Regards, Luca
Re: dacapo bloat benchmark fails on openjdk7
Tom Rodriguez wrote: Could you be more specific about what bits you are running, i.e. which changeset of hotspot? The official openjdk bits run fine for me with at least b30-b33. It seems b33. [EMAIL PROTECTED]:~/mp/interfaces/OpenJDK/jdk7/hotspot$ hg log | head -n 6 changeset: 241:5b3b8a69f10f tag: tip user:xdono date:Thu Aug 14 09:26:23 2008 -0700 summary: Added tag jdk7-b33 for changeset 585535ec8a14 Bloat is not the only dacapo benchmark to fail, but also antlr, chart, eclipse, fop, jython, lusearch, pmd, xalan. hsqldb and luindex do not fail. I think it is definitively a problem of my OpenJDK build. Once the build finished, to use the new VM, is it necessary to set some environemnt variable or invoking the new java executable will suffice? I also put on pastebin my make sanity: http://pastebin.com/f52ad5606 Do you see something strange? Thanks for all your help. Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
Re: dacapo bloat benchmark fails on openjdk7
I built a copy of hotspot from b33 on ubuntu with gcc 4.2 and I'm seeing the same failure you are. I believe this is the same issue covered by 6741642 which was found and fixed by the IcedTea folks. Applying the fix to a copy of b33 allowed dacapo to complete normally. 6741642 is in the hotspot-comp repository but won't trickle up into a promoted build for a couple weeks so you may need to apply this by hand to your version. http://hg.openjdk.java.net/jdk7/hotspot-comp/hotspot/rev/fa4d1d240383 is the changeset of interest. tom On Aug 28, 2008, at 1:34 AM, Luca Martini wrote: Tom Rodriguez wrote: Could you be more specific about what bits you are running, i.e. which changeset of hotspot? The official openjdk bits run fine for me with at least b30-b33. It seems b33. [EMAIL PROTECTED]:~/mp/interfaces/OpenJDK/jdk7/hotspot$ hg log | head -n 6 changeset: 241:5b3b8a69f10f tag: tip user:xdono date:Thu Aug 14 09:26:23 2008 -0700 summary: Added tag jdk7-b33 for changeset 585535ec8a14 Bloat is not the only dacapo benchmark to fail, but also antlr, chart, eclipse, fop, jython, lusearch, pmd, xalan. hsqldb and luindex do not fail. I think it is definitively a problem of my OpenJDK build. Once the build finished, to use the new VM, is it necessary to set some environemnt variable or invoking the new java executable will suffice? I also put on pastebin my make sanity: http://pastebin.com/f52ad5606 Do you see something strange? Thanks for all your help. Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
dacapo bloat benchmark fails on openjdk7
Apologies for cross posting to the dacapobench-researchers and [EMAIL PROTECTED] lists. Hi all, I recently tried to execute the DaCapo benchmark on OpenJDK 7 and I found that bloat benchmark fails with the following exception: $ OpenJDK/jdk7/build/linux-i586/j2re-image/bin/java-jar ~/dacapo-2006-10-MR2.jar bloat = DaCapo bloat starting = ** Exception while optimizing transform(LEDU/purdue/cs/bloat/editor/MethodEditor;)Z of class EDU/purdue/cs/bloat/trans/CompactArrayInitializer String index out of range: -1802716504 java.lang.StringIndexOutOfBoundsException: String index out of range: -1802716504 at java.lang.String.getChars(String.java:860) at java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:408) at java.lang.StringBuilder.append(StringBuilder.java:136) at java.lang.StringBuilder.append(StringBuilder.java:132) at java.util.AbstractCollection.toString(AbstractCollection.java:439) at EDU.purdue.cs.bloat.codegen.RegisterAllocator$IGNode.toString(RegisterAllocator.java:570) at java.lang.String.valueOf(String.java:2838) at java.lang.StringBuffer.append(StringBuffer.java:236) at EDU.purdue.cs.bloat.util.Graph$4.next(Graph.java:983) at java.util.AbstractCollection.addAll(AbstractCollection.java:322) at EDU.purdue.cs.bloat.codegen.RegisterAllocator.init(RegisterAllocator.java:288) at EDU.purdue.cs.bloat.optimize.Main.bloatMethod(Main.java:1078) at EDU.purdue.cs.bloat.optimize.Main.editClass(Main.java:688) at EDU.purdue.cs.bloat.optimize.Main.main(Main.java:446) at dacapo.bloat.BloatHarness.iterate(BloatHarness.java:25) at dacapo.Benchmark.run(Benchmark.java:126) at dacapo.TestHarness.runBenchmark(TestHarness.java:302) at dacapo.TestHarness.main(TestHarness.java:242) at Harness.main(Harness.java:5) I really don't know if it's a DaCapo problem or if I missed something in the build process of OpenJDK. The version of the DaCapo is 2006-10-MR2. I built OpenJDK both using mercurial repository and source bundle, and the build runs fine but when executing the benchmark the same error come up. These are the options I used for building OpenJDK 7: LANG=C FINDBUGS_HOME= local path for findbugs ALT_BOOTDIR=/usr/lib/jvm/java-6-sun ALT_BINARY_PLUGS_PATH= local path for binary plugs ANT_HOME=/usr/share/ant ARCH_DADA_MODEL=32 Here there are some information on my machine: $ uname -a Linux 2.6.24-21-generic #1 SMP Tue Aug 12 13:37:22 UTC 2008 i686 GNU/Linux Any clues? Can anybody confirm such error? Thanks for your attention, Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
Re: [dacapobench-researchers] dacapo bloat benchmark fails on openjdk7
Luca, This is not a known problem with DaCapo, so on the face of it, this looks like a bug in your version of the VM (perhaps a bug in how you built it). FYI, we maintain 12 hourly performance regressions of a number of VMs here: http://dacapo.anu.edu.au/regression/perf/2006-10-MR2.html (the latest DaCapo release) and here: http://dacapo.anu.edu.au/regression/perf/head.html (the current DaCapo svn head). If there is interest and I can find the time, I may add the OpenJDK head to the list of tested benchmarks. Cheers, --Steve On 27/08/2008, at 6:34 PM, Luca Martini wrote: Apologies for cross posting to the dacapobench-researchers and [EMAIL PROTECTED] lists. Hi all, I recently tried to execute the DaCapo benchmark on OpenJDK 7 and I found that bloat benchmark fails with the following exception: $ OpenJDK/jdk7/build/linux-i586/j2re-image/bin/java-jar ~/dacapo-2006-10-MR2.jar bloat = DaCapo bloat starting = ** Exception while optimizing transform(LEDU/purdue/cs/bloat/editor/MethodEditor;)Z of class EDU/purdue/cs/bloat/trans/CompactArrayInitializer String index out of range: -1802716504 java.lang.StringIndexOutOfBoundsException: String index out of range: -1802716504 at java.lang.String.getChars(String.java:860) at java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:408) at java.lang.StringBuilder.append(StringBuilder.java:136) at java.lang.StringBuilder.append(StringBuilder.java:132) at java.util.AbstractCollection.toString(AbstractCollection.java:439) at EDU.purdue.cs.bloat.codegen.RegisterAllocator $IGNode.toString(RegisterAllocator.java:570) at java.lang.String.valueOf(String.java:2838) at java.lang.StringBuffer.append(StringBuffer.java:236) at EDU.purdue.cs.bloat.util.Graph$4.next(Graph.java:983) at java.util.AbstractCollection.addAll(AbstractCollection.java:322) at EDU .purdue .cs.bloat.codegen.RegisterAllocator.init(RegisterAllocator.java:288) at EDU.purdue.cs.bloat.optimize.Main.bloatMethod(Main.java:1078) at EDU.purdue.cs.bloat.optimize.Main.editClass(Main.java:688) at EDU.purdue.cs.bloat.optimize.Main.main(Main.java:446) at dacapo.bloat.BloatHarness.iterate(BloatHarness.java:25) at dacapo.Benchmark.run(Benchmark.java:126) at dacapo.TestHarness.runBenchmark(TestHarness.java:302) at dacapo.TestHarness.main(TestHarness.java:242) at Harness.main(Harness.java:5) I really don't know if it's a DaCapo problem or if I missed something in the build process of OpenJDK. The version of the DaCapo is 2006-10- MR2. I built OpenJDK both using mercurial repository and source bundle, and the build runs fine but when executing the benchmark the same error come up. These are the options I used for building OpenJDK 7: LANG=C FINDBUGS_HOME= local path for findbugs ALT_BOOTDIR=/usr/lib/jvm/java-6-sun ALT_BINARY_PLUGS_PATH= local path for binary plugs ANT_HOME=/usr/share/ant ARCH_DADA_MODEL=32 Here there are some information on my machine: $ uname -a Linux 2.6.24-21-generic #1 SMP Tue Aug 12 13:37:22 UTC 2008 i686 GNU/ Linux Any clues? Can anybody confirm such error? Thanks for your attention, Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] * - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ dacapobench-researchers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dacapobench-researchers
Re: [dacapobench-researchers] dacapo bloat benchmark fails on openjdk7
Steve, Thanks for your reply. I hope that someone of the OpenJDK folk could tell me if it's only a problem of my build or not. Regards, Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
Re: dacapo bloat benchmark fails on openjdk7
Could you be more specific about what bits you are running, i.e. which changeset of hotspot? The official openjdk bits run fine for me with at least b30-b33. tom On Aug 27, 2008, at 1:34 AM, Luca Martini wrote: Apologies for cross posting to the dacapobench-researchers and [EMAIL PROTECTED] lists. Hi all, I recently tried to execute the DaCapo benchmark on OpenJDK 7 and I found that bloat benchmark fails with the following exception: $ OpenJDK/jdk7/build/linux-i586/j2re-image/bin/java-jar ~/dacapo-2006-10-MR2.jar bloat = DaCapo bloat starting = ** Exception while optimizing transform(LEDU/purdue/cs/bloat/editor/MethodEditor;)Z of class EDU/purdue/cs/bloat/trans/CompactArrayInitializer String index out of range: -1802716504 java.lang.StringIndexOutOfBoundsException: String index out of range: -1802716504 at java.lang.String.getChars(String.java:860) at java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:408) at java.lang.StringBuilder.append(StringBuilder.java:136) at java.lang.StringBuilder.append(StringBuilder.java:132) at java.util.AbstractCollection.toString(AbstractCollection.java:439) at EDU.purdue.cs.bloat.codegen.RegisterAllocator $IGNode.toString(RegisterAllocator.java:570) at java.lang.String.valueOf(String.java:2838) at java.lang.StringBuffer.append(StringBuffer.java:236) at EDU.purdue.cs.bloat.util.Graph$4.next(Graph.java:983) at java.util.AbstractCollection.addAll(AbstractCollection.java:322) at EDU .purdue .cs.bloat.codegen.RegisterAllocator.init(RegisterAllocator.java:288) at EDU.purdue.cs.bloat.optimize.Main.bloatMethod(Main.java:1078) at EDU.purdue.cs.bloat.optimize.Main.editClass(Main.java:688) at EDU.purdue.cs.bloat.optimize.Main.main(Main.java:446) at dacapo.bloat.BloatHarness.iterate(BloatHarness.java:25) at dacapo.Benchmark.run(Benchmark.java:126) at dacapo.TestHarness.runBenchmark(TestHarness.java:302) at dacapo.TestHarness.main(TestHarness.java:242) at Harness.main(Harness.java:5) I really don't know if it's a DaCapo problem or if I missed something in the build process of OpenJDK. The version of the DaCapo is 2006-10- MR2. I built OpenJDK both using mercurial repository and source bundle, and the build runs fine but when executing the benchmark the same error come up. These are the options I used for building OpenJDK 7: LANG=C FINDBUGS_HOME= local path for findbugs ALT_BOOTDIR=/usr/lib/jvm/java-6-sun ALT_BINARY_PLUGS_PATH= local path for binary plugs ANT_HOME=/usr/share/ant ARCH_DADA_MODEL=32 Here there are some information on my machine: $ uname -a Linux 2.6.24-21-generic #1 SMP Tue Aug 12 13:37:22 UTC 2008 i686 GNU/ Linux Any clues? Can anybody confirm such error? Thanks for your attention, Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
Re: dacapo bloat benchmark fails on openjdk7
I concur with Tom. I went back to b25 and didn't have any problems running dacapo bloat with JDK7. -- Chuck On 08/27/08 11:35, Tom Rodriguez wrote: Could you be more specific about what bits you are running, i.e. which changeset of hotspot? The official openjdk bits run fine for me with at least b30-b33. tom On Aug 27, 2008, at 1:34 AM, Luca Martini wrote: Apologies for cross posting to the dacapobench-researchers and [EMAIL PROTECTED] lists. Hi all, I recently tried to execute the DaCapo benchmark on OpenJDK 7 and I found that bloat benchmark fails with the following exception: $ OpenJDK/jdk7/build/linux-i586/j2re-image/bin/java-jar ~/dacapo-2006-10-MR2.jar bloat = DaCapo bloat starting = ** Exception while optimizing transform(LEDU/purdue/cs/bloat/editor/MethodEditor;)Z of class EDU/purdue/cs/bloat/trans/CompactArrayInitializer String index out of range: -1802716504 java.lang.StringIndexOutOfBoundsException: String index out of range: -1802716504 at java.lang.String.getChars(String.java:860) at java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:408) at java.lang.StringBuilder.append(StringBuilder.java:136) at java.lang.StringBuilder.append(StringBuilder.java:132) at java.util.AbstractCollection.toString(AbstractCollection.java:439) at EDU.purdue.cs.bloat.codegen.RegisterAllocator$IGNode.toString(RegisterAllocator.java:570) at java.lang.String.valueOf(String.java:2838) at java.lang.StringBuffer.append(StringBuffer.java:236) at EDU.purdue.cs.bloat.util.Graph$4.next(Graph.java:983) at java.util.AbstractCollection.addAll(AbstractCollection.java:322) at EDU.purdue.cs.bloat.codegen.RegisterAllocator.init(RegisterAllocator.java:288) at EDU.purdue.cs.bloat.optimize.Main.bloatMethod(Main.java:1078) at EDU.purdue.cs.bloat.optimize.Main.editClass(Main.java:688) at EDU.purdue.cs.bloat.optimize.Main.main(Main.java:446) at dacapo.bloat.BloatHarness.iterate(BloatHarness.java:25) at dacapo.Benchmark.run(Benchmark.java:126) at dacapo.TestHarness.runBenchmark(TestHarness.java:302) at dacapo.TestHarness.main(TestHarness.java:242) at Harness.main(Harness.java:5) I really don't know if it's a DaCapo problem or if I missed something in the build process of OpenJDK. The version of the DaCapo is 2006-10-MR2. I built OpenJDK both using mercurial repository and source bundle, and the build runs fine but when executing the benchmark the same error come up. These are the options I used for building OpenJDK 7: LANG=C FINDBUGS_HOME= local path for findbugs ALT_BOOTDIR=/usr/lib/jvm/java-6-sun ALT_BINARY_PLUGS_PATH= local path for binary plugs ANT_HOME=/usr/share/ant ARCH_DADA_MODEL=32 Here there are some information on my machine: $ uname -a Linux 2.6.24-21-generic #1 SMP Tue Aug 12 13:37:22 UTC 2008 i686 GNU/Linux Any clues? Can anybody confirm such error? Thanks for your attention, Luca -- * Ing. Luca Martini Dipartimento di Ing. dell'Informazione Università di Pisa Via Diotisalvi, 2, I-56126 Pisa, Italy Tel: +39 050 2217 465 Fax: +39 050 2217 600 e-mail: [EMAIL PROTECTED] *
