You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Xiao-Feng Li (JIRA)" <ji...@apache.org> on 2007/01/06 15:00:27 UTC
[jira] Created: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
[DRLVM][GCv5] parallel generational or non-generational collection
------------------------------------------------------------------
Key: HARMONY-2945
URL: https://issues.apache.org/jira/browse/HARMONY-2945
Project: Harmony
Issue Type: Improvement
Components: DRLVM
Environment: Windows and Linux IA32
Reporter: Xiao-Feng Li
Attachments: GCv5-parallelization-1-5.patch
Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
There are two dependences:
1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
Thanks,
xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (HARMONY-2945) [DRLVM][GCv5] parallel
generational or non-generational collection
Posted by "weldon washburn (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12463716 ]
weldon washburn commented on HARMONY-2945:
------------------------------------------
this patch does not work on the current svn HEAD. Please update
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Assigned: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "weldon washburn (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
weldon washburn reassigned HARMONY-2945:
----------------------------------------
Assignee: weldon washburn
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Closed: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "weldon washburn (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
weldon washburn closed HARMONY-2945.
------------------------------------
Resolution: Fixed
01....patch was committed. The change to gc_platform.h for MAP_FIXED was applied manually. Passes on winxp and rhel4 w/ gcc 4.0.2
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: 01-GCv5-parallelization-1-11.patch, 02-GCv5-parallelization-1-11.patch, GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: GCv5-parallelization-1-5.patch
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Attachments: GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (HARMONY-2945) [DRLVM][GCv5] parallel
generational or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12463791 ]
Xiao-Feng Li commented on HARMONY-2945:
---------------------------------------
Hi, Weldon, updated patch is attached as "01-GCv5-parallelization-1-11.patch". It's built against r495093.
To save your commit efforts, you probably can apply this patch and H2965 together for commit.
Thanks,
xiaofeng
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: 01-GCv5-parallelization-1-11.patch, GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: 02-GCv5-parallelization-1-11.patch
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: 01-GCv5-parallelization-1-11.patch, 02-GCv5-parallelization-1-11.patch, GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: 01-GCv5-parallelization-1-11.patch
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: 01-GCv5-parallelization-1-11.patch, GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (HARMONY-2945) [DRLVM][GCv5] parallel
generational or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12463849 ]
Xiao-Feng Li commented on HARMONY-2945:
---------------------------------------
A minor fix in the patch: change FIXED to MAP_FIXED in line 2313.
02-GCv5-parallelization-1-11.patch is the one has this fix. Thanks.
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: 01-GCv5-parallelization-1-11.patch, 02-GCv5-parallelization-1-11.patch, GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: (was: GCv5-parallelization-1-5.patch)
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Attachments: GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: GCv5-parallelization-1-5-02.patch
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Updated: (HARMONY-2945) [DRLVM][GCv5] parallel generational
or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li updated HARMONY-2945:
----------------------------------
Attachment: GCv5-parallelization-1-5.patch
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Attachments: GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] Commented: (HARMONY-2945) [DRLVM][GCv5] parallel
generational or non-generational collection
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-2945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12463211 ]
Xiao-Feng Li commented on HARMONY-2945:
---------------------------------------
Dear committer, please use "GCv5-parallelization-1-5-02.patch" for patching.
Thanks,
xiaofeng
> [DRLVM][GCv5] parallel generational or non-generational collection
> ------------------------------------------------------------------
>
> Key: HARMONY-2945
> URL: https://issues.apache.org/jira/browse/HARMONY-2945
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Windows and Linux IA32
> Reporter: Xiao-Feng Li
> Assigned To: weldon washburn
> Attachments: GCv5-parallelization-1-5-02.patch, GCv5-parallelization-1-5.patch
>
>
> Hi, attached is a patch for GCv5 which achieves good parallelization scalability in real SMP machine. It can works in generational or nongenerational mode. I have tested it on a Intel Tulsa platform which has four Pentium-D dual-core processors (totally 8 cores). It performs pretty well. With more processors and larger heap size, this parallel GC can demonstrate more benefits.
> There are two dependences:
> 1. To fully exploit this GC's adavantage, helper inlining should be enabled for it. It depends on Harmony-2652 which inlines write barrier.
> 2. Some JIT optimizations do not implement write barrer for reference field update. These opts have to be disabled to make the generational mode work.
> At the same time before the two dependences are resolved, people still can play with it in non-generational mode. But the performance will be impacted seriously without helper inlinings.
> Dear commiter, please help to integrate it into Harmony SVN. Please apply the patch at working_vm directory.
> Thanks,
> xiaofeng
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira