You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Chen Yang (JIRA)" <ji...@apache.org> on 2007/11/19 07:43:43 UTC
[jira] Created: (HARMONY-5146) [drlvm][gc]Allocation Prefetch in
Garbage Collector
[drlvm][gc]Allocation Prefetch in Garbage Collector
---------------------------------------------------
Key: HARMONY-5146
URL: https://issues.apache.org/jira/browse/HARMONY-5146
Project: Harmony
Issue Type: Improvement
Components: DRLVM
Environment: Win32/Win64/Linux32 have already been verified
Reporter: Chen Yang
Allocation Prefetch is known to be effective in improving the application performance, this patch is based on the study on current allocation implementation in garbage collector and has already been evaluated. Some options are provided:
-XX:gc.prefetch=true/false ; to enable or disable the allocation prefetch, default false
-XX:gc.zeroing_size=N ; to specify the number of bytes in group memset, default 256
-XX:gc.prefetch_stride=N ; to specify the number of bytes of prefetch stride, default 64
-XX:gc.prefetch_distance=N ; to specify the number of bytes to be prefetched, default 1024
To get the best performance, it is expected to select appropriate parameters.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-5146) [drlvm][gc]Allocation Prefetch in
Garbage Collector
Posted by "Chen Yang (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-5146?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chen Yang updated HARMONY-5146:
-------------------------------
Attachment: alloc_prefetch.patch
Alloc prefetch support for GC of DRLVM
> [drlvm][gc]Allocation Prefetch in Garbage Collector
> ---------------------------------------------------
>
> Key: HARMONY-5146
> URL: https://issues.apache.org/jira/browse/HARMONY-5146
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Win32/Win64/Linux32 have already been verified
> Reporter: Chen Yang
> Attachments: alloc_prefetch.patch
>
>
> Allocation Prefetch is known to be effective in improving the application performance, this patch is based on the study on current allocation implementation in garbage collector and has already been evaluated. Some options are provided:
> -XX:gc.prefetch=true/false ; to enable or disable the allocation prefetch, default false
> -XX:gc.zeroing_size=N ; to specify the number of bytes in group memset, default 256
> -XX:gc.prefetch_stride=N ; to specify the number of bytes of prefetch stride, default 64
> -XX:gc.prefetch_distance=N ; to specify the number of bytes to be prefetched, default 1024
> To get the best performance, it is expected to select appropriate parameters.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Closed: (HARMONY-5146) [drlvm][gc]Allocation Prefetch in
Garbage Collector
Posted by "Xiao-Feng Li (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-5146?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiao-Feng Li closed HARMONY-5146.
---------------------------------
Resolution: Fixed
Assignee: Xiao-Feng Li
committed in r596220
> [drlvm][gc]Allocation Prefetch in Garbage Collector
> ---------------------------------------------------
>
> Key: HARMONY-5146
> URL: https://issues.apache.org/jira/browse/HARMONY-5146
> Project: Harmony
> Issue Type: Improvement
> Components: DRLVM
> Environment: Win32/Win64/Linux32 have already been verified
> Reporter: Chen Yang
> Assignee: Xiao-Feng Li
> Attachments: alloc_prefetch.patch
>
>
> Allocation Prefetch is known to be effective in improving the application performance, this patch is based on the study on current allocation implementation in garbage collector and has already been evaluated. Some options are provided:
> -XX:gc.prefetch=true/false ; to enable or disable the allocation prefetch, default false
> -XX:gc.zeroing_size=N ; to specify the number of bytes in group memset, default 256
> -XX:gc.prefetch_stride=N ; to specify the number of bytes of prefetch stride, default 64
> -XX:gc.prefetch_distance=N ; to specify the number of bytes to be prefetched, default 1024
> To get the best performance, it is expected to select appropriate parameters.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.