It's J9. If DRLVM can tap on the improvement on performance. I am ok not to patch this. Thanks for your comments!
On Wed, Jan 7, 2009 at 2:58 AM, Aleksey Shipilev (JIRA) <[email protected]>wrote: > > [ > https://issues.apache.org/jira/browse/HARMONY-6056?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12661257#action_12661257] > > Aleksey Shipilev commented on HARMONY-6056: > ------------------------------------------- > > Just flashed into my mind. > Kevin, what javac/JVM you use to make the performance runs? Is it J9 or > DRLVM? Are there any steps to reproduce the result? > > AFAIK, DRLVM does not substitute the "String s2 = (String)s1 + (int)v1;" to > "new String(s1, v1)", is it J9 who does this thing? If so the another > benefit of generalizing your solution in Integer.toString(), because DRLVM > can then tap on your probably-optimized radix-10 conversion when > concatenating strings via StringBuilder. > > > [classlib][jit][opt][performance] Optimize heap allocations in > String(String, int) constructor for JIT > > > ------------------------------------------------------------------------------------------------------ > > > > Key: HARMONY-6056 > > URL: https://issues.apache.org/jira/browse/HARMONY-6056 > > Project: Harmony > > Issue Type: Improvement > > Components: Classlib > > Affects Versions: 5.0M8 > > Reporter: Kevin Zhou > > Assignee: Tim Ellison > > Fix For: 5.0M9 > > > > Attachments: HARMONY-6056.diff > > > > Original Estimate: 24h > > Remaining Estimate: 24h > > > > This patch optimizes heap allocations in String(String, int) constructor > for JIT performance improvement. > > -- > This message is automatically generated by JIRA. > - > You can reply to this email to add a comment to the issue online. > >
