[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16478803#comment-16478803
]
ASF GitHub Bot commented on IO-468:
---
Github user berndhopp closed the pull request at:
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14332404#comment-14332404
]
Thomas Neidhart commented on IO-468:
Well, my arguments against this patch have certainly
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14332363#comment-14332363
]
Bernd Hopp commented on IO-468:
---
as far as I can see, the numbers show two things:
- fixed arra
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14325626#comment-14325626
]
Thomas Neidhart commented on IO-468:
The Error is the score inside the 99.9% confidence in
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14325308#comment-14325308
]
Sebb commented on IO-468:
-
I assume the Error column is the margin of error?
This increases dramatical
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14314993#comment-14314993
]
Thomas Neidhart commented on IO-468:
I think it is a pretty easy trap to fall into and man
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14314989#comment-14314989
]
Thomas Neidhart commented on IO-468:
ok, I will see if I find some time the next days to c
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14313901#comment-14313901
]
Bernd Hopp commented on IO-468:
---
@Thomas Neidhart
Thank you for your effort, could you share yo
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14313023#comment-14313023
]
Bernd Hopp commented on IO-468:
---
this is interesting. Awkward for me of course to not have spott
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312870#comment-14312870
]
Thomas Neidhart commented on IO-468:
Some benchmarks I did with my own test harness. The n
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312325#comment-14312325
]
Bernd Hopp commented on IO-468:
---
Thomas, I DID read the material. I responded to every point mad
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312252#comment-14312252
]
Thomas Neidhart commented on IO-468:
I am sorry, but you got so many pointers to the actua
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312248#comment-14312248
]
Bernd Hopp commented on IO-468:
---
So the "compromise" I made was to use WeakReference? And your p
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312219#comment-14312219
]
Joerg Schaible commented on IO-468:
---
Sorry, but Sebb is not alone here. We're very defensive
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312215#comment-14312215
]
Bernd Hopp commented on IO-468:
---
I will try that
> Avoid allocating memory for method internal
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312189#comment-14312189
]
Bernd Hopp commented on IO-468:
---
[~s...@apache.org], you are one of the persons responsible for
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312153#comment-14312153
]
Sebb commented on IO-468:
-
When I said that buffer allocation is not slow, I meant that it is not slow
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312141#comment-14312141
]
Bernd Hopp commented on IO-468:
---
allright, thank you
> Avoid allocating memory for method inter
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312138#comment-14312138
]
Mark Thomas commented on IO-468:
@[~bernd_hopp] Thsi issue with the current implementation is
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312133#comment-14312133
]
Bernd Hopp commented on IO-468:
---
@Sebb
"think the benchmark is not really the main issue here,
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311673#comment-14311673
]
Sebb commented on IO-468:
-
I think the benchmark is not really the main issue here, although it will b
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311594#comment-14311594
]
Emmanuel Bourg commented on IO-468:
---
Bernd I suggest you rewrite your micro benchmark with J
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311587#comment-14311587
]
Bernd Hopp commented on IO-468:
---
[~javaguru]
Thank you for your comment. By 'the patch as curren
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311580#comment-14311580
]
Mark Thomas commented on IO-468:
The patch as currently would trigger a class loader memory le
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311562#comment-14311562
]
Thomas Neidhart commented on IO-468:
You did not fully understand the article: it is not a
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311556#comment-14311556
]
Bernd Hopp commented on IO-468:
---
[~tn]
please look again, there is a jit-phase at the start of m
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311513#comment-14311513
]
Thomas Neidhart commented on IO-468:
additional note: your micro-benchmarking test is also
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311511#comment-14311511
]
Thomas Neidhart commented on IO-468:
btw. this Stackoverflow question about ThreadLocals s
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311504#comment-14311504
]
Thomas Neidhart commented on IO-468:
The drawbacks of ThreadLocals are that they might lea
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311352#comment-14311352
]
Bernd Hopp commented on IO-468:
---
I'm sorry, I attached an outdated 'performancetest.ods'. I have
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311350#comment-14311350
]
Bernd Hopp commented on IO-468:
---
[~tn]
"ThreadLocals have drawbacks, and I would not use them f
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311337#comment-14311337
]
Bernd Hopp commented on IO-468:
---
"Trying to do some internal optimization for the general use ca
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311334#comment-14311334
]
Bernd Hopp commented on IO-468:
---
this is my software setup:
mvn -version
Apache Maven 3.2.2 (NO
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311332#comment-14311332
]
Thomas Neidhart commented on IO-468:
if there are already method allowing to pass an array
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311312#comment-14311312
]
Sebb commented on IO-468:
-
@[~tn]
The IOUtils class already provides access to methods that accept a
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311308#comment-14311308
]
Sebb commented on IO-468:
-
What version of Java did you use for compile/test?
As I already wrote, I d
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311300#comment-14311300
]
Thomas Neidhart commented on IO-468:
There are use-cases where users want to prevent memor
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311285#comment-14311285
]
Bernd Hopp commented on IO-468:
---
testing was more difficult than I expected, but heres some resu
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14307092#comment-14307092
]
Bernd Hopp commented on IO-468:
---
I'm gonna do some more testing on:
1: Is there a reproducable
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14307074#comment-14307074
]
Sebb commented on IO-468:
-
As to the use of ThreadLocal for the other buffers:
I agree that it does s
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14307032#comment-14307032
]
Bernd Hopp commented on IO-468:
---
you're right, this does not make sense, I'm gonna revert these
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14307005#comment-14307005
]
Sebb commented on IO-468:
-
Regarding the Skip buffer: this is currently a static buffer, and is shared
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14306930#comment-14306930
]
Bernd Hopp commented on IO-468:
---
I referred to the current PR at
https://github.com/apache/comm
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14305390#comment-14305390
]
Sebb commented on IO-468:
-
Where is the new patch?
Note that with the current code, the memory is rel
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14305183#comment-14305183
]
Bernd Hopp commented on IO-468:
---
Hello Sebb, thanks for your comments.
Your concerns regarding
[
https://issues.apache.org/jira/browse/IO-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14303254#comment-14303254
]
Sebb commented on IO-468:
-
AIUI memory allocation via new byte[] is not cpu-intensive.
Also the memory
46 matches
Mail list logo