[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16653717#comment-16653717
]
ASF GitHub Bot commented on IGNITE-9272:
Github user asfgit closed the pull request at:
https://github.com/apache/ignite/pull/4619
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results (less - better):
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16653643#comment-16653643
]
Ignite TC Bot commented on IGNITE-9272:
---
{panel:title=Possible
Blockers|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1}
{color:#d04437}Hadoop{color} [[tests 0 TIMEOUT , Out Of Memory Error , Exit
Code , [Artifacts publishing failed]
|https://ci.ignite.apache.org/viewLog.html?buildId=2105003]]
* HadoopMapReduceErrorResilienceTest.testRecoveryAfterAnError0_Error (last
started)
{panel}
[TeamCity Run
All|http://ci.ignite.apache.org/viewLog.html?buildId=2105086buildTypeId=IgniteTests24Java8_RunAll]
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results (less - better):
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16653642#comment-16653642
]
Ignite TC Bot commented on IGNITE-9272:
---
{panel:title=Possible
Blockers|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1}
{color:#d04437}Hadoop{color} [[tests 0 TIMEOUT , Out Of Memory Error , Exit
Code , [Artifacts publishing failed]
|https://ci.ignite.apache.org/viewLog.html?buildId=2105003]]
{panel}
[TeamCity Run
All|http://ci.ignite.apache.org/viewLog.html?buildId=2105086buildTypeId=IgniteTests24Java8_RunAll]
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results (less - better):
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16651704#comment-16651704
]
Nikolay Izhikov commented on IGNITE-9272:
-
Latest Run All - https://ci.ignite.apache.org/viewQueued.html?itemId=2098115
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results (less - better):
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16641638#comment-16641638
]
Nikolay Izhikov commented on IGNITE-9272:
-
[~zstan] I looked to possible blockers.
Looks like it a bot error.
The only test we should check is .Net.
Please, rerun this suite and ensure there is no new failures.
Please, attach a link to a build here.
Your changes look good to me.
I will merge it after we will have test results.
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results:
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16641617#comment-16641617
]
Ignite TC Bot commented on IGNITE-9272:
---
{panel:title=Possible
Blockers|borderStyle=dashed|borderColor=#ccc|titleBGColor=#F7D6C1}
{color:#d04437}MVCC Cache{color} [[tests
7|https://ci.ignite.apache.org/viewLog.html?buildId=2028675]]
* IgniteCacheMvccTestSuite: CacheMvccTransactionsTest.testRebalanceSimple -
0,0% fails in last 100 master runs.
* IgniteCacheMvccTestSuite: DataStreamProcessorMvccSelfTest.testColocated -
0,0% fails in last 100 master runs.
* IgniteCacheMvccTestSuite: DataStreamProcessorMvccSelfTest.testPartitioned -
0,0% fails in last 100 master runs.
* IgniteCacheMvccTestSuite: DataStreamProcessorMvccSelfTest.testReplicated -
0,0% fails in last 100 master runs.
{color:#d04437}Platform .NET (Core Linux){color} [[tests
8|https://ci.ignite.apache.org/viewLog.html?buildId=2028626]]
* dll: IgniteConfigurationTest.TestDefaultSpi - 0,0% fails in last 100 master
runs.
{color:#d04437}Cache 3{color} [[tests
2|https://ci.ignite.apache.org/viewLog.html?buildId=2028657]]
* IgniteBinaryObjectsCacheTestSuite3:
IgniteCacheGroupsTest.testRestartsAndCacheCreateDestroy - 0,0% fails in last
100 master runs.
{panel}
[TeamCity Run
All|http://ci.ignite.apache.org/viewLog.html?buildId=2028678buildTypeId=IgniteTests24Java8_RunAll]
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results:
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16639963#comment-16639963
]
Nikolay Izhikov commented on IGNITE-9272:
-
[~zstan]
Generally looks good!
I left some comments in PR, mostly minor.
Please, address them.
I think we can consider to include this ticket to 2.7 scope.
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.8
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> {noformat}
> jmh results:
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> {noformat}
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16639322#comment-16639322 ] Stanilovsky Evgeny commented on IGNITE-9272: plz review ticket. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.8 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16632893#comment-16632893 ] Stanilovsky Evgeny commented on IGNITE-9272: [~NIzhikov], thanks for review, hope all your comments resolved besides crcWithReset func namig. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.8 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16632809#comment-16632809 ] Nikolay Izhikov commented on IGNITE-9272: - [~zstan] Please, reolve my comments in PR: https://github.com/apache/ignite/pull/4619 > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.8 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631967#comment-16631967 ] Vladimir Ozerov commented on IGNITE-9272: - Agreed to move the ticket to the next release. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.8 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631798#comment-16631798 ] Vladimir Ozerov commented on IGNITE-9272: - On the other hand, I see that you did multiple runs, and performance improvement is consistent between runs. I would suggest to do the following: 1) If TC is OK, ticket is reviewed and storage compatibility is not broken by 30 Sept - then merge to AI 2.7 2) Otherwise - move to AI 2.8. What do you think? > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631795#comment-16631795 ] Vladimir Ozerov commented on IGNITE-9272: - 4-5% is very close to typical measurement error. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[
https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631745#comment-16631745
]
Stanilovsky Evgeny commented on IGNITE-9272:
yep, 25% i mentioned above, it`s our inner benchmarks, yardstick shows:
atomic-get-and-put 6000s duration.
Throughput (more - better).
{code:java}
ignite-9272:
65530.42
65523.72
65536.26
65487.34
65811.76
65202.05
65372.89
65627.32
master:
62740.62
62772.30
62928.18
62989.85
62926.43
62668.95
63026.49
62885.89
{code}
= 4-5% profit, thoughts?
> PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
> --
>
> Key: IGNITE-9272
> URL: https://issues.apache.org/jira/browse/IGNITE-9272
> Project: Ignite
> Issue Type: Improvement
> Components: general
>Affects Versions: 2.6
>Reporter: Stanilovsky Evgeny
>Assignee: Stanilovsky Evgeny
>Priority: Major
> Fix For: 2.7
>
> Attachments: BenchmarkCRC.java
>
>
> I see that Ignite has its own crc32 realization called: PureJavaCrc32 and
> from desc it seems to be : _The current version is ~10x to 1.8x as fast as
> Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite
> results.
> + If it really so, looks like backward compatibility would be easy, all that
> need is just to take lower part of long form zip.crc32 realization.
> jmh results:
> Benchmark Mode CntScoreError Units
> BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op
> BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op
> JMH version: 1.21
> VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11
> VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java
> op system : ubuntu 16.10
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631392#comment-16631392 ] Vladimir Ozerov commented on IGNITE-9272: - [~zstan], [~NIzhikov], folks, ticket doesn't seem to be ready at the moment. I propose to move it to AI 2.8 unless we have clear indication that the fix gives significant performance boost and doesn't break compatibiltiy. [~zstan], could you please share your benchmark results? Was it Yardstick or something else? > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16631338#comment-16631338 ] Nikolay Izhikov commented on IGNITE-9272: - Hello [~zstan] I have done the review. Please, take a look at my comments in PR > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16626828#comment-16626828 ] Stanilovsky Evgeny commented on IGNITE-9272: [~NIzhikov] i found speed improvement reach 25% on some grid operations. Who can review ? > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16625072#comment-16625072 ] Stanilovsky Evgeny commented on IGNITE-9272: [~NIzhikov] tomorrow a plan to finish with benchmarks on real data, will publish them here. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16623595#comment-16623595 ] Nikolay Izhikov commented on IGNITE-9272: - [~zstan] Do we have a chance to resolve this ticket until the code freeze of 2.7? > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 2.7 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16592582#comment-16592582 ] ASF GitHub Bot commented on IGNITE-9272: GitHub user zstan opened a pull request: https://github.com/apache/ignite/pull/4619 IGNITE-9272 j.u.zip.CRC32algo instead of PureJavaCrc32 You can merge this pull request into a Git repository by running: $ git pull https://github.com/gridgain/apache-ignite ignite-9272 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/ignite/pull/4619.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #4619 commit 7479203c7bb4ffa6fecfa6a3d604f1c61a8ddfdb Author: Evgeny Stanilovskiy Date: 2018-08-24T14:34:05Z IGNITE-9272 j.u.zip.CRC32algo instead of PureJavaCrc32 > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 3.0 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (IGNITE-9272) PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace.
[ https://issues.apache.org/jira/browse/IGNITE-9272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16586490#comment-16586490 ] Dmitriy Setrakyan commented on IGNITE-9272: --- Can we add this option without breaking compatibility with previous page/storage formats? If not, then this should support both implementation. The default should be the new fastest implementation, but if we encounter the older, slower one, then we should print out a warning in the log and automatically switch to the older implementation. > PureJavaCrc32 vs j.u.zip.CRC32 benchmark and probably replace. > -- > > Key: IGNITE-9272 > URL: https://issues.apache.org/jira/browse/IGNITE-9272 > Project: Ignite > Issue Type: Improvement > Components: general >Affects Versions: 2.6 >Reporter: Stanilovsky Evgeny >Assignee: Stanilovsky Evgeny >Priority: Major > Fix For: 3.0 > > Attachments: BenchmarkCRC.java > > > I see that Ignite has its own crc32 realization called: PureJavaCrc32 and > from desc it seems to be : _The current version is ~10x to 1.8x as fast as > Sun's native java.util.zip.CRC32 in Java 1.6_ But my jmh tests show opposite > results. > + If it really so, looks like backward compatibility would be easy, all that > need is just to take lower part of long form zip.crc32 realization. > jmh results: > Benchmark Mode CntScoreError Units > BenchmarkCRC.Crc32 avgt5 1521060.716 ± 44083.424 ns/op > BenchmarkCRC.pureJavaCrc32 avgt5 4657756.671 ± 177243.254 ns/op > JMH version: 1.21 > VM version: JDK 1.8.0_131, Java HotSpot(TM) 64-Bit Server VM, 25.131-b11 > VM invoker: /usr/lib/jvm/java-8-oracle/jre/bin/java > op system : ubuntu 16.10 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
