You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "SammiChen (JIRA)" <ji...@apache.org> on 2018/05/29 09:18:00 UTC
[jira] [Commented] (HADOOP-15499) Performance severe drop when
running RawErasureCoderBenchmark with NativeRSRawErasureCoder
[ https://issues.apache.org/jira/browse/HADOOP-15499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16493287#comment-16493287 ]
SammiChen commented on HADOOP-15499:
------------------------------------
Performance data before the patch,
bin/hadoop jar ./share/hadoop/common/hadoop-common-3.0.0-alpha2.jar org.apache.hadoop.io.erasurecode.rawcoder.RawErasureCoderBenchmark encode 3 50 1024 64
Using 126MB buffer.
ISA-L coder encode 50400MB data, with chunk size 64KB
Total time: 0.98 s.
Total throughput: 51639.34 MB/s
Threads statistics:
50 threads in total.
Performance data after the patch,
bin/hadoop jar ./share/hadoop/common/hadoop-common-3.0.0-alpha2.jar org.apache.hadoop.io.erasurecode.rawcoder.RawErasureCoderBenchmark encode 3 72 10240 4096
Using 120MB buffer.
ISA-L coder encode 734400MB data, with chunk size 4096KB
Total time: 8.11 s.
Total throughput: 90521.39 MB/s
Threads statistics:
72 threads in total.
Min: 6.78 s, Max: 7.93 s, Avg: 7.36 s, 90th Percentile: 7.66 s.
I also compared the performance data of two scenarios, one is remove all the synchronized key words, another is the current ReentrantReadWriteLock solution.
The performance of ReentrantReadWriteLock solution is like less than 5% degrade than the remove synchronized key words case. It's acceptable for me.
> Performance severe drop when running RawErasureCoderBenchmark with NativeRSRawErasureCoder
> ------------------------------------------------------------------------------------------
>
> Key: HADOOP-15499
> URL: https://issues.apache.org/jira/browse/HADOOP-15499
> Project: Hadoop Common
> Issue Type: Improvement
> Affects Versions: 3.0.0, 3.0.1, 3.0.2
> Reporter: SammiChen
> Assignee: SammiChen
> Priority: Major
> Attachments: HADOOP-15499.001.patch
>
>
> Run RawErasureCoderBenchmark which is a micro-benchmark to test EC codec encoding/decoding performance.
> 50 concurrency Native ISA-L coder has the less throughput than 1 concurrency Native ISA-L case. It's abnormal.
>
> bin/hadoop jar ./share/hadoop/common/hadoop-common-3.2.0-SNAPSHOT-tests.jar org.apache.hadoop.io.erasurecode.rawcoder.RawErasureCoderBenchmark encode 3 1 1024 1024
> Using 126MB buffer.
> ISA-L coder encode 1008MB data, with chunk size 1024KB
> Total time: 0.19 s.
> Total throughput: 5390.37 MB/s
> Threads statistics:
> 1 threads in total.
> Min: 0.18 s, Max: 0.18 s, Avg: 0.18 s, 90th Percentile: 0.18 s.
>
> bin/hadoop jar ./share/hadoop/common/hadoop-common-3.2.0-SNAPSHOT-tests.jar org.apache.hadoop.io.erasurecode.rawcoder.RawErasureCoderBenchmark encode 3 50 1024 10240
> Using 120MB buffer.
> ISA-L coder encode 54000MB data, with chunk size 10240KB
> Total time: 11.58 s.
> Total throughput: 4662 MB/s
> Threads statistics:
> 50 threads in total.
> Min: 0.55 s, Max: 11.5 s, Avg: 6.32 s, 90th Percentile: 10.45 s.
>
> RawErasureCoderBenchmark shares a single coder between all concurrent threads. While
> NativeRSRawEncoder and NativeRSRawDecoder has synchronized key work on doDecode and doEncode function. So 50 concurrent threads are forced to use the shared coder encode/decode function one by one.
>
> To resolve the issue, there are two approaches.
> # Refactor RawErasureCoderBenchmark to use dedicated coder for each concurrent thread.
> # Refactor NativeRSRawEncoder and NativeRSRawDecoder to get better concurrency. Since the synchronized key work is to try to protect the private variable nativeCoder from being checked in doEncode/doDecode and being modified in release. We can use reentrantReadWriteLock to increase the concurrency since doEncode/doDecode can be called multiple times without change the nativeCoder state.
> I prefer approach 2 and will upload a patch later.
>
>
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org