6:12 a.m.
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:14 a.m.
New subject: [JBoss JIRA] Updated: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban updated JGRP-827:
--------------------------
Attachment: JGroups.fast.zip
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:18 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
Correction: with 2.6 we get ca 71'000
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:50 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
The version from 2008-09-15 yiels 77'351 msgs/sec !
cvs -d:ext:belaban@javagroups.cvs.sf.net:/cvsroot/javagroups co -D2008-09-15 -d
JGroups-2008-09-15 JGroups
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:53 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
Version 2008-09-20 (1 week later): 76'432 ! So the regression must have happened
between the 20th and 25th. I'm further drilling down into this...
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:05 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
2008-09-24: 77'342: OK
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:05 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
2008-09-22 is fine too.
Also 2008-09-23: 75'976 msgs/sec
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:11 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
2008-09-27 is slow: 65'940 msgs/sec ! So the regression must have happened between 27
and 24, going back..
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:13 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
2008-09-25 is OK: 77'615 msgs/sec
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
7:13 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
2008-09-26 is slow: 66'031 msgs/sec ! So the regression must have happend between 25
and 26, comparing the 2 snapshots now...
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
9:57 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
The 2 files which cause the regression are UDP and TP. With the following versions,
performance is great (78'108 msgs/sec !):
- UDP.java,v 1.183 2008/09/05 14:59:03 vlada Exp (latest: 1.185)
- TP.java,v 1.228 2008/09/25 13:15:06 belaban Exp (latest: 1.229)
Later versions show the regression.
Looking into what happend between UDP 1.183 and 1.185 and TP 1.228 and 1.229
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
10:21 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
The culprit is ... UDP !
The change from sending multicasts out over mcast_sock to sock caused the regression. Now
I need to find out why...
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
11:05 a.m.
New subject: [JBoss JIRA] Commented: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban commented on JGRP-827:
-------------------------------
OK, reverted back to using mcast_sock for sending multicast packets. Got 81'042
msgs/sec ! The config used is JGroups/conf/fast-local.xml and is attached
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: JGroups.fast.zip
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
11:07 a.m.
New subject: [JBoss JIRA] Updated: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban updated JGRP-827:
--------------------------
Attachment: fast-local.xml
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: fast-local.xml
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
11:07 a.m.
New subject: [JBoss JIRA] Updated: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban updated JGRP-827:
--------------------------
Attachment: (was: JGroups.fast.zip)
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: fast-local.xml
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
11:34 a.m.
New subject: [JBoss JIRA] Resolved: (JGRP-827) Performance regression
[
https://jira.jboss.org/jira/browse/JGRP-827?page=com.atlassian.jira.plugi...
]
Bela Ban resolved JGRP-827.
---------------------------
Resolution: Done
Performance regression
----------------------
Key: JGRP-827
URL: https://jira.jboss.org/jira/browse/JGRP-827
Project: JGroups
Issue Type: Bug
Reporter: Bela Ban
Assignee: Bela Ban
Priority: Critical
Fix For: 2.7
Attachments: fast-local.xml
There is a perf regression in 2.7: start 2 senders (perf.Test), send 1M 1K messages. With
2.6, we get ca 68'000 msgs/sec, with JGroups.fast (snapshot ca 2 weeks ago), we get
74'000 and with CVS head we get 65'000 msgs/sec.
Attached is JGroups.fast.zip
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
6521
days inactive
6521
days old
15 comments
1 participants
participants (1)
-
Bela Ban (JIRA)