[
https://issues.jboss.org/browse/JGRP-1998?page=com.atlassian.jira.plugin....
]
Bela Ban edited comment on JGRP-1998 at 12/22/15 3:58 AM:
----------------------------------------------------------
h3. Status
* {{TP.receive()}} now de-serializes both single messages and message batches
* TCP uses a pre-allocated buffer for each connection ({{TcpConnection}}), which grows on
demand, rather than creating a new one for each message. This reduced memory allocation
(UPerf, 4 nodes, 2M msgs/node) by ca. 1.8GB
* TCP_NIO2 now also pre-allocates a buffer (in {{NioConnection}}) and grows it on demand
was (Author: belaban):
h3. Status
* {{TP.receive()}} now de-serializes both single messages and message batches
* TCP uses a pre-allocated buffer for each connection ({{TcpConnection}}), which grows on
demand, rather than creating a new one for each message. This reduced memory allocation
(UPerf, 4 nodes, 2M msgs/node) by ca. 1.8GB
Transport: reuse of incoming buffers
------------------------------------
Key: JGRP-1998
URL:
https://issues.jboss.org/browse/JGRP-1998
Project: JGroups
Issue Type: Enhancement
Reporter: Bela Ban
Assignee: Bela Ban
Fix For: 3.6.7
Investigate whether buffers can be reused on the receive side. JGRP-1989 already provides
this on the send side. Goal: reduction of memory allocation rate.
--
This message was sent by Atlassian JIRA
(v6.4.11#64026)