Author: sshafroi
Date: 2008-12-02 12:02:02 +0100 (Tue, 02 Dec 2008)
New Revision: 7009
Modified:
branches/2.18/data-model-api/src/main/java/no/sesat/search/datamodel/generic/MapDataObjectSupport.java
Log:
It looks like we would get better performance by increasing number of expected
simuntaniously threads a bit. I noticed a lot of synchronization on this while
stress-testing.
Modified:
branches/2.18/data-model-api/src/main/java/no/sesat/search/datamodel/generic/MapDataObjectSupport.java
===================================================================
---
branches/2.18/data-model-api/src/main/java/no/sesat/search/datamodel/generic/MapDataObjectSupport.java
2008-12-02 10:16:44 UTC (rev 7008)
+++
branches/2.18/data-model-api/src/main/java/no/sesat/search/datamodel/generic/MapDataObjectSupport.java
2008-12-02 11:02:02 UTC (rev 7009)
@@ -1,4 +1,4 @@
-/* Copyright (2007) Schibsted Søk AS
+/* Copyright (2007-2008) Schibsted Søk AS
* This file is part of SESAT.
*
* SESAT is free software: you can redistribute it and/or modify
@@ -49,7 +49,7 @@
// Most MapDataObjectSupport instances only contain an item or two.
// Max currency in any mode is typically ~20, but unlikely for even two
threads to update at the same time.
- private final Map<String,V> map = new ConcurrentHashMap<String,V>(5,
0.75f, 2){
+ private final Map<String,V> map = new ConcurrentHashMap<String,V>(5,
0.75f, 10){
@Override
public V put(final String key, final V value){
_______________________________________________
Kernel-commits mailing list
[email protected]
http://sesat.no/mailman/listinfo/kernel-commits
