Alawn/Chronicle-Map
1
0
Fork 0
mirror of https://github.com/OpenHFT/Chronicle-Map.git synced 2021-09-13 01:26:12 +00:00
Code Releases Activity

Update README.md

Browse Source
This commit is contained in:
Rob Austin
2014-09-23 15:30:44 +01:00
parent f39eabd95b
commit 511f195a9f
1 changed files with 13 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+13 -13
View File
@@ -57,23 +57,10 @@ A Chronicle Map implements the java.util.concurrent.ConcurrentMap, so on the fac
## When to use
#### When to use HashMap
Most of the time.
If you compare HashMap, ConcurrentHashMap and Chronicle Map, most of the maps in your system are likely to be HashMap. This is because HashMap is lightweight and synchronized HashMap works well for lightly contended use cases. By contention I mean, how many threads on average are trying to use a Map. One reason you can't have many contended resources, is that you only have so many CPUs and they can only be accessing so many resources at once (ideally no more than one or two per thread at a time)
#### Known Issues
Chronicle map lets you assign a map larger than your available memory, This will however impact performance as Chronicle Map will have to page the segments to and from disk as data is written or read from the map. This feature works brilliantly on Linux but unfortunately its not supported on Windows, if you use more memory than is physically available you will experience the following error :
```java
Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
j sun.misc.Unsafe.compareAndSwapLong(Ljava/lang/Object;JJJ)Z+0
j net.openhft.lang.io.NativeBytes.compareAndSwapLong(JJJ)Z+13
j net.openhft.lang.io.AbstractBytes.tryLockNanos8a(JJ)Z+12
j net.openhft.lang.io.AbstractBytes.tryLockNanosLong(JJ)Z+41
j net.openhft.collections.AbstractVanillaSharedHashMap$Segment.lock()V+12
```
#### When to use ConcurrentHashMap
Some of the time.
@@ -449,6 +436,19 @@ cluster = clusterReplicatorBuilder.create();
map = clusterReplicatorBuilder.create((short) 1, builder);
```
#### Known Issues
Chronicle map lets you assign a map larger than your available memory, This will however impact performance as Chronicle Map will have to page the segments to and from disk as data is written or read from the map. This feature works brilliantly on Linux but unfortunately its not supported on Windows, if you use more memory than is physically available you will experience the following error :
```java
Java frames: (J=compiled Java code, j=interpreted, Vv=VM code)
j sun.misc.Unsafe.compareAndSwapLong(Ljava/lang/Object;JJJ)Z+0
j net.openhft.lang.io.NativeBytes.compareAndSwapLong(JJJ)Z+13
j net.openhft.lang.io.AbstractBytes.tryLockNanos8a(JJ)Z+12
j net.openhft.lang.io.AbstractBytes.tryLockNanosLong(JJ)Z+41
j net.openhft.collections.AbstractVanillaSharedHashMap$Segment.lock()V+12
```
# Example : Replicating data between process on different servers
Lets assume that we had two server, lets call them server1 and server2, if we wished to share a map between them, this is how we could set it up