SharedHashMap has been reengineered and moved to ChronicleMap under the new GitHub project https://github.com/OpenHFT/Chronicle-Map
All new development of SharedHashMap will be done on the ChronicleMap project, the old ShareHashMap code will remain here, all new enhancements will be added to ChronicleMap.
If you are running this in production and wish to have professionally supported code, please see http://openhft.net/support
Huge Collections for Java using efficient off heap storage
HugeHashMap supports concurrent access to very large, off heap key/value pairs. From HugeHashMapTest
HugeConfig config = HugeConfig.DEFAULT.clone()
.setSegments(128)
.setSmallEntrySize(128)
.setCapacity(count);
final HugeHashMap<CharSequence, SampleValues> map =
new HugeHashMap<CharSequence, SampleValues>(
config, CharSequence.class, SampleValues.class);
final SampleValues value = new SampleValues();
StringBuilder user = new StringBuilder();
for (int i = 0; i < count; i++) {
value.ee = i;
value.gg = i;
value.ii = i;
map.put(users(user, i), value);
}
for (int i = 0; i < count; i++) {
assertNotNull(map.get(users(user, i), value));
// test we get back the values we stored.
assertEquals(i, value.ee);
assertEquals(i, value.gg, 0.0);
assertEquals(i, value.ii);
}
for (int i = 0; i < count; i++)
map.remove(users(user, i));
HugeHashMapTest.testPutPerf() with the options "-ea -mx128m -Xmn64m -verbose:gc" prints on an Hex core i7 @ 3.5 GHz
Starting test
Put/get 14,020 K operations per second
This test adds, gets twice and removes 100 million keys without triggering a GC with a young generation of 64 MB.
The use of StringBuilder as the key allows the key to recycled efficiently. String can be used but creating millions of Strings can trigger GCs.
Also note that the value object can be recycled. There is also a get(key) method which returns a new object each time, however this can trigger GCs.
Check out our documentation at [JavaDoc] (http://openhft.github.io/HugeCollections/apidocs/)