org.mentaqueue.test.messages
Class TestAll

java.lang.Object
  org.mentaqueue.test.messages.TestAll

public class TestAll
extends Object

To run:

 java -Xms1g -Xmx4g -XX:NewSize=512m -XX:MaxNewSize=1024m -DtestNonBatchingQueues=true -cp target/classes org.mentaqueue.test.messages.TestAll 20 5 50 1000 1024
 
 format: java -Xms1g -Xmx4g -XX:NewSize=512m -XX:MaxNewSize=1024m -DtestUncheckedQueues=true -DtestNonBatchingQueues=true -cp target/classes org.mentaqueue.test.messages.TestAll IGNORE OUTLIERS RUNS MESSAGES BUFFER_SIZE
 IGNORE => How many initial runs will be ignored when calculating the average (warmup runs)?
 OUTLIERS => How many best and worst to remove?
 RUNS => How many times to run the test?
 MESSAGES => How many messages to send to the other thread? (in thousands)
 BUFFER_SIZE => How many messages can the buffer hold? (in multiples of 1024)
 

Latest results:

 ******************************************************************
   LINUX
 ******************************************************************
 
 Intel(R) Xeon(R) CPU E5345 @ 2.33GHz
 One physical processor with 6 logical processors
 Java HotSpot(TM) 64-Bit Server VM (build 20.2-b06, mixed mode)
 Linux soliveira-linux 2.6.32-30-generic x86_64 GNU/Linux
 
 Final results:
 
 1: BrokenFastQueue => 1,416,666,666 messages/sec
 2: AtomicQueue => 138,133,212 messages/sec
 3: VolatileQueue => 90,336,277 messages/sec (1.53x)
 4: LockedQueue => 25,278,134 messages/sec (5.46x)
 5: AtomicQueue (*) => 21,829,598 messages/sec (6.33x)
 6: SynchronizedQueue => 14,059,058 messages/sec (9.83x)
 7: ConcurrentLinkedQueue => 13,392,916 messages/sec (10.31x)
 8: VolatileQueue (*) => 13,367,912 messages/sec (10.33x)
 9: PooledConcurrentLinkedQueue => 6,802,307 messages/sec (20.31x)
 10: SynchronizedQueue (*) => 5,668,692 messages/sec (24.37x)
 11: PooledBlockingQueue => 5,662,282 messages/sec (24.4x)
 12: BlockingQueue => 5,289,564 messages/sec (26.11x)
 13: BlockingLinkedQueue => 4,151,632 messages/sec (33.27x)
 14: PooledBlockingLinkedQueue => 3,971,971 messages/sec (34.78x)
 15: BlockingLinkedQueue (*) => 3,909,367 messages/sec (35.33x)
 16: BlockingQueue (*) => 3,730,679 messages/sec (37.03x)
 17: BlockingArrayQueue => 3,568,064 messages/sec (38.71x)
 18: PooledBlockingLinkedQueue (*) => 3,546,631 messages/sec (38.95x)
 19: LockedQueue (*) => 3,350,480 messages/sec (41.23x)
 20: PooledBlockingArrayQueue => 3,044,092 messages/sec (45.38x)
 21: PooledBlockingQueue (*) => 3,038,979 messages/sec (45.45x)
 22: BlockingArrayQueue (*) => 2,972,741 messages/sec (46.47x)
 23: PooledBlockingArrayQueue (*) => 2,347,767 messages/sec (58.84x)
 
 (*) => Non-batching mode
 
 Test details:
 
 Warmup iterations per queue: 20
 Outliers ignored (worst and best): 10
 Total number of iterations: 60 (considered: 30)
 Total number of messages sent to the other thread: 10,000,000
 Queue buffer size: 1,048,576
 Also test non-batching queues: true
 
 

 ******************************************************************
   MAC OSX
 ******************************************************************
 
 MacBook Pro Intel Core 2 Duo 2.53 GHz
 One physical processor with 2 logical processors
 Java HotSpot(TM) 64-Bit Server VM (build 20.4-b02-402, mixed mode)
 Darwin soliveira.local 10.8.0 Darwin Kernel Version 10.8.0 root:xnu-1504.15.3~1/RELEASE_I386 i386
 
 Final results:
 
 1: BrokenFastQueue => 500,000,000 messages/sec
 2: AtomicQueue => 55,116,958 messages/sec
 3: VolatileQueue => 33,513,162 messages/sec (1.64x)
 4: AtomicQueue (*) => 28,655,461 messages/sec (1.92x)
 5: SynchronizedQueue => 26,991,465 messages/sec (2.04x)
 6: VolatileQueue (*) => 19,823,529 messages/sec (2.78x)
 7: LockedQueue => 18,451,178 messages/sec (2.99x)
 8: ConcurrentLinkedQueue => 10,900,873 messages/sec (5.06x)
 9: PooledConcurrentLinkedQueue => 6,377,823 messages/sec (8.64x)
 10: SynchronizedQueue (*) => 5,756,309 messages/sec (9.58x)
 11: LockedQueue (*) => 4,662,285 messages/sec (11.82x)
 12: BlockingLinkedQueue => 4,111,964 messages/sec (13.4x)
 13: BlockingLinkedQueue (*) => 3,978,599 messages/sec (13.85x)
 14: PooledBlockingLinkedQueue => 3,833,007 messages/sec (14.38x)
 15: BlockingQueue (*) => 3,659,262 messages/sec (15.06x)
 16: PooledBlockingLinkedQueue (*) => 3,645,013 messages/sec (15.12x)
 17: BlockingArrayQueue => 3,610,799 messages/sec (15.26x)
 18: BlockingQueue => 3,427,360 messages/sec (16.08x)
 19: PooledBlockingArrayQueue => 3,286,787 messages/sec (16.77x)
 20: PooledBlockingQueue (*) => 3,161,825 messages/sec (17.43x)
 21: PooledBlockingQueue => 3,118,518 messages/sec (17.67x)
 22: PooledBlockingArrayQueue (*) => 2,695,151 messages/sec (20.45x)
 23: BlockingArrayQueue (*) => 2,682,131 messages/sec (20.55x)
 
 (*) => Non-batching mode
 
 Test details:
 
 Warmup iterations per queue: 20
 Outliers ignored (worst and best): 10
 Total number of iterations: 50 (considered: 20)
 Total number of messages sent to the other thread: 1,000,000
 Queue buffer size: 1,048,576
 Also test non-batching queues: true
 
 

Author:

Sergio Oliveira Jr.

Constructor Summary

TestAll()

Method Summary

static void

main(String[] args)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TestAll

public TestAll()

Method Detail

main

public static void main(String[] args)
                 throws Exception

Throws:

Exception