-
Notifications
You must be signed in to change notification settings - Fork 40
Performance Test on Windows 7
#Performance Test on Windows 7
In the following tests, one producer process, one consumer process and one broker process all run on my server(Win7, HD with NTFS), and there is only one active topic on broker for testing in each round. These tests are performed to check the pure performance of Luxun messaging system by separating the impact of network. See full hardware spec for benchmark here.
JVM setting:
- producer : -Xmx1g
- consuemr : -Xmx1g
- broker : -Xmx4g
###Test Suite 1 - Standalone Async Producing(batch size = 200, flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC11 | 100 | 1 | 2000000 | 110.8924 | 1162790.6977 |
| TC12 | 1024 | 1 | 2000000 | 336.1661 | 344234.0792 |
| TC13 | 2048 | 1 | 2000000 | 394.5707 | 202020.2020 |
| TC14 | 4096 | 1 | 2000000 | 403.9555 | 103412.6163 |
| TC15 | 100 | 10 | 2000000 | 244.5319 | 2564102.5641 |
| TC16 | 1024 | 10 | 2000000 | 453.1613 | 464037.1230 |
| TC17 | 2048 | 10 | 2000000 | 438.4119 | 224466.8911 |
| TC18 | 4096 | 10 | 2000000 | 423.4417 | 108401.0840 |
###Test Suite 2 - Standalone Sync Producing(flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC21 | 100 | 1 | 2000000 | 10.6914 | 112107.6233 |
| TC22 | 1024 | 1 | 2000000 | 91.1398 | 93327.1115 |
| TC23 | 2048 | 1 | 2000000 | 144.2485 | 73855.2437 |
| TC24 | 4096 | 1 | 2000000 | 177.6780 | 45485.5583 |
| TC25 | 100 | 10 | 2000000 | 21.3112 | 223463.6872 |
| TC26 | 1024 | 10 | 2000000 | 124.8801 | 127877.2379 |
| TC27 | 2048 | 10 | 2000000 | 162.5572 | 83229.2967 |
| TC28 | 4096 | 10 | 2000000 | 139.0125 | 35587.1886 |
###Test Suite 3 - Standalone Consuming(fetch size = 1024 * 1024bytes)
| Test Case | messge size(bytes) | thread number | nMsgs fetched | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC31 | 100 | 1 | 2000000 | 222.3017 | 2331002.3310 |
| TC32 | 1024 | 1 | 2000000 | 331.9947 | 339962.6041 |
| TC33 | 2048 | 1 | 2000000 | 340.7406 | 174459.1766 |
| TC34 | 4096 | 1 | 2000000 | 274.8558 | 70363.0735 |
| TC35 | 100 | 10 | 2000000 | 210.9899 | 2212389.3805 |
| TC36 | 1024 | 10 | 2000000 | 319.7030 | 327375.8865 |
| TC37 | 2048 | 10 | 2000000 | 342.8517 | 175540.0952 |
| TC38 | 4096 | 10 | 2000000 | 327.7624 | 83907.1854 |
###Test Suite 4 - Concurrent Producing(Async) & Consuming(batch size = 200, fetch size = 1024 * 1024bytes, flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent/fetched | producer throughput(MB/s) | producer speed(nMsg/s) | consumer throughput(MB/s) | consumer speed(nMsg/s) |
|---|---|---|---|---|---|---|---|
| TC41 | 1024 | 20(p:10, c:10) | 50000000 | 117.0237 | 119832.2349 | 121.5719 | 124489.5927 |
In the following tests, one producer process, one consumer process and one broker process all run on my server(Win7, HD with NTFS), and there is only one active topic on broker for testing in each round. These tests are performed to check the pure performance of Luxun messaging system with default JVM setting. See full hardware spec for benchmark here.
JVM setting: default
###Test Suite 1 - Standalone Async Producing(batch size = 200, flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC11 | 100 | 1 | 2000000 | 147.8565 | 1550387.5969 |
| TC12 | 1024 | 1 | 2000000 | 390.6250 | 400000 |
| TC13 | 2048 | 1 | 2000000 | 478.7071 | 245098.0392 |
| TC14 | 4096 | 1 | 2000000 | 232.3075 | 59470.7107 |
| TC15 | 100 | 10 | 2000000 | 164.4266 | 1724137.9310 |
| TC16 | 1024 | 10 | 2000000 | 231.6874 | 237247.9241 |
| TC17 | 2048 | 10 | 2000000 | 338.2035 | 173160.1732 |
| TC18 | 4096 | 10 | 2000000 | 139.1362 | 35618.8780 |
###Test Suite 2 - Standalone Sync Producing(flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC21 | 100 | 1 | 2000000 | 10.4341 | 109409.1904 |
| TC22 | 1024 | 1 | 2000000 | 91.7822 | 93984.9624 |
| TC23 | 2048 | 1 | 2000000 | 128.8766 | 65984.8235 |
| TC24 | 4096 | 1 | 2000000 | 138.2010 | 35379.4445 |
| TC25 | 100 | 10 | 2000000 | 20.8453 | 218579.2350 |
| TC26 | 1024 | 10 | 2000000 | 115.3647 | 118133.4908 |
| TC27 | 2048 | 10 | 2000000 | 123.8899 | 63431.6524 |
| TC28 | 4096 | 10 | 2000000 | 136.1276 | 34848.6697 |
###Test Suite 3 - Standalone Consuming(fetch size = 1024 * 1024bytes)
| Test Case | messge size(bytes) | thread number | nMsgs fetched | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC31 | 100 | 1 | 2000000 | 190.7349 | 2000000 |
| TC32 | 1024 | 1 | 2000000 | 306.1804 | 313528.7663 |
| TC33 | 2048 | 1 | 2000000 | 217.1949 | 111203.7809 |
| TC34 | 4096 | 1 | 2000000 | 324.1432 | 82980.6655 |
| TC35 | 100 | 10 | 2000000 | 231.1938 | 2424242.4242 |
| TC36 | 1024 | 10 | 2000000 | 332.6167 | 340599.4550 |
| TC37 | 2048 | 10 | 2000000 | 360.9212 | 184791.6474 |
| TC38 | 4096 | 10 | 2000000 | 340.4584 | 87157.3626 |
In the following tests, one broker runs on my server, one producer process and one consumer process run on my notebook(Win7), and there is only one active topic on borker for testing in each round. These tests are performed to check the performance of Luxun messaging system in networking environment. The switch used is a very cheap 1g switch, see full hardware spec for benchmark here.
JVM setting:
- producer : -Xmx1g
- consuemr : -Xmx1g
- broker : -Xmx4g
###Test Suite 1 - Standalone Async Producing(batch size = 200, flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC11 | 100 | 1 | 2000000 | 18.4964 | 193948.7975 |
| TC12 | 1024 | 1 | 2000000 | 55.0564 | 56377.7308 |
| TC13 | 2048 | 1 | 2000000 | 59.4351 | 30430.7472 |
| TC14 | 4096 | 1 | 2000000 | 75.6607 | 19369.1469 |
| TC15 | 100 | 10 | 2000000 | 76.8782 | 806126.5619 |
| TC16 | 1024 | 10 | 2000000 | 107.5628 | 110144.2890 |
| TC17 | 2048 | 10 | 2000000 | 111.0424 | 56853.7154 |
| TC18 | 4096 | 10 | 2000000 | 106.1020 | 27162.1034 |
###Test Suite 2 - Standalone Sync Producing(flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC21 | 100 | 1 | 2000000 | 6.3057 | 66120.0741 |
| TC22 | 1024 | 1 | 2000000 | 16.2444 | 16634.2299 |
| TC23 | 2048 | 1 | 2000000 | 15.9745 | 8178.9555 |
| TC24 | 4096 | 1 | 2000000 | 15.5229 | 3973.8678 |
| TC25 | 100 | 10 | 2000000 | 7.7975 | 81762.8061 |
| TC26 | 1024 | 10 | 2000000 | 57.6960 | 59080.7042 |
| TC27 | 2048 | 10 | 2000000 | 73.8212 | 37796.4660 |
| TC28 | 4096 | 10 | 2000000 | 83.7739 | 21446.1113 |
###Test Suite 3 - Standalone Consuming(fetch size = 1024 * 1024bytes)
| Test Case | messge size(bytes) | thread number | nMsgs fetched | throughput(MB/s) | speed(nMsg/s) |
|---|---|---|---|---|---|
| TC31 | 100 | 1 | 2000000 | 58.7238 | 615763.5468 |
| TC32 | 1024 | 1 | 2000000 | 75.1433 | 76946.7528 |
| TC33 | 2048 | 1 | 2000000 | 75.9380 | 38880.2488 |
| TC34 | 4096 | 1 | 2000000 | 75.7568 | 19393.7513 |
| TC35 | 100 | 10 | 2000000 | 62.8837 | 658544.6164 |
| TC36 | 1024 | 10 | 2000000 | 74.5438 | 76332.9644 |
| TC37 | 2048 | 10 | 2000000 | 76.1497 | 38988.6348 |
| TC38 | 4096 | 10 | 2000000 | 74.8639 | 19165.1654 |
###Test Suite 4 - Concurrent Producing(Async) & Consuming(batch size = 200, fetch size = 1024 * 1024bytes, flush disabled on broker)
| Test Case | messge size(bytes) | thread number | nMsgs sent/fetched | producer throughput(MB/s) | producer speed(nMsg/s) | consumer throughput(MB/s) | consumer speed(nMsg/s) |
|---|---|---|---|---|---|---|---|
| TC41 | 1024 | 20(p:10, c:10) | 5000000 | 107.9313 | 110521.6622 | 59.7257 | 61159.0870 |