To benchmark IO on Linux and MySQL transaction processing, SysBench is a popular choice that can do both. After poking around at the source code, it seems PostgreSQL and Oracle are also included for transaction processing testing if you have the proper header files, but I didn’t test those.

1. Cached
2. Sysbench Download (DEB, RPM, TGZ, TXZ, XZ, ZST)
3. MySQL :: MySQL Benchmark Tool
4. Sysbench Windows 安装

To benchmark IO on Windows and SQL Server transaction processing, Microsoft provides two tools, SQLIO and SQLIOSim. SQLIO is a misnomer in that it really doesn’t have much to do with SQL Server. It is a general purpose disk IO benchmark tool.

1. Jun 28, 2020 Sysbench is a multi-threaded benchmark tool based on luaJIT it’s the actual standard for MySQL benchmarks, it needs to be able to connect to the database. Sysbench Installation. First, we need to install the sysbench, I am installing sysbench on another server so that we can test the actual impact of load on our MySQL server.
2. -warmup-time=N execute events for this many seconds with statistics disabled before the actual benchmark run with statistics enabled 0 This is self-explanatory. SysBench generates statistical results from the tests and those results may be affected if MySQL is in a cold state.
3. How To Build On Windows: 1.Open the command prompt ( cmd.exe ) and go to the downloaded sysbench directory. 2.To Build with MYSQL Support,you will need mysql.h header file and client library libmysqld.lib. 3.In the sysbench directory execute the cmake command. 4.Once the cmake command is.

Sep 19, 2010 SQLIO is using one file and sysbench is using 128 files for the test Good point, Vladislav. I did another test with SysBench with one single file 3G in size. Sysbench –test=fileio –file-total-size=3G –file-num=1 prepare sysbench 0.4: multi-threaded system evaluation benchmark. 1 files, 3145728Kb each, 3072Mb total Creating files for the. Arch Linux Community aarch64 Official sysbench-1.0.20-1-aarch64.pkg.tar.xz: Scriptable multi-threaded benchmark tool for databases and systems: Arch Linux Community x8664 Official.

So today I was playing with SysBench and noticed that I can compile and build it on Windows as well. I decided I should run IO benchmark on a single machine with both tools (SQLIO and SysBench), and see if I could reconcile the results.

To make things simple, I thought I would just benchmark random read of 3G (orders of magnitude bigger than disk controller cache) files for 5 minutes (300 seconds) with a single thread using 16Kb block size, without adding any additional queue. I tested this on both my laptop and an Amazon EC2 instance. The commands for both tools are listed below, and they should perform the same thing, as far as I can tell. Let me know if you have any comments/pointers or if I missed anything.

SysBench commands:

Fro SQLIO, here is the line in param.txt and command used:

As this is a quick test, I ran the same test twice and took the average value for comparison purposes. The detailed output is pasted at the end of this post.

On my Windows XP Pro Service Pack 2 laptop with Intel X-25 SSD:

So on my laptop, SQLIO’s results are 13% higher than that of SysBench.

On Amazon EC2 ami-c3e40daa with EBS device running Windows Server 2008 Datacenter Edition Service Pack 2, whose results varied widely between my two runs:

On this machine, SQLIO results are 66% higher than that of SysBench.

Below is the gory details.

Here are the detailed output on my laptop:
SQLIO
C:Program FilesSQLIO>sqlio -kR -s300 -dc -b16 -frandom -Fparam.txt
sqlio v1.5.SG
parameter file used: param.txt
file c:testfile.dat with 1 thread (0) using mask 0x0 (0)
1 thread reading for 300 secs from file c:testfile.dat
using 16KB random IOs
using specified size: 3072 MB for file: c:testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec: 3835.39
MBs/sec: 59.92

C:Program FilesSQLIO>sqlio -kR -s300 -dc -b16 -frandom -Fparam.txt
sqlio v1.5.SG
parameter file used: param.txt
file c:testfile.dat with 1 thread (0) using mask 0x0 (0)
1 thread reading for 300 secs from file c:testfile.dat
using 16KB random IOs
using specified size: 3072 MB for file: c:testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec: 3832.00
MBs/sec: 59.87

SysBench
C:MessAroundsysbench-0.4.12sysbench-0.4.12sysbenchRelWithDebInfo>sysbench.e
xe –test=fileio –file-total-size=3G –file-test-mode=rndrd –max-time=300 run
sysbench 0.4: multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
128 files, 24Mb each
3Gb total file size
Block size 16Kb
Number of random requests for random IO: 10000
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random read test
Threads started!
WARNING: Operation time (18446744073709226000.000000) is greater than maximal co
unted value, counting as 10000000000000.000000
WARNING: Percentile statistics will be inaccurate
Done.

Operations performed: 10000 Read, 0 Write, 0 Other = 10000 Total
Read 156.25Mb Written 0b Total transferred 156.25Mb (52.143Mb/sec)
3337.16 Requests/sec executed

Test execution summary:
total time: 2.9966s
total number of events: 10000
total time taken by event execution: 2.9343
per-request statistics:
min: 0.01ms
avg: 0.29ms
max: 18446744073709.47ms
approx. 95 percentile: 0.48ms

Threads fairness:
events (avg/stddev): 10000.0000/0.00
execution time (avg/stddev): 2.9343/0.00

C:MessAroundsysbench-0.4.12sysbench-0.4.12sysbenchRelWithDebInfo>sysbench.e
xe –test=fileio –file-total-size=3G –file-test-mode=rndrd –max-time=300 run
sysbench 0.4: multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
128 files, 24Mb each
3Gb total file size
Block size 16Kb
Number of random requests for random IO: 10000
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random read test
Threads started!
WARNING: Operation time (18446744073694841000.000000) is greater than maximal co
unted value, counting as 10000000000000.000000
WARNING: Percentile statistics will be inaccurate
Done.

Operations performed: 10000 Read, 0 Write, 0 Other = 10000 Total
Read 156.25Mb Written 0b Total transferred 156.25Mb (53.818Mb/sec)
3444.38 Requests/sec executed

Test execution summary:
total time: 2.9033s
total number of events: 10000
total time taken by event execution: 2.8777
per-request statistics:
min: 0.01ms
avg: 0.29ms
max: 18446744073696.34ms
approx. 95 percentile: 15.39ms

Threads fairness:
events (avg/stddev): 10000.0000/0.00
execution time (avg/stddev): 2.8777/0.00

Here are the detailed output from Amazon EC2 ami-c3e40daa with EBS device:
SQLIO
c:Program FilesSQLIO>sqlio -kR -t1 -s300 -dC -frandom -b16 -Fparam.txt -BH -LS

sqlio v1.5.SG
using system counter for latency timings, 3579545 counts per second
parameter file used: param.txt
file c:testfile.dat with 1 thread (0) using mask 0x0 (0)
1 thread reading for 300 secs from file c:testfile.dat
using 16KB random IOs
buffering set to use hardware disk cache (but not file cache)
size of file c:testfile.dat needs to be: 3221225472 bytes
current file size: 0 bytes
need to expand by: 3221225472 bytes
expanding c:testfile.dat … done.
using specified size: 3072 MB for file: c:testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec: 1230.94
MBs/sec: 19.23
latency metrics:
Min_Latency(ms): 0
Avg_Latency(ms): 0
Max_Latency(ms): 204
histogram:
ms: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%: 98 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

c:Program FilesSQLIO>sqlio -kR -t1 -s300 -dC -frandom -b16 -Fparam.txt -BH -LS

Cached

sqlio v1.5.SG
using system counter for latency timings, 3579545 counts per second
parameter file used: param.txt
file c:testfile.dat with 1 thread (0) using mask 0x0 (0)
1 thread reading for 300 secs from file c:testfile.dat
using 16KB random IOs
buffering set to use hardware disk cache (but not file cache)
using specified size: 3072 MB for file: c:testfile.dat
initialization done
CUMULATIVE DATA:
throughput metrics:
IOs/sec: 126.88
MBs/sec: 1.98
latency metrics:
Min_Latency(ms): 0
Avg_Latency(ms): 7
Max_Latency(ms): 497
histogram:
ms: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24+
%: 13 9 0 3 7 8 8 8 8 8 8 8 2 1 1 1 1 1 1 1 1 0 0 0 2

C:UsersAdministratorDocumentssysbench-0.4.12sysbenchRelWithDebInfo>sysbenc
h.exe –test=fileio –file-total-size=3G –file-test-mode=rndrd –max-time=300 r
un
sysbench 0.4: multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
128 files, 24Mb each
3Gb total file size
Block size 16Kb
Number of random requests for random IO: 10000
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random read test
Threads started!
Done.

Operations performed: 10000 Read, 0 Write, 0 Other = 10000 Total
Read 156.25Mb Written 0b Total transferred 156.25Mb (10.64Mb/sec)
680.95 Requests/sec executed

Test execution summary:
total time: 14.6854s
total number of events: 10000
total time taken by event execution: 14.6048
per-request statistics:
min: 0.01ms
avg: 1.46ms
max: 150.29ms
approx. 95 percentile: 4.77ms

Threads fairness:
events (avg/stddev): 10000.0000/0.00
execution time (avg/stddev): 14.6048/0.00

C:UsersAdministratorDocumentssysbench-0.4.12sysbenchRelWithDebInfo>sysbenc
h.exe –test=fileio –file-total-size=3G –file-test-mode=rndrd –max-time=300 r
un
sysbench 0.4: multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 1

Extra file open flags: 0
128 files, 24Mb each
3Gb total file size
Block size 16Kb
Number of random requests for random IO: 10000
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random read test
Threads started!
Done.

Sysbench Download (DEB, RPM, TGZ, TXZ, XZ, ZST)

Operations performed: 10000 Read, 0 Write, 0 Other = 10000 Total
Read 156.25Mb Written 0b Total transferred 156.25Mb (2.1371Mb/sec)
136.77 Requests/sec executed

MySQL :: MySQL Benchmark Tool

Test execution summary:
total time: 73.1139s
total number of events: 10000
total time taken by event execution: 73.0284
per-request statistics:
min: 0.02ms
avg: 7.30ms
max: 728.84ms
approx. 95 percentile: 23.08ms

Sysbench Windows 安装

Threads fairness:
events (avg/stddev): 10000.0000/0.00
execution time (avg/stddev): 73.0284/0.00