鲲鹏平台centos7.6使用UnixBench进行性能测试

1. 环境如下：

• 硬件：华为鲲鹏云 ECS（4vCPUs | 16GB）
• OS：CentOS 7.6 / Kernel 4.18

2. 安装支持库

yum groupinstall "Development Tools" -y
yum install libX11-devel mesa-libGL-devel perl-Time-HiRes -y

3. 下载UnixBench(正常情况不用不可描述的工具是下载不了的)

cd /opt/source
wget https://storage.googleapis.com/google-code-archive-downloads/v2/code.google.com/byte-unixbench/UnixBench5.1.3.tgz
tar xvfz UnixBench5.1.3.tgz
mv UnixBench /opt/tools/installed/

4. 修改make文件关闭图形测试

cd /opt/tools/installed/UnixBench
vim Makefile

注释掉 GRAPHICS_TESTS = defined

# GRAPHICS TESTS: Uncomment the definition of "GRAPHIC_TESTS" to enable
# the building of the graphics benchmarks.  This will require the
# X11 libraries on your system.
#
# Comment the line out to disable these tests.
# GRAPHIC_TESTS = defined

5. 执行测试

./Run 

6. 测试结果(大于1000分应该就还好)。从结果看来，感觉没有认出cpu数，一直按照单核测试方法测试，所有结果应该不是特别准确。

========================================================================
   BYTE UNIX Benchmarks (Version 5.1.3)

   System: test-1: GNU/Linux
   OS: GNU/Linux -- 4.18.0-80.7.2.el7.aarch64 -- #1 SMP Thu Sep 12 16:13:20 UTC 2019
   Machine: aarch64 (aarch64)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   14:38:19 up 1 day,  3:33,  2 users,  load average: 0.00, 0.00, 0.00; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Thu Jul 16 2020 14:38:19 - 15:06:28
0 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       27969004.9 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3497.4 MWIPS (10.0 s, 7 samples)
Execl Throughput                               4935.3 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        516728.7 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          139055.7 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1657275.1 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1167334.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 116831.8 lps   (10.0 s, 7 samples)
Process Creation                               9625.6 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   7959.4 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2368.1 lpm   (60.0 s, 2 samples)
System Call Overhead                        1026532.0 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   27969004.9   2396.7
Double-Precision Whetstone                       55.0       3497.4    635.9
Execl Throughput                                 43.0       4935.3   1147.7
File Copy 1024 bufsize 2000 maxblocks          3960.0     516728.7   1304.9
File Copy 256 bufsize 500 maxblocks            1655.0     139055.7    840.2
File Copy 4096 bufsize 8000 maxblocks          5800.0    1657275.1   2857.4
Pipe Throughput                               12440.0    1167334.6    938.4
Pipe-based Context Switching                   4000.0     116831.8    292.1
Process Creation                                126.0       9625.6    763.9
Shell Scripts (1 concurrent)                     42.4       7959.4   1877.2
Shell Scripts (8 concurrent)                      6.0       2368.1   3946.8
System Call Overhead                          15000.0    1026532.0    684.4
                                                                   ========
System Benchmarks Index Score                                        1158.0

说明：