很久没有写点什么了,列一下这几年对各家VPS及云服务器的性能测试吧。

很多都不包含硬盘测试,因为SSD的话一般都在300以上,至少200+,只要没有特别大的出入,则实际使用中根本感觉不出什么区别,而HDD的话,很少见到超过120的,只要不是慢的离谱,也可以接受。

 

先列一下scaleway的几个常见型号:

ARM64-2GB:

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

   System: scw-8afeb0: GNU/Linux
   OS: GNU/Linux -- 4.9.41-mainline-rev1 -- #1 SMP Mon Aug 7 03:24:06 UTC 2017
   Machine: aarch64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   07:46:24 up 3 min,  1 user,  load average: 0.18, 0.12, 0.04; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Sat Aug 12 2017 07:46:24 - 08:14:23
0 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables        8960310.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                      991.4 MWIPS (9.9 s, 7 samples)
Execl Throughput                               1156.4 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        131939.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           48281.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        316398.3 KBps  (30.0 s, 2 samples)
Pipe Throughput                              455721.8 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  13913.2 lps   (10.0 s, 7 samples)
Process Creation                               2254.7 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   2560.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    767.5 lpm   (60.0 s, 2 samples)
System Call Overhead                         944730.7 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0    8960310.8    767.8
Double-Precision Whetstone                       55.0        991.4    180.2
Execl Throughput                                 43.0       1156.4    268.9
File Copy 1024 bufsize 2000 maxblocks          3960.0     131939.0    333.2
File Copy 256 bufsize 500 maxblocks            1655.0      48281.4    291.7
File Copy 4096 bufsize 8000 maxblocks          5800.0     316398.3    545.5
Pipe Throughput                               12440.0     455721.8    366.3
Pipe-based Context Switching                   4000.0      13913.2     34.8
Process Creation                                126.0       2254.7    178.9
Shell Scripts (1 concurrent)                     42.4       2560.0    603.8
Shell Scripts (8 concurrent)                      6.0        767.5   1279.1
System Call Overhead                          15000.0     944730.7    629.8
                                                                   ========
System Benchmarks Index Score                                         337.6

 

硬盘写入:

root@scw-8afeb0:~/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.61336 s, 119 MB/s

real    0m8.619s
user    0m0.280s
sys     0m8.308s
root@scw-8afeb0:~/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.92673 s, 115 MB/s

real    0m9.685s
user    0m0.240s
sys     0m8.552s
root@scw-8afeb0:~/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.77997 s, 117 MB/s

real    0m9.454s
user    0m0.196s
sys     0m8.512s

 

读取

root@scw-8afeb0:~/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied2000000+0 records in
2000000+0 records out
, 8.88969 s, 115 MB/s
1024000000 bytes (1.0 GB) copied, 8.88981 s, 115 MB/s
root@scw-8afeb0:~/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied, 9.27607 s, 110 MB/s
2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 9.27627 s, 110 MB/s
root@scw-8afeb0:~/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied, 8.70826 s, 118 MB/s
2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.7084 s, 118 MB/s

 

 

 

vc1s型号:

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

   System: web: GNU/Linux
   OS: GNU/Linux -- 4.4.80-mainline-rev1 -- #1 SMP Mon Aug 7 02:42:53 UTC 2017
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: Intel(R) Atom(TM) CPU C2750 @ 2.40GHz (4787.8 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Atom(TM) CPU C2750 @ 2.40GHz (4787.8 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   06:00:43 up 1 min,  2 users,  load average: 0.12, 0.03, 0.01; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Sat Aug 12 2017 06:00:43 - 06:28:34
2 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       12028204.7 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     1986.0 MWIPS (9.9 s, 7 samples)
Execl Throughput                               2371.8 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        387524.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          113878.1 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1009884.2 KBps  (30.0 s, 2 samples)
Pipe Throughput                              977398.8 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  43325.2 lps   (10.0 s, 7 samples)
Process Creation                               4644.4 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   3668.1 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    809.8 lpm   (60.0 s, 2 samples)
System Call Overhead                        1725605.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   12028204.7   1030.7
Double-Precision Whetstone                       55.0       1986.0    361.1
Execl Throughput                                 43.0       2371.8    551.6
File Copy 1024 bufsize 2000 maxblocks          3960.0     387524.5    978.6
File Copy 256 bufsize 500 maxblocks            1655.0     113878.1    688.1
File Copy 4096 bufsize 8000 maxblocks          5800.0    1009884.2   1741.2
Pipe Throughput                               12440.0     977398.8    785.7
Pipe-based Context Switching                   4000.0      43325.2    108.3
Process Creation                                126.0       4644.4    368.6
Shell Scripts (1 concurrent)                     42.4       3668.1    865.1
Shell Scripts (8 concurrent)                      6.0        809.8   1349.7
System Call Overhead                          15000.0    1725605.1   1150.4
                                                                   ========
System Benchmarks Index Score                                         682.1

------------------------------------------------------------------------
Benchmark Run: Sat Aug 12 2017 06:28:34 - 06:56:28
2 CPUs in system; running 2 parallel copies of tests

Dhrystone 2 using register variables       24059549.2 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3972.2 MWIPS (9.9 s, 7 samples)
Execl Throughput                               4141.1 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        431084.2 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          124690.9 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1234728.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1954477.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 200515.8 lps   (10.0 s, 7 samples)
Process Creation                              11479.8 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   5581.1 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    841.4 lpm   (60.1 s, 2 samples)
System Call Overhead                        2707984.2 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   24059549.2   2061.7
Double-Precision Whetstone                       55.0       3972.2    722.2
Execl Throughput                                 43.0       4141.1    963.1
File Copy 1024 bufsize 2000 maxblocks          3960.0     431084.2   1088.6
File Copy 256 bufsize 500 maxblocks            1655.0     124690.9    753.4
File Copy 4096 bufsize 8000 maxblocks          5800.0    1234728.5   2128.8
Pipe Throughput                               12440.0    1954477.4   1571.1
Pipe-based Context Switching                   4000.0     200515.8    501.3
Process Creation                                126.0      11479.8    911.1
Shell Scripts (1 concurrent)                     42.4       5581.1   1316.3
Shell Scripts (8 concurrent)                      6.0        841.4   1402.3
System Call Overhead                          15000.0    2707984.2   1805.3
                                                                   ========
System Benchmarks Index Score                                        1160.8

 

写入:

root@web:~/z/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.1613 s, 125 MB/s

real    0m8.164s
user    0m0.610s
sys     0m6.120s
root@web:~/z/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 6.93709 s, 148 MB/s

real    0m7.335s
user    0m0.320s
sys     0m5.460s
root@web:~/z/z/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 7.84341 s, 131 MB/s

real    0m8.059s
user    0m0.420s
sys     0m5.390s

读取:

root@web:~/z/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 5.52652 s, 185 MB/s
, 5.52806 s, 185 MB/s
root@web:~/z/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 5.08371 s, 201 MB/s
, 5.08531 s, 201 MB/s
root@web:~/z/z/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 4.97994 s, 206 MB/s

 

C1型号:

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

   System: scw-1c5a9e: GNU/Linux
   OS: GNU/Linux -- 4.4.80-mainline-rev1 -- #1 SMP Mon Aug 7 02:31:43 UTC 2017
   Machine: armv7l (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: ARMv7 Processor rev 2 (v7l) (0.0 bogomips)
          
   CPU 1: ARMv7 Processor rev 2 (v7l) (0.0 bogomips)
          
   CPU 2: ARMv7 Processor rev 2 (v7l) (0.0 bogomips)
          
   CPU 3: ARMv7 Processor rev 2 (v7l) (0.0 bogomips)
          
   07:48:08 up 4 min,  1 user,  load average: 0.99, 0.65, 0.28; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Sat Aug 12 2017 07:48:08 - 08:16:23
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables        3714895.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                      556.5 MWIPS (10.0 s, 7 samples)
Execl Throughput                                970.3 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        138715.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           33466.3 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        351646.7 KBps  (30.0 s, 2 samples)
Pipe Throughput                              288027.5 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  51566.7 lps   (10.0 s, 7 samples)
Process Creation                               1714.2 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   1503.1 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    498.3 lpm   (60.1 s, 2 samples)
System Call Overhead                         677612.3 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0    3714895.6    318.3
Double-Precision Whetstone                       55.0        556.5    101.2
Execl Throughput                                 43.0        970.3    225.6
File Copy 1024 bufsize 2000 maxblocks          3960.0     138715.0    350.3
File Copy 256 bufsize 500 maxblocks            1655.0      33466.3    202.2
File Copy 4096 bufsize 8000 maxblocks          5800.0     351646.7    606.3
Pipe Throughput                               12440.0     288027.5    231.5
Pipe-based Context Switching                   4000.0      51566.7    128.9
Process Creation                                126.0       1714.2    136.0
Shell Scripts (1 concurrent)                     42.4       1503.1    354.5
Shell Scripts (8 concurrent)                      6.0        498.3    830.4
System Call Overhead                          15000.0     677612.3    451.7
                                                                   ========
System Benchmarks Index Score                                         272.6

------------------------------------------------------------------------
Benchmark Run: Sat Aug 12 2017 08:16:23 - 08:44:42
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       14897276.4 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     2226.7 MWIPS (9.9 s, 7 samples)
Execl Throughput                               2725.9 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        186024.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           61822.8 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        585502.3 KBps  (30.0 s, 2 samples)
Pipe Throughput                              844841.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 195209.4 lps   (10.0 s, 7 samples)
Process Creation                               3929.4 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   4053.4 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    569.2 lpm   (60.4 s, 2 samples)
System Call Overhead                        2423953.9 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   14897276.4   1276.5
Double-Precision Whetstone                       55.0       2226.7    404.9
Execl Throughput                                 43.0       2725.9    633.9
File Copy 1024 bufsize 2000 maxblocks          3960.0     186024.5    469.8
File Copy 256 bufsize 500 maxblocks            1655.0      61822.8    373.6
File Copy 4096 bufsize 8000 maxblocks          5800.0     585502.3   1009.5
Pipe Throughput                               12440.0     844841.6    679.1
Pipe-based Context Switching                   4000.0     195209.4    488.0
Process Creation                                126.0       3929.4    311.9
Shell Scripts (1 concurrent)                     42.4       4053.4    956.0
Shell Scripts (8 concurrent)                      6.0        569.2    948.6
System Call Overhead                          15000.0    2423953.9   1616.0
                                                                   ========
System Benchmarks Index Score                                         674.0

 

写入:

root@scw-1c5a9e:~/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 13.9705 s, 73.3 MB/s

real    0m13.974s
user    0m0.880s
sys     0m13.030s
root@scw-1c5a9e:~/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 14.3673 s, 71.3 MB/s

real    0m14.982s
user    0m0.860s
sys     0m13.570s
root@scw-1c5a9e:~/UnixBench# time dd if=/dev/zero bs=1024 count=1000000 of=/1Gb.file
1000000+0 records in
1000000+0 records out
1024000000 bytes (1.0 GB) copied, 14.9229 s, 68.6 MB/s

real    0m15.522s
user    0m0.720s
sys     0m14.370s

读取:

root@scw-1c5a9e:~/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied, 9.4293 s, 109 MB/s
2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 9.43144 s, 109 MB/s
root@scw-1c5a9e:~/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied, 8.13743 s, 126 MB/s
2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.13825 s, 126 MB/s
root@scw-1c5a9e:~/UnixBench# dd if=/1Gb.file bs=64k |dd of=/dev/null
15625+0 records in
15625+0 records out
1024000000 bytes (1.0 GB) copied2000000+0 records in
2000000+0 records out
1024000000 bytes (1.0 GB) copied, 8.43918 s, 121 MB/s

 

法国机房的网速测试:

Speedtest (IPv4 only)
---------------------
Your public IPv4 is 51.15.140.234

Location                Provider        Speed
CDN                     Cachefly        96.4MB/s

Atlanta, GA, US         Coloat          4.56MB/s 
Dallas, TX, US          Softlayer       9.58MB/s 
Seattle, WA, US         Softlayer       8.72MB/s 
San Jose, CA, US        Softlayer       8.49MB/s 
Washington, DC, US      Softlayer       9.57MB/s 

Tokyo, Japan            Linode          8.91MB/s 
Singapore               Softlayer       3.19MB/s 

Rotterdam, Netherlands  id3.net         27.9MB/s
Haarlem, Netherlands    Leaseweb        70.5MB/s

 

荷兰增强版的VC1S:

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

   System: eu-free-02: GNU/Linux
   OS: GNU/Linux -- 4.5.7-std-3 -- #1 SMP Tue Jul 12 09:56:30 UTC 2016
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: Intel(R) Xeon(R) CPU D-1531 @ 2.20GHz (4400.0 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Xeon(R) CPU D-1531 @ 2.20GHz (4400.0 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   07:46:00 up 16 min,  1 user,  load average: 0.15, 0.18, 0.10; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Sat Mar 24 2018 07:46:00 - 08:14:04
2 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       27433625.9 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4407.2 MWIPS (9.7 s, 7 samples)
Execl Throughput                               5125.6 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        933049.1 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          258986.3 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       2168940.3 KBps  (30.0 s, 2 samples)
Pipe Throughput                             2042835.9 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 342742.7 lps   (10.0 s, 7 samples)
Process Creation                              10944.6 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   8163.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2015.8 lpm   (60.0 s, 2 samples)
System Call Overhead                        3342947.2 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   27433625.9   2350.8
Double-Precision Whetstone                       55.0       4407.2    801.3
Execl Throughput                                 43.0       5125.6   1192.0
File Copy 1024 bufsize 2000 maxblocks          3960.0     933049.1   2356.2
File Copy 256 bufsize 500 maxblocks            1655.0     258986.3   1564.9
File Copy 4096 bufsize 8000 maxblocks          5800.0    2168940.3   3739.6
Pipe Throughput                               12440.0    2042835.9   1642.2
Pipe-based Context Switching                   4000.0     342742.7    856.9
Process Creation                                126.0      10944.6    868.6
Shell Scripts (1 concurrent)                     42.4       8163.0   1925.2
Shell Scripts (8 concurrent)                      6.0       2015.8   3359.6
System Call Overhead                          15000.0    3342947.2   2228.6
                                                                   ========
System Benchmarks Index Score                                        1690.7

------------------------------------------------------------------------
Benchmark Run: Sat Mar 24 2018 08:14:04 - 08:42:06
2 CPUs in system; running 2 parallel copies of tests

Dhrystone 2 using register variables       54299327.6 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     8791.0 MWIPS (9.4 s, 7 samples)
Execl Throughput                               9858.2 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1042399.1 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          279293.7 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       3091948.6 KBps  (30.0 s, 2 samples)
Pipe Throughput                             3938964.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 763087.5 lps   (10.0 s, 7 samples)
Process Creation                              24464.4 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  15176.9 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2147.4 lpm   (60.0 s, 2 samples)
System Call Overhead                        4676475.2 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   54299327.6   4652.9
Double-Precision Whetstone                       55.0       8791.0   1598.4
Execl Throughput                                 43.0       9858.2   2292.6
File Copy 1024 bufsize 2000 maxblocks          3960.0    1042399.1   2632.3
File Copy 256 bufsize 500 maxblocks            1655.0     279293.7   1687.6
File Copy 4096 bufsize 8000 maxblocks          5800.0    3091948.6   5330.9
Pipe Throughput                               12440.0    3938964.6   3166.4
Pipe-based Context Switching                   4000.0     763087.5   1907.7
Process Creation                                126.0      24464.4   1941.6
Shell Scripts (1 concurrent)                     42.4      15176.9   3579.5
Shell Scripts (8 concurrent)                      6.0       2147.4   3578.9
System Call Overhead                          15000.0    4676475.2   3117.7
                                                                   ========
System Benchmarks Index Score                                        2752.5

 

新版c3955的vc1s:

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

   System: eu-free: GNU/Linux
   OS: GNU/Linux -- 4.15.11-mainline-rev1 -- #1 SMP Mon Mar 19 08:22:48 UTC 2018
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: Intel(R) Atom(TM) CPU C3955 @ 2.10GHz (4200.0 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Atom(TM) CPU C3955 @ 2.10GHz (4200.0 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   04:02:35 up 4 min,  1 user,  load average: 0.28, 0.11, 0.03; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Sat Mar 31 2018 04:02:35 - 04:30:47
2 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       21754711.0 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3208.3 MWIPS (9.9 s, 7 samples)
Execl Throughput                               2329.0 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        154341.9 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           40693.1 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        506775.8 KBps  (30.0 s, 2 samples)
Pipe Throughput                              202597.7 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  47205.9 lps   (10.0 s, 7 samples)
Process Creation                               5313.1 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   4971.1 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   1010.8 lpm   (60.0 s, 2 samples)
System Call Overhead                         169301.6 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   21754711.0   1864.2
Double-Precision Whetstone                       55.0       3208.3    583.3
Execl Throughput                                 43.0       2329.0    541.6
File Copy 1024 bufsize 2000 maxblocks          3960.0     154341.9    389.8
File Copy 256 bufsize 500 maxblocks            1655.0      40693.1    245.9
File Copy 4096 bufsize 8000 maxblocks          5800.0     506775.8    873.8
Pipe Throughput                               12440.0     202597.7    162.9
Pipe-based Context Switching                   4000.0      47205.9    118.0
Process Creation                                126.0       5313.1    421.7
Shell Scripts (1 concurrent)                     42.4       4971.1   1172.4
Shell Scripts (8 concurrent)                      6.0       1010.8   1684.6
System Call Overhead                          15000.0     169301.6    112.9
                                                                   ========
System Benchmarks Index Score                                         459.7

------------------------------------------------------------------------
Benchmark Run: Sat Mar 31 2018 04:30:47 - 04:58:58
2 CPUs in system; running 2 parallel copies of tests

Dhrystone 2 using register variables       43470827.4 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     6410.6 MWIPS (9.9 s, 7 samples)
Execl Throughput                               4274.2 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        283951.8 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           72614.0 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        942333.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                              404175.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 121001.9 lps   (10.0 s, 7 samples)
Process Creation                              11243.7 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   7693.6 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   1034.5 lpm   (60.0 s, 2 samples)
System Call Overhead                         336578.4 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   43470827.4   3725.0
Double-Precision Whetstone                       55.0       6410.6   1165.6
Execl Throughput                                 43.0       4274.2    994.0
File Copy 1024 bufsize 2000 maxblocks          3960.0     283951.8    717.0
File Copy 256 bufsize 500 maxblocks            1655.0      72614.0    438.8
File Copy 4096 bufsize 8000 maxblocks          5800.0     942333.5   1624.7
Pipe Throughput                               12440.0     404175.1    324.9
Pipe-based Context Switching                   4000.0     121001.9    302.5
Process Creation                                126.0      11243.7    892.4
Shell Scripts (1 concurrent)                     42.4       7693.6   1814.5
Shell Scripts (8 concurrent)                      6.0       1034.5   1724.1
System Call Overhead                          15000.0     336578.4    224.4
                                                                   ========
System Benchmarks Index Score                                         846.7

 

 

一台自用的AMD Phenom(tm) II X4 965物理机:

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

   System: us2.tingtao.org: GNU/Linux
   OS: GNU/Linux -- 3.16.0-4-amd64 -- #1 SMP Debian 3.16.51-3 (2017-12-13)
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: AMD Phenom(tm) II X4 965 Processor (6830.4 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 1: AMD Phenom(tm) II X4 965 Processor (6830.4 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 2: AMD Phenom(tm) II X4 965 Processor (6830.4 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 3: AMD Phenom(tm) II X4 965 Processor (6830.4 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   20:32:22 up  7:48,  1 user,  load average: 0.23, 0.30, 0.41; runlevel 5

------------------------------------------------------------------------
Benchmark Run: Wed May 09 2018 20:32:22 - 21:00:51
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       30443823.4 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3753.0 MWIPS (9.8 s, 7 samples)
Execl Throughput                               1744.7 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1025499.3 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          311434.7 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1398652.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                             2130794.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 135683.0 lps   (10.0 s, 7 samples)
Process Creation                               7635.3 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   4919.2 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2559.7 lpm   (60.0 s, 2 samples)
System Call Overhead                        4049371.0 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   30443823.4   2608.7
Double-Precision Whetstone                       55.0       3753.0    682.4
Execl Throughput                                 43.0       1744.7    405.7
File Copy 1024 bufsize 2000 maxblocks          3960.0    1025499.3   2589.6
File Copy 256 bufsize 500 maxblocks            1655.0     311434.7   1881.8
File Copy 4096 bufsize 8000 maxblocks          5800.0    1398652.5   2411.5
Pipe Throughput                               12440.0    2130794.6   1712.9
Pipe-based Context Switching                   4000.0     135683.0    339.2
Process Creation                                126.0       7635.3    606.0
Shell Scripts (1 concurrent)                     42.4       4919.2   1160.2
Shell Scripts (8 concurrent)                      6.0       2559.7   4266.2
System Call Overhead                          15000.0    4049371.0   2699.6
                                                                   ========
System Benchmarks Index Score                                        1359.7

------------------------------------------------------------------------
Benchmark Run: Wed May 09 2018 21:00:51 - 21:29:37
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables      122656916.4 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    15039.3 MWIPS (9.8 s, 7 samples)
Execl Throughput                              16738.9 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        932668.9 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          237327.0 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1798153.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                             8491247.2 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                1294803.7 lps   (10.0 s, 7 samples)
Process Creation                              39717.6 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  21292.5 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2836.8 lpm   (60.1 s, 2 samples)
System Call Overhead                        3831416.6 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0  122656916.4  10510.4
Double-Precision Whetstone                       55.0      15039.3   2734.4
Execl Throughput                                 43.0      16738.9   3892.8
File Copy 1024 bufsize 2000 maxblocks          3960.0     932668.9   2355.2
File Copy 256 bufsize 500 maxblocks            1655.0     237327.0   1434.0
File Copy 4096 bufsize 8000 maxblocks          5800.0    1798153.5   3100.3
Pipe Throughput                               12440.0    8491247.2   6825.8
Pipe-based Context Switching                   4000.0    1294803.7   3237.0
Process Creation                                126.0      39717.6   3152.2
Shell Scripts (1 concurrent)                     42.4      21292.5   5021.8
Shell Scripts (8 concurrent)                      6.0       2836.8   4728.1
System Call Overhead                          15000.0    3831416.6   2554.3
                                                                   ========
System Benchmarks Index Score                                        3613.0

 

一台自用的i3-3225:

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

   System: us1.tingtao.org: GNU/Linux
   OS: GNU/Linux -- 3.16.0-4-amd64 -- #1 SMP Debian 3.16.51-3 (2017-12-13)
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: Intel(R) Core(TM) i3-3225 CPU @ 3.30GHz (6585.0 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET,                     Intel virtualization
   CPU 1: Intel(R) Core(TM) i3-3225 CPU @ 3.30GHz (6585.0 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET,                     Intel virtualization
   CPU 2: Intel(R) Core(TM) i3-3225 CPU @ 3.30GHz (6585.0 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET,                     Intel virtualization
   CPU 3: Intel(R) Core(TM) i3-3225 CPU @ 3.30GHz (6585.0 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET,                     Intel virtualization
   10:41:24 up 2 days, 11:50,  1 user,  load average: 0.20, 0.06, 0.02; runlevel 5

------------------------------------------------------------------------
Benchmark Run: Thu May 10 2018 10:41:24 - 11:09:40
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       37551342.2 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4657.7 MWIPS (9.9 s, 7 samples)
Execl Throughput                               6733.2 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1140393.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          337710.5 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       2357659.4 KBps  (30.0 s, 2 samples)
Pipe Throughput                             2303563.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 245947.2 lps   (10.0 s, 7 samples)
Process Creation                              16197.1 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   9554.4 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2433.5 lpm   (60.0 s, 2 samples)
System Call Overhead                        4528742.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   37551342.2   3217.8
Double-Precision Whetstone                       55.0       4657.7    846.9
Execl Throughput                                 43.0       6733.2   1565.9
File Copy 1024 bufsize 2000 maxblocks          3960.0    1140393.0   2879.8
File Copy 256 bufsize 500 maxblocks            1655.0     337710.5   2040.5
File Copy 4096 bufsize 8000 maxblocks          5800.0    2357659.4   4064.9
Pipe Throughput                               12440.0    2303563.6   1851.7
Pipe-based Context Switching                   4000.0     245947.2    614.9
Process Creation                                126.0      16197.1   1285.5
Shell Scripts (1 concurrent)                     42.4       9554.4   2253.4
Shell Scripts (8 concurrent)                      6.0       2433.5   4055.8
System Call Overhead                          15000.0    4528742.1   3019.2
                                                                   ========
System Benchmarks Index Score                                        2000.8

------------------------------------------------------------------------
Benchmark Run: Thu May 10 2018 11:09:40 - 11:38:17
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       63398601.2 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    12656.4 MWIPS (11.0 s, 7 samples)
Execl Throughput                              14180.5 lps   (29.5 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1035170.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          294558.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       2625675.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                             4572994.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 988944.8 lps   (10.0 s, 7 samples)
Process Creation                              35239.9 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  13963.3 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   1683.2 lpm   (60.1 s, 2 samples)
System Call Overhead                        8852674.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   63398601.2   5432.6
Double-Precision Whetstone                       55.0      12656.4   2301.2
Execl Throughput                                 43.0      14180.5   3297.8
File Copy 1024 bufsize 2000 maxblocks          3960.0    1035170.0   2614.1
File Copy 256 bufsize 500 maxblocks            1655.0     294558.4   1779.8
File Copy 4096 bufsize 8000 maxblocks          5800.0    2625675.5   4527.0
Pipe Throughput                               12440.0    4572994.4   3676.0
Pipe-based Context Switching                   4000.0     988944.8   2472.4
Process Creation                                126.0      35239.9   2796.8
Shell Scripts (1 concurrent)                     42.4      13963.3   3293.2
Shell Scripts (8 concurrent)                      6.0       1683.2   2805.3
System Call Overhead                          15000.0    8852674.1   5901.8
                                                                   ========
System Benchmarks Index Score                                        3211.5

 

 

简单解释一下,其中scaleway的VC1S这个型号现在法国已经取消这个型号了,在我最后用的时间(大概5月份)法国那个新的start什么的型号是c3955的cpu,以前荷兰和法国的vc1s是统一的c2750,理论上3955比2750要牛b一些,但估计这家的软件还没优化好,实际效果是2750比3955强不少。但荷兰一直到5月份都还能时不时的开到xeon核心的vc1s机器,价格相同但性能提升巨大,直线提升150%,但是这个要看运气,实在想开的时候,连续开,多开,不停的开,直到开出来为止,运气不好的时候折腾两三天也开不到这个配置的。

关于vultr和do的测试,可以参考之前的文章:《Vultr和digitalocean的性能测试》,但最近vultr >=10美元的机器都改用新核心了,没测试过,应该性能有提升。

还有些常识需要注意的,我特意放出两个物理机的性能测试对比,可以看到两个整体得分一个是3600,一个是3200基本没区别,但单核心的2000和1300这个区别是很大的,我们一般操作和使用主要看的是单核心性能,能直接影响程序执行效率的也是单核心,而多核心性能只有在负载较高的时候才能得以体现,甚至在负载巨大的时候才会有意义。   就好比上面的c2750的多核心效率是1100,和物理机的单核1300理论上是差不多的,但实际效果差距很大,根本没有可比性。

最后就是,scaleway的机器呢,跑个博客网站下载什么的肯定够,再多的就别想了,访问量上来以后就不能用这个了,那cpu真的干不了什么。而且他的系统做的很奇怪,是他们自己编译的内核,有些莫名其妙的问题,还有些软件运行不了,真心不太明白他们这么折腾图个什么。而且5月份我决定把这里的机器全清了是因为他们出现过一次严重故障,荷兰机房的全部断掉,断了十几个小时,虽然说他母公司是欧洲数一数二的托管商,但是对于设备故障都要整个机房断掉十几个小时,我是受不了。

By 听涛

发表评论

电子邮件地址不会被公开。 必填项已用*标注