CPU unixbench 对比: Intel N6000 VS N3540 - 软路由&小主机&家用服务器

  |   0 评论   |   2,802 浏览

平台基本信息

测试平台: Ubuntu 20.04 LTS

N6000 平台信息:

  1. 内存 DDR4-3200HZ (实际运行频率2933) / 16G
  2. 磁盘: 致钛 TiPlus 5000 1T

N3540 平台信息:

  1. 内存: DDR3L-1600MHZ 8G (镁光)
  2. Kingston KC600 SATA 1T (金士顿)

测试报告

N3540 平台

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

   System: HomeNas: GNU/Linux
   OS: GNU/Linux -- 5.15.0-91-generic -- #101~20.04.1-Ubuntu SMP Thu Nov 16 14:22:28 UTC 2023
   Machine: x86_64 (x86_64)
   Language: C.UTF-8 (charmap="UTF-8", collate="ANSI_X3.4-1968")
   CPU 0: Intel(R) Pentium(R) CPU N3540 @ 2.16GHz (4333.3 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Pentium(R) CPU N3540 @ 2.16GHz (4333.3 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 2: Intel(R) Pentium(R) CPU N3540 @ 2.16GHz (4333.3 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 3: Intel(R) Pentium(R) CPU N3540 @ 2.16GHz (4333.3 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   19:02:35 up 20 min,  0 users,  load average: 0.72, 0.33, 0.29; runlevel

------------------------------------------------------------------------
Benchmark Run: Sun Dec 24 2023 19:02:35 - 19:26:54
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       13604744.1 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     2298.3 MWIPS (10.1 s, 7 samples)
Execl Throughput                                838.6 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        128043.9 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           33354.2 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        417489.8 KBps  (30.0 s, 2 samples)
Pipe Throughput                              157660.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  35332.5 lps   (10.0 s, 7 samples)
Process Creation                               1940.6 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   1924.0 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    865.5 lpm   (60.0 s, 2 samples)
System Call Overhead                          98581.6 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   13604744.1   1165.8
Double-Precision Whetstone                       55.0       2298.3    417.9
Execl Throughput                                 43.0        838.6    195.0
File Copy 1024 bufsize 2000 maxblocks          3960.0     128043.9    323.3
File Copy 256 bufsize 500 maxblocks            1655.0      33354.2    201.5
File Copy 4096 bufsize 8000 maxblocks          5800.0     417489.8    719.8
Pipe Throughput                               12440.0     157660.6    126.7
Pipe-based Context Switching                   4000.0      35332.5     88.3
Process Creation                                126.0       1940.6    154.0
Shell Scripts (1 concurrent)                     42.4       1924.0    453.8
Shell Scripts (8 concurrent)                      6.0        865.5   1442.5
System Call Overhead                          15000.0      98581.6     65.7
                                                                   ========
System Benchmarks Index Score                                         288.4

------------------------------------------------------------------------
Benchmark Run: Sun Dec 24 2023 19:26:54 - 19:51:13
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       54093690.3 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     9135.1 MWIPS (10.1 s, 7 samples)
Execl Throughput                               2680.3 lps   (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        302028.4 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           77325.5 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1109521.4 KBps  (30.0 s, 2 samples)
Pipe Throughput                              313467.6 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 106772.7 lps   (10.0 s, 7 samples)
Process Creation                               9710.9 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   7391.6 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    986.4 lpm   (60.2 s, 2 samples)
System Call Overhead                         156807.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   54093690.3   4635.3
Double-Precision Whetstone                       55.0       9135.1   1660.9
Execl Throughput                                 43.0       2680.3    623.3
File Copy 1024 bufsize 2000 maxblocks          3960.0     302028.4    762.7
File Copy 256 bufsize 500 maxblocks            1655.0      77325.5    467.2
File Copy 4096 bufsize 8000 maxblocks          5800.0    1109521.4   1913.0
Pipe Throughput                               12440.0     313467.6    252.0
Pipe-based Context Switching                   4000.0     106772.7    266.9
Process Creation                                126.0       9710.9    770.7
Shell Scripts (1 concurrent)                     42.4       7391.6   1743.3
Shell Scripts (8 concurrent)                      6.0        986.4   1644.1
System Call Overhead                          15000.0     156807.1    104.5
                                                                   ========
System Benchmarks Index Score                                         780.1

N6000 平台报告

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

   System: NAS-N6000-4L: GNU/Linux
   OS: GNU/Linux -- 5.15.0-46-generic -- #49~20.04.1-Ubuntu SMP Thu Aug 4 19:15:44 UTC 2022
   Machine: x86_64 (x86_64)
   Language: en_US.UTF-8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
   CPU 0: Intel(R) Pentium(R) Silver N6000 @ 1.10GHz (2227.2 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Pentium(R) Silver N6000 @ 1.10GHz (2227.2 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 2: Intel(R) Pentium(R) Silver N6000 @ 1.10GHz (2227.2 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 3: Intel(R) Pentium(R) Silver N6000 @ 1.10GHz (2227.2 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   18:37:28 up 14 min,  0 users,  load average: 0.73, 0.88, 0.55; runlevel

------------------------------------------------------------------------
Benchmark Run: Sun Dec 24 2023 18:37:28 - 19:01:55
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       39811766.2 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4647.6 MWIPS (10.0 s, 7 samples)
Execl Throughput                               3897.5 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1033014.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          305061.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       2258300.9 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1755039.9 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 167553.7 lps   (10.0 s, 7 samples)
Process Creation                               9326.5 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  10948.8 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   3050.0 lpm   (60.0 s, 2 samples)
System Call Overhead                        1796560.3 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   39811766.2   3411.5
Double-Precision Whetstone                       55.0       4647.6    845.0
Execl Throughput                                 43.0       3897.5    906.4
File Copy 1024 bufsize 2000 maxblocks          3960.0    1033014.0   2608.6
File Copy 256 bufsize 500 maxblocks            1655.0     305061.4   1843.3
File Copy 4096 bufsize 8000 maxblocks          5800.0    2258300.9   3893.6
Pipe Throughput                               12440.0    1755039.9   1410.8
Pipe-based Context Switching                   4000.0     167553.7    418.9
Process Creation                                126.0       9326.5    740.2
Shell Scripts (1 concurrent)                     42.4      10948.8   2582.3
Shell Scripts (8 concurrent)                      6.0       3050.0   5083.4
System Call Overhead                          15000.0    1796560.3   1197.7
                                                                   ========
System Benchmarks Index Score                                        1624.0

------------------------------------------------------------------------
Benchmark Run: Sun Dec 24 2023 19:01:55 - 19:26:28
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables      136531621.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    17086.2 MWIPS (9.9 s, 7 samples)
Execl Throughput                              12111.3 lps   (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks       1852416.0 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          541963.2 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       3906333.8 KBps  (30.0 s, 2 samples)
Pipe Throughput                             5007870.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 290175.0 lps   (10.0 s, 7 samples)
Process Creation                              20995.7 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  22688.9 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   3216.2 lpm   (60.0 s, 2 samples)
System Call Overhead                        5319351.5 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0  136531621.8  11699.4
Double-Precision Whetstone                       55.0      17086.2   3106.6
Execl Throughput                                 43.0      12111.3   2816.6
File Copy 1024 bufsize 2000 maxblocks          3960.0    1852416.0   4677.8
File Copy 256 bufsize 500 maxblocks            1655.0     541963.2   3274.7
File Copy 4096 bufsize 8000 maxblocks          5800.0    3906333.8   6735.1
Pipe Throughput                               12440.0    5007870.1   4025.6
Pipe-based Context Switching                   4000.0     290175.0    725.4
Process Creation                                126.0      20995.7   1666.3
Shell Scripts (1 concurrent)                     42.4      22688.9   5351.2
Shell Scripts (8 concurrent)                      6.0       3216.2   5360.4
System Call Overhead                          15000.0    5319351.5   3546.2
                                                                   ========
System Benchmarks Index Score                                        3629.6

报告分析与对比

MWIPS: 每秒百万次韦氏指令
这个单位来源于一种被称为韦氏（Whetstone）的早期浮点基准测试。这个测量标准在1970年代由英国的沃森实验室（National Physical Laboratory）首次发布，用于评估一台计算机每秒钟可以执行多少次特定的算术和逻辑操作。

韦氏基准通过一组具有代表性的算术和逻辑操作对CPU进行测试，然后以每秒执行多少次这些操作的速度来评估性能。这种评估方式受许多因素（如系统内存、系统架构、指令集等）的影响。

因此，MWIPS并不是用来衡量CPU频率或者核心数量的，而是一种评估计算机进行浮点运算（如小数运算和科学计数法运算）的性能的单位。

定点性能 (单核)

  • N3540 :

    • 测试结果: 13604744.1 lps
    • 评分基线: 116700.0 (11.67万)
    • 实际得分: 13604744.1 (1360万)
    • Index积分: 1165.8分
  • N6000 :
    • 测试结果: 39811766.2
    • 评分基线: 116700.0 (11.67万)
    • 实际得分: 39811766.2 (3981万)
    • Index积分: 3411.5分

定点性能 (多核)

  • N3540 :
    • 测试结果: 54093690.3 lps
    • 评分基线: 116700.0 (11.67万)
    • 实际得分: 54093690.3 (5409万)
    • Index积分: 4635.3分
  • N6000
    • 测试结果: 136531621.8 lps
    • 评分基线: 116700.0 (11.67万)
    • 实际得分: 136531621.8 (13653万)
    • Index积分: 11699.4分

浮点性能(单核)

  • N3540
    • 测试结果: 2298.3 MWIPS
    • 评分基线: 55.0
    • 实际得分: 2298.3
    • Index积分: 417.9分
  • N6000
    • 测试结果: 4647.6 MWIPS
    • 评分基线: 55.0
    • 实际得分: 4647.6
    • Index积分: 845.0分

浮点性能(多核)

  • N3540
    • 测试结果: 9135.1 MWIPS
    • 评分基线: 55.0
    • 实际得分: 9135.1
    • Index积分: 1660.9分
  • N6000
    • 测试结果: 17086.2 MWIPS
    • 评分基线: 55.0
    • 实际得分: 17086.2 (17086.2)
    • Index积分: 3106.6分

结论

定点:

  • 单核: n6000 相对于 N3540 定点性能提升: 3981/1360 = 2.92倍;
  • 多核:n6000相对于N3540 定点性能提升: *13653/5409=2.52倍;

浮点:

  • 单核: n6000 相对于 N3540 定点性能提升4647.6/2298.3=2.02倍
  • 多核: n6000 相对于 N3540 定点性能提升17086.2/9135.1=1.87倍

综合得分:

单核: N6000相对于N3540 提升: 1624.0/288.4=5.63倍
多核: N6000相对于N3540 提升: 3629.6/780.1=4.65倍

评论

发表评论


取消