Результат работы UnixBench - может ли быть лучше на VDS?

Малыш, хватит уже плодить глупые темы на форуме. Если нечем заняться, лучше сделай домашнюю работу.

Ну вот вам (на продакшен ноде с нагрузкой по процу 60-80%).
И думаю, это далеко не предел. Я на пустой ноде в районе 5000 поинтов выжимал с VDS.
Debian 7.1 64bit
гипервизор VMWare
1 ядро, 1 ГБ памяти, 20 ГБ SSD с внешней системы хранения по iSCSI 10 Gbit

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: debian: GNU/Linux

OS: GNU/Linux -- 3.2.0-4-amd64 -- #1 SMP Debian 3.2.46-1+deb7u1

Machine: x86_64 (unknown)

Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")

CPU 0: Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz (5200.0 bogomips)

x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET

03:58:33 up 3 min, 1 user, load average: 0.00, 0.01, 0.01; runlevel 2

------------------------------------------------------------------------

Benchmark Run: Tue Dec 03 2013 03:58:33 - 04:26:36

1 CPU in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 35966098.2 lps (10.0 s, 7 samples)

Double-Precision Whetstone 4431.9 MWIPS (9.8 s, 7 samples)

Execl Throughput 4250.1 lps (29.6 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks 1107042.5 KBps (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks 338621.1 KBps (30.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks 1987290.5 KBps (30.0 s, 2 samples)

Pipe Throughput 2946954.1 lps (10.0 s, 7 samples)

Pipe-based Context Switching 226604.6 lps (10.0 s, 7 samples)

Process Creation 14540.4 lps (30.0 s, 2 samples)

Shell Scripts (1 concurrent) 7980.3 lpm (60.0 s, 2 samples)

Shell Scripts (8 concurrent) 1034.5 lpm (60.0 s, 2 samples)

System Call Overhead 4753371.3 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX

Dhrystone 2 using register variables 116700.0 35966098.2 3081.9

Double-Precision Whetstone 55.0 4431.9 805.8

Execl Throughput 43.0 4250.1 988.4

File Copy 1024 bufsize 2000 maxblocks 3960.0 1107042.5 2795.6

File Copy 256 bufsize 500 maxblocks 1655.0 338621.1 2046.0

File Copy 4096 bufsize 8000 maxblocks 5800.0 1987290.5 3426.4

Pipe Throughput 12440.0 2946954.1 2368.9

Pipe-based Context Switching 4000.0 226604.6 566.5

Process Creation 126.0 14540.4 1154.0

Shell Scripts (1 concurrent) 42.4 7980.3 1882.1

Shell Scripts (8 concurrent) 6.0 1034.5 1724.1

System Call Overhead 15000.0 4753371.3 3168.9

========

System Benchmarks Index Score 1739.2

Вот еще другая нода (процессор другой), та же машинка:
Только теперь с внешней системы хранения по iSCSI 1Gbit (у прошлой 10 гигабит)

========================================================================

BYTE UNIX Benchmarks (Version 5.1.3)

System: debian: GNU/Linux

OS: GNU/Linux -- 3.2.0-4-amd64 -- #1 SMP Debian 3.2.46-1+deb7u1

Machine: x86_64 (unknown)

Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")

CPU 0: Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz (5200.0 bogomips)

x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET

04:34:23 up 39 min, 2 users, load average: 0.00, 0.89, 1.23; runlevel 2

------------------------------------------------------------------------

Benchmark Run: Tue Dec 03 2013 04:34:23 - 05:02:42

1 CPU in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 36801856.8 lps (10.0 s, 7 samples)

Double-Precision Whetstone 4366.4 MWIPS (10.9 s, 7 samples)

Execl Throughput 2392.8 lps (30.0 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks 1446607.2 KBps (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks 449367.5 KBps (30.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks 2870211.9 KBps (30.0 s, 2 samples)

Pipe Throughput 3159549.4 lps (10.0 s, 7 samples)

Pipe-based Context Switching 308665.7 lps (10.0 s, 7 samples)

Process Creation 15362.1 lps (30.0 s, 2 samples)

Shell Scripts (1 concurrent) 6685.3 lpm (60.0 s, 2 samples)

Shell Scripts (8 concurrent) 899.7 lpm (60.0 s, 2 samples)

System Call Overhead 5455667.8 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX

Dhrystone 2 using register variables 116700.0 36801856.8 3153.5

Double-Precision Whetstone 55.0 4366.4 793.9

Execl Throughput 43.0 2392.8 556.5

File Copy 1024 bufsize 2000 maxblocks 3960.0 1446607.2 3653.0

File Copy 256 bufsize 500 maxblocks 1655.0 449367.5 2715.2

File Copy 4096 bufsize 8000 maxblocks 5800.0 2870211.9 4948.6

Pipe Throughput 12440.0 3159549.4 2539.8

Pipe-based Context Switching 4000.0 308665.7 771.7

Process Creation 126.0 15362.1 1219.2

Shell Scripts (1 concurrent) 42.4 6685.3 1576.7

Shell Scripts (8 concurrent) 6.0 899.7 1499.6

System Call Overhead 15000.0 5455667.8 3637.1

========

System Benchmarks Index Score 1829.4

VPS/VDS на SSD в Германии

Виртуализация KVM

Тариф Start - 3400 MHz / 512 Mb / 5 Gb SSD

BYTE UNIX Benchmarks (Version 5.1.2)

System: static.***.***.**.**.clients.your-server.de: GNU/Linux

OS: GNU/Linux -- 2.6.32-279.el6.x86_64 -- #1 SMP Fri Jun 22 12:19:21 UTC 2012

Machine: x86_64 (x86_64)

Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")

CPU 0: QEMU Virtual CPU version (cpu64-rhel6) (6800.0 bogomips)

x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET

05:11:01 up 21 min, 1 user, load average: 0.07, 0.02, 0.00; runlevel 3

------------------------------------------------------------------------

Benchmark Run: Tue Sep 03 2013 05:11:01 - 05:39:04

1 CPU in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 43992375.8 lps (10.0 s, 7 samples)

Double-Precision Whetstone 4322.4 MWIPS (10.1 s, 7 samples)

Execl Throughput 6537.5 lps (30.0 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks 1434486.6 KBps (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks 397676.3 KBps (30.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks 3085004.8 KBps (30.0 s, 2 samples)

Pipe Throughput 2398207.2 lps (10.0 s, 7 samples)

Pipe-based Context Switching 494601.1 lps (10.0 s, 7 samples)

Process Creation 25353.4 lps (30.0 s, 2 samples)

Shell Scripts (1 concurrent) 8315.6 lpm (60.0 s, 2 samples)

Shell Scripts (8 concurrent) 1117.5 lpm (60.1 s, 2 samples)

System Call Overhead 3629299.6 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX

Dhrystone 2 using register variables 116700.0 43992375.8 3769.7

Double-Precision Whetstone 55.0 4322.4 785.9

Execl Throughput 43.0 6537.5 1520.3

File Copy 1024 bufsize 2000 maxblocks 3960.0 1434486.6 3622.4

File Copy 256 bufsize 500 maxblocks 1655.0 397676.3 2402.9

File Copy 4096 bufsize 8000 maxblocks 5800.0 3085004.8 5319.0

Pipe Throughput 12440.0 2398207.2 1927.8

Pipe-based Context Switching 4000.0 494601.1 1236.5

Process Creation 126.0 25353.4 2012.2

Shell Scripts (1 concurrent) 42.4 8315.6 1961.2

Shell Scripts (8 concurrent) 6.0 1117.5 1862.5

System Call Overhead 15000.0 3629299.6 2419.5

========

System Benchmarks Index Score 2132.4

System Benchmarks Index Score 2351.1

Но UnixBench очень далек от реальных показателей, и ссылаться на его результаты весьма некорректно.

Попробовал дать 4 ядра:

------------------------------------------------------------------------

Benchmark Run: Tue Dec 03 2013 06:54:51 - 07:22:56

4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 149654574.3 lps (10.0 s, 7 samples)

Double-Precision Whetstone 17592.3 MWIPS (9.8 s, 7 samples)

Execl Throughput 17062.5 lps (30.0 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks 771158.5 KBps (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks 225154.5 KBps (30.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks 2021097.4 KBps (30.0 s, 2 samples)

Pipe Throughput 10662355.9 lps (10.0 s, 7 samples)

Pipe-based Context Switching 1237089.7 lps (10.0 s, 7 samples)

Process Creation 50256.4 lps (30.0 s, 2 samples)

Shell Scripts (1 concurrent) 30983.8 lpm (60.0 s, 2 samples)

Shell Scripts (8 concurrent) 3918.6 lpm (60.0 s, 2 samples)

System Call Overhead 5089713.2 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX

Dhrystone 2 using register variables 116700.0 149654574.3 12823.9

Double-Precision Whetstone 55.0 17592.3 3198.6

Execl Throughput 43.0 17062.5 3968.0

File Copy 1024 bufsize 2000 maxblocks 3960.0 771158.5 1947.4

File Copy 256 bufsize 500 maxblocks 1655.0 225154.5 1360.5

File Copy 4096 bufsize 8000 maxblocks 5800.0 2021097.4 3484.7

Pipe Throughput 12440.0 10662355.9 8571.0

Pipe-based Context Switching 4000.0 1237089.7 3092.7

Process Creation 126.0 50256.4 3988.6

Shell Scripts (1 concurrent) 42.4 30983.8 7307.5

Shell Scripts (8 concurrent) 6.0 3918.6 6531.0

System Call Overhead 15000.0 5089713.2 3393.1

========

System Benchmarks Index Score 4145.1

Я так понял, оно практически не зависит от скорости дисковой подсистемы.

System Benchmarks Index Score 100500.5 - или больше написать?

sladkydze, хорошая писькамерка :)

Спокуха, у меня щаз тест 40-ядерной идёт :) Ыыыыы. Облом, не поняла она 40 ядер :( запускаю на 32. Так, 32 тоже не понимает :(

Правда нода нагружена на 60%, боюсь, особо много там не наберется.

AdminVPS, если бы я писал ради пиара, то ни в коем случае не критиковал бы этот тест. Глупо, не?

Ради прикола тоже померили :)

Тестовый FPS-1 на продакшн-ноде:

------------------------------------------------------------------------

Benchmark Run: Tue Dec 03 2013 17:24:53 - 17:54:30

4 CPUs in system; running 4 parallel copies of tests



Dhrystone 2 using register variables       35691561.0 lps   (10.0 s, 7 samples)

Double-Precision Whetstone                     8642.0 MWIPS (9.8 s, 7 samples)

Execl Throughput                               5536.8 lps   (29.6 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks        263319.6 KBps  (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks           75357.8 KBps  (30.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks        889063.4 KBps  (30.0 s, 2 samples)

Pipe Throughput                             1899538.4 lps   (10.0 s, 7 samples)

Pipe-based Context Switching                 301516.9 lps   (10.0 s, 7 samples)

Process Creation                              12042.2 lps   (30.0 s, 2 samples)

Shell Scripts (1 concurrent)                   4991.6 lpm   (60.1 s, 2 samples)

Shell Scripts (8 concurrent)                    931.5 lpm   (60.1 s, 2 samples)

System Call Overhead                        2756816.3 lps   (10.0 s, 7 samples)



System Benchmarks Index Values               BASELINE       RESULT    INDEX

Dhrystone 2 using register variables         116700.0   35691561.0   3058.4

Double-Precision Whetstone                       55.0       8642.0   1571.3

Execl Throughput                                 43.0       5536.8   1287.6

File Copy 1024 bufsize 2000 maxblocks          3960.0     263319.6    664.9

File Copy 256 bufsize 500 maxblocks            1655.0      75357.8    455.3

File Copy 4096 bufsize 8000 maxblocks          5800.0     889063.4   1532.9

Pipe Throughput                               12440.0    1899538.4   1527.0

Pipe-based Context Switching                   4000.0     301516.9    753.8

Process Creation                                126.0      12042.2    955.7

Shell Scripts (1 concurrent)                     42.4       4991.6   1177.3

Shell Scripts (8 concurrent)                      6.0        931.5   1552.5

System Call Overhead                          15000.0    2756816.3   1837.9

                                                                   ========

System Benchmarks Index Score                                        1217.5



root@test:/root/unixbench-5.1.3#

Если делать быструю оценку адекватности теста можно вглянуть на замеры скорости диска, где мелькают даже гигабайты. Дальше можно и не смотреть :)