Server benchmark: Hetzner VQ 12 VPS

  • 2012
  • Hetzner VQ 12 VPS
  • Single Core
  • 1 GB RAM
  • 40 GB HD on some sort of RAID
  • no backup space
  • 2 TB traffic/mo
  • Debian 6.0 64 bit
  • Tests run 2012-01-31
UnixBench 5.1.2
./Run dhry2reg whetstone-double syscall pipe context1 spawn execl shell1 shell8 shell16

========================================================================
   BYTE UNIX Benchmarks (Version 5.1.2)

   System: ********: GNU/Linux
   OS: GNU/Linux -- 2.6.32-5-amd64 -- #1 SMP Mon Jan 16 16:22:28 UTC 2012
   Machine: x86_64 (unknown)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: QEMU Virtual CPU version 0.12.3 (6148.8 bogomips)
          x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   23:33:01 up 12 min,  1 user,  load average: 0.10, 0.03, 0.01; runlevel 2

------------------------------------------------------------------------
Benchmark Run: Mon Jan 30 2012 23:33:01 - 23:57:04
1 CPU in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       29946843.0 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     3894.4 MWIPS (10.0 s, 7 samples)
Execl Throughput                               4957.1 lps   (30.0 s, 2 samples)
Pipe Throughput                             2817011.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 465003.1 lps   (10.0 s, 7 samples)
Process Creation                              17985.7 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   8123.4 lpm   (60.0 s, 2 samples)
Shell Scripts (16 concurrent)                   528.5 lpm   (60.1 s, 2 samples)
Shell Scripts (8 concurrent)                    967.9 lpm   (60.0 s, 2 samples)
System Call Overhead                        4939570.4 lps   (10.0 s, 7 samples)

System Benchmarks Partial Index              BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   29946843.0   2566.1
Double-Precision Whetstone                       55.0       3894.4    708.1
Execl Throughput                                 43.0       4957.1   1152.8
Pipe Throughput                               12440.0    2817011.4   2264.5
Pipe-based Context Switching                   4000.0     465003.1   1162.5
Process Creation                                126.0      17985.7   1427.4
Shell Scripts (1 concurrent)                     42.4       8123.4   1915.9
Shell Scripts (16 concurrent)                     ---        528.5      ---
Shell Scripts (8 concurrent)                      6.0        967.9   1613.2
System Call Overhead                          15000.0    4939570.4   3293.0
                                                                   ========
System Benchmarks Index Score (Partial Only)                         1627.5

the best of three subsequent test runs of

dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
1073741824 bytes (1.1 GB) copied, 8.39155 s, 128 MB/s

This must be a some sort of peak , the other two were around 70 MB/s.

the best of three subsequent test runs of

hdparm -t dev
Timing buffered disk reads: 858 MB in  3.00 seconds = 285.75 MB/sec

Disclaimer: based on my experience disk performance testing with dd and hdparm are crude methods. The real-life disk load of a web service cannot be emulated with dd or hdparm.