Benchmark as Seen on WHT?

[therandthem]

Hello,

Can you run this popular benchmark (http://www.webhostingtalk.com/showthread.php?s=&threadid=308055) on your 32MB and 64MB VPS? This has been very helpful in shopping for a VPS.

Thanks,

therandthem

[matta]

This was run on our Fedora Core 3 (minimal) image. Note this was a 32 VM that I upgraded the RAM/CPU on to make a 64 hence why the disk is 3GB in both tests. This test is good to judge performance in regards to CPU and disk, but remember if you overload and your VM starts swapping these numbers will drop substantially. Right now the host looks to be fairly idle, although on another server with more VM's the numbers are still very good.

The 32:
==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux 32mbtest.com 2.6.10-xenU #3 Wed Feb 23 13:37:39 EST 2005 i686 athlon i386 GNU/Linux
/dev/sda1 3084240 944628 1982940 33% /

Start Benchmark Run: Tue Mar 1 15:02:50 EST 2005
15:02:50 up 5 min, 1 user, load average: 0.10, 0.06, 0.01

End Benchmark Run: Tue Mar 1 15:14:19 EST 2005
15:14:19 up 16 min, 1 user, load average: 14.57, 6.76, 3.17


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 3950918.9 104.9
Double-Precision Whetstone 83.1 1079.6 129.9
Execl Throughput 188.3 2135.1 113.4
File Copy 1024 bufsize 2000 maxblocks 2672.0 14659.0 54.9
File Copy 256 bufsize 500 maxblocks 1077.0 31293.0 290.6
File Read 4096 bufsize 8000 maxblocks 15382.0 13838.0 9.0
Pipe Throughput 111814.6 1017798.0 91.0
Pipe-based Context Switching 15448.6 216380.9 140.1
Process Creation 569.3 4427.8 77.8
Shell Scripts (8 concurrent) 44.8 466.8 104.2
System Call Overhead 114433.5 903423.2 78.9
=========
FINAL SCORE 85.6


The 64:
==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux 64mbtest.com 2.6.10-xenU #3 Wed Feb 23 13:37:39 EST 2005 i686 athlon i386 GNU/Linux
/dev/sda1 3084240 944684 1982884 33% /

Start Benchmark Run: Tue Mar 1 15:17:34 EST 2005
15:17:34 up 1 min, 1 user, load average: 0.12, 0.05, 0.01

End Benchmark Run: Tue Mar 1 15:29:03 EST 2005
15:29:03 up 13 min, 1 user, load average: 17.85, 7.25, 3.21


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 3974371.5 105.5
Double-Precision Whetstone 83.1 1079.7 129.9
Execl Throughput 188.3 2135.9 113.4
File Copy 1024 bufsize 2000 maxblocks 2672.0 70830.0 265.1
File Copy 256 bufsize 500 maxblocks 1077.0 32210.0 299.1
File Read 4096 bufsize 8000 maxblocks 15382.0 264287.0 171.8
Pipe Throughput 111814.6 1026877.8 91.8
Pipe-based Context Switching 15448.6 219003.4 141.8
Process Creation 569.3 4435.5 77.9
Shell Scripts (8 concurrent) 44.8 466.9 104.2
System Call Overhead 114433.5 904642.2 79.1
=========
FINAL SCORE 129.9

The 128 (This VM is on a much more loaded host that hosts over 3 times as many VM's. It is also only Opteron 240's compared to the 242's in the newer production servers):
==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht)
System -- Linux server.xxxxxxxx.com 2.6.10-xenU #3 Wed Feb 23 13:37:39 EST 2005 i686 athlon i386 GNU/Linux
/dev/sda1 6047492 2754008 2986284 48% /

Start Benchmark Run: Tue Mar 1 15:38:06 EST 2005
15:38:06 up 1 day, 13:34, 1 user, load average: 0.32, 0.14, 0.05

End Benchmark Run: Tue Mar 1 15:49:41 EST 2005
15:49:41 up 1 day, 13:45, 1 user, load average: 16.92, 7.39, 3.33


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 3486185.8 92.5
Double-Precision Whetstone 83.1 718.1 86.4
Execl Throughput 188.3 1989.1 105.6
File Copy 1024 bufsize 2000 maxblocks 2672.0 58591.0 219.3
File Copy 256 bufsize 500 maxblocks 1077.0 24529.0 227.8
File Read 4096 bufsize 8000 maxblocks 15382.0 225276.0 146.5
Pipe Throughput 111814.6 754473.3 67.5
Pipe-based Context Switching 15448.6 113450.0 73.4
Process Creation 569.3 3417.3 60.0
Shell Scripts (8 concurrent) 44.8 411.4 91.8
System Call Overhead 114433.5 750759.2 65.6
=========
FINAL SCORE 100.9


So as you can see, it does hold it's performance pretty well.

[therandthem]

Excellent. You did it. unixshell# is probably the first VPS offering that competes with dedicated servers . Obviously, the prices for memory are higher with unixshell# than a bottom-of-the-line dedicated server. That being said, however, after a few minutes in my VPS control panel I feel like I am virtually in the server room. I could not feel more confident in a VPS provider than I do right now.

Kindest regards,

therandthem

[matta]

Thanks! It's appreciated.