SliceHost
==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht.1)
System -- Linux sparta.dillio.net 2.6.18-xen #1 SMP Tue Feb 12 06:40:50 UTC 2008 x86_64 x86_64 x86_64 GNU/Linux
/dev/sda1 9805144 3825344 5481728 42% /

Start Benchmark Run: Tue Jun 17 16:00:53 EDT 2008
16:00:53 up 36 days, 23:01, 1 user, load average: 0.00, 0.00, 0.00

End Benchmark Run: Tue Jun 17 16:11:41 EDT 2008
16:11:42 up 36 days, 23:12, 2 users, load average: 14.76, 6.36, 2.82


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 9345690.3 248.0
Double-Precision Whetstone 83.1 2060.5 248.0
Execl Throughput 188.3 1385.8 73.6
File Copy 1024 bufsize 2000 maxblocks 2672.0 47061.0 176.1
File Copy 256 bufsize 500 maxblocks 1077.0 15932.0 147.9
File Read 4096 bufsize 8000 maxblocks 15382.0 390576.0 253.9
Pipe Throughput 111814.6 499706.2 44.7
Pipe-based Context Switching 15448.6 118041.6 76.4
Process Creation 569.3 2205.0 38.7
Shell Scripts (8 concurrent) 44.8 410.5 91.6
System Call Overhead 114433.5 421841.5 36.9
=========
FINAL SCORE 102.7



Linode
==============================================================
BYTE UNIX Benchmarks (Version 4.1-wht.1)
System -- Linux rome.dillio.net 2.6.18.8-domU-linode7 #1 SMP Mon Feb 11 13:45:54 EST 2008 i686 i686 i386 GNU/Linux
/dev/xvda 11936416 1897552 10038864 16% /

Start Benchmark Run: Tue Jun 17 16:00:30 EDT 2008
16:00:30 up 12:46, 1 user, load average: 0.00, 0.00, 0.00

End Benchmark Run: Tue Jun 17 16:10:43 EDT 2008
16:10:43 up 12:56, 1 user, load average: 14.29, 6.09, 2.66


INDEX VALUES
TEST BASELINE RESULT INDEX

Dhrystone 2 using register variables 376783.7 38725011.7 1027.8
Double-Precision Whetstone 83.1 1992.8 239.8
Execl Throughput 188.3 3079.5 163.5
File Copy 1024 bufsize 2000 maxblocks 2672.0 132814.0 497.1
File Copy 256 bufsize 500 maxblocks 1077.0 37258.0 345.9
File Read 4096 bufsize 8000 maxblocks 15382.0 1219939.0 793.1
Pipe-based Context Switching 15448.6 361016.5 233.7
Pipe Throughput 111814.6 2418718.5 216.3
Process Creation 569.3 4223.0 74.2
Shell Scripts (8 concurrent) 44.8 965.7 215.6
System Call Overhead 114433.5 1959239.5 171.2
=========
FINAL SCORE 277.8

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

Either the host node are setup differently to manage performance or Linode is just using better hardware hands down....This can also play out opteron (slicehost) vs xeon (linode). Just to be fair the linode vps gets a little more ram as well.

If you look a lot of this win was in the processor and you can see why:

SliceHost
[root@sparta unixbench-4.1.0-wht-1]# cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2212
stepping : 2
cpu MHz : 2010.300
cache size : 1024 KB
physical id : 0
siblings : 1
core id : 1
cpu cores : 1
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu tsc msr pae mce cx8 apic mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8_legacy
bogomips : 4031.05
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc

processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2212
stepping : 2
cpu MHz : 2010.300
cache size : 1024 KB
physical id : 0
siblings : 1
core id : 1
cpu cores : 1
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu tsc msr pae mce cx8 apic mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow up pni cx16 lahf_lm cmp_legacy svm cr8_legacy
bogomips : 4031.05
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc

processor : 2
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2212
stepping : 2
cpu MHz : 2010.300
cache size : 1024 KB
physical id : 0
siblings : 1
core id : 1
cpu cores : 1
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu tsc msr pae mce cx8 apic mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow up pni cx16 lahf_lm cmp_legacy svm cr8_legacy
bogomips : 4031.05
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc

processor : 3
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2212
stepping : 2
cpu MHz : 2010.300
cache size : 1024 KB
physical id : 0
siblings : 1
core id : 1
cpu cores : 1
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu tsc msr pae mce cx8 apic mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow up pni cx16 lahf_lm cmp_legacy svm cr8_legacy
bogomips : 4031.05
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc








Linode
[root@rome unixbench-4.1.0-wht-1]# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU L5420 @ 2.50GHz
stepping : 6
cpu MHz : 2500.088
cache size : 6144 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu de tsc msr pae cx8 apic cmov pat clflush acpi mmx fxsr sse sse2 ss ht nx constant_tsc pni
bogomips : 5007.44

processor : 1
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU L5420 @ 2.50GHz
stepping : 6
cpu MHz : 2500.088
cache size : 6144 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu de tsc msr pae cx8 apic cmov pat clflush acpi mmx fxsr sse sse2 ss ht nx constant_tsc pni
bogomips : 5007.44

processor : 2
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU L5420 @ 2.50GHz
stepping : 6
cpu MHz : 2500.088
cache size : 6144 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu de tsc msr pae cx8 apic cmov pat clflush acpi mmx fxsr sse sse2 ss ht nx constant_tsc pni
bogomips : 5007.44

processor : 3
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU L5420 @ 2.50GHz
stepping : 6
cpu MHz : 2500.088
cache size : 6144 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu de tsc msr pae cx8 apic cmov pat clflush acpi mmx fxsr sse sse2 ss ht nx constant_tsc pni
bogomips : 5007.44

You will see quite a bit of variability around cpu on those types of tests, since other slices can drastically impact the results. If there are no slices doing much of anything you could end up with a test run that uses close to 400% of 400% available processing power. On the server your test slice is on Ross, it appears that at least 2 slices are very active with cpu which would cut down the ceiling of burstability by a factor of 3 (given that all three slices in question are 256, you could expect to be able to burst to approximately 133%).

These types of tests are difficult to use as a basis of comparison with VPS’s that can burst. For example, you may have been the only customer on the Linode server, which means you effectively had use of the entire physical server. If you repeat that test in a few months time, you could see dramatically different results.

I have heard good things about Linode. They have been around longer than Slicehost, but only recently began using Xen in production – so I feel that Slicehost is the more experienced Xen host of the two. Looking over Linode’s new Xen-based offerings, I don’t see anything that would compel me to switch.

Numbers don't really mean much to me. As long as my slice performs at a level I am comfortable with, I don't care what the numbers look like. If my slice starts getting more traffic and I notice a degradation in performance, it's either time to upgrade to a larger slice or dedicated server. Slice has been great for my needs and by far out performs my old shared host. I'm not trying to pump up Slice or put down Linode, just tossing in my two cents.

I would agree that a benchmark like this does not give any valuable information. A lot of times the first limiting factor on a server will be the available ram. Slicehost also happens to offer larger ram options than linode.

It is good to see more competition. It gives me more options to diversify. What happened at ThePlanet could happend to anyone. I personally really like the cheaper bandwidth rates that linode is offering and may have to use them for some transfer intensive projects in the future.

another thing.
linode gives out a lot of free upgrade, seem at least once a year you get 10% more ram or 20% more space or 30% more bandwidth etc... etc...
http://blog.linode.com/category/upgrades/

I have used both providers and one thing I don't like about linode, it feels like roulet when picking your datacenter. The Atlanta sites seems to be have more than its fair share of network issues, though they claim to be fixing this sites issue. I know some of these are circumstances that may be out of there control. I also am not a big fan of how they load all modules into the kernel. You can still load modules, but nearly every base modules is loaded, whether you need it or not. But the notable plus is they do not have a hardcap on max bandwidth on the nics. The management interface is also top notch as well. But with that said slicehost sells a great vps and easily meets or exceeds linode in most areas, in my opinion.

Justin

I am loyal to Slicehost purely for the amazing second to none support and service. I remember sending a support ticket for the most stupid thing and Jason treated me seriously. :)

Hey – the hardware is good to too.

I've been wondering why my slice seems to be out performed by my media temple vps with similar specs and a plesk hog . Is slicehost planning to upgrade to better processors anytime in the near future?