[citation][nom]Wittermark[/nom]for those people making false claims on which browser is faster, i think this graph can put all the BS to rest:clearly firefox 4 is nowhere close to chrome, not even version 5.[/citation]
You must not understand benchmarking at all.
One benchmark can not definitively show which is the fastest.
For one, workloads can vary.
Secondly, hardware can vary, and even that could have an impact. I could run something on an Athlon and get different results from a Phenom, depending upon how it uses memory and how much the L3 cache helps. Or, Intel chips, which have a different architecture. Or, different memory amounts, which could involve caching. Or different disks - one may benefit from SSD more than the other, or a specific way a hard disk caches, etc... Or some might run better on a different OS than this was tested on. Even if it's 1% or 2% here and there, and it can be quite a bit higher, there are so many variables results can vary based on changing parameters.
There are a lot of things that can change a result. Benchmarking is extremely difficult because of that, and weird things like I mention above do have an impact.
So, sorry, your one benchmark doesn't eliminate any discussion on the subject. No single benchmark ever did. Especially for software that runs on software.
You must not understand benchmarking at all.
One benchmark can not definitively show which is the fastest.
For one, workloads can vary.
Secondly, hardware can vary, and even that could have an impact. I could run something on an Athlon and get different results from a Phenom, depending upon how it uses memory and how much the L3 cache helps. Or, Intel chips, which have a different architecture. Or, different memory amounts, which could involve caching. Or different disks - one may benefit from SSD more than the other, or a specific way a hard disk caches, etc... Or some might run better on a different OS than this was tested on. Even if it's 1% or 2% here and there, and it can be quite a bit higher, there are so many variables results can vary based on changing parameters.
There are a lot of things that can change a result. Benchmarking is extremely difficult because of that, and weird things like I mention above do have an impact.
So, sorry, your one benchmark doesn't eliminate any discussion on the subject. No single benchmark ever did. Especially for software that runs on software.