Are these benchmarks credible ?

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
A couple of months ago, a computer magazine published their own
comparative benchmarks of several budget motherboards.

I understand that the chipset and its implementation by the mobo
manufacturer make a difference, but I was surprised by the very large
differences in benchmark figures. Here are a few sample figures for AMD
mobos (RAM, HDD, PSU, etc are the same for all tests) :

Quake 3 fps :

For S754 3000+, GF6200 AGP -
Asus K8N (NF3) - 44 fps
Gigabyte K8N-E (NF4) - 157.4 fps
Winfast NF4K8AB (NF4) - 170.4 fps

For S939 FX-53, GF6200 PCI-e -
Elitegroup RS480M (RS480) - 66.9 fps
Asrock 939A8X-M (ULi1689) - 132.4 fps
Winfast NF4K8MC (NF4) - 185.6 fps

There were similar differences for other games. There were also major
differences in video encoding times, though less than those for game
fps figures.

Are these figures credible ?

Re: Are these benchmarks credible ? wrote:
Quoted text here. Click to load it

Correction : These figures are for Serious Sam Second Encounter. The
figures for Q3 show similar ratios. Sorry.

Re: Are these benchmarks credible ?

On 7 Oct 2005 13:19:29 -0700, wrote:

Quoted text here. Click to load it

No they didn't have some of them set up correctly.  Using
same CPU and (close enough) video card will not result in
more than a couple dozen % difference and usually
significantly less than that.  More than anything it might
indicate the suitablity of the particular board (and it's
bios version) default settings and how those setting effect
the performance.  Then again, more aggressive settings may
be instable with some hardware, like memory, so one cannot
automatically assume any give score would be what they will
get if they only buy same motherboard.

Re: Are these benchmarks credible ?

Quoted text here. Click to load it

Do you have a link to these tests??
I'd like to see them try and explain the 100+ fps difference shown in the
Socket 754 test.

"I don't cheat to survive. I cheat to LIVE!!"
 - Alceryes

Re: Are these benchmarks credible ?

Alceryes wrote:
Quoted text here. Click to load it

Afraid not. Their website doesn't have an online copy of their mag's
contents. And since it's an Indian publication, I don't think you'll
have access to it even if you were willing to buy a copy just to check.

Quoted text here. Click to load it

See the same scale of difference for S939 too. They don't really try to
explain anythng in the text. The benchmarks are presented as a
double-page chart and some selected benchmarks as bar graphs. The text
simply makes comments like this mobo is the winner and that one did
poorly and so on, quoting the same figures here and there.

There were other tests and many other mobos involved. The ones I quoted
are just a few samples.

One thing that's clear is that they are not typos, as the same figures
are repeated in the text. The mag is called 'Digit' and their website
is The issue in question is the July issue.

So the general idea is that these tests were not competently run ??

Site Timeline