VIA KT400 Chipset Overclocking
 

Nice work !

Should rename it:
"VIA KT400 Chipset si teh suXor ! " :D

What about hyperion (http://www.viaarena.com/?PageID=211) ?
What's the influence on the benchies ?

would be cool if:
a) I had the motherboards & other hardware still in my possesion
b) the R9700Pro worked on the KT400 :)

:-)

why didN'T you put a nforceII in the review?

so that we good see how mutch the kt400 sucks ;)

silencer

Hmmm... not a bad idea silencer.

JMKe, shall I make a quick run of DCDDR running at 133/166/200FSB? It would make a nice comparison. :)

Awesome work by the way. How did you make those 1337 graphs? :ws:

you're kidding right.. about the graphs? :p
Excell man.. please.. go out and buy it, there's a students pack out that includes alot of otherwise very $$$ stuff, for only 100€ or so :)

yes please & add them to the mentioned thread :)

anywayz to answer to Silencer's question: I didn't add the nForce2 cause at the time of testing the nForce2 wasn't out yet :)

but Comtech has shown interest in this, so I might be reviewing 2-3 nForce2 mobo's very soon
Epox & Asus are pretty sure, maybe the Abit will follow also :)

thx for the answer and correction of my English.

btw nice review

silencer

Here are some epox 8rda+ benches at 7-2-2 CL2:

@133fsb 2075/1997
@166fsb 2582/2468
@200fsb 3104/2920


I don't think sisoft is a good way to bench an nforce mobo though. It cearly states to use 64bit data packets to bench, which is the bandwith of single channel ddr... i wouldn't be suprised if i would get exactly the same results with just 1 stick.


And yes i'll stop being a cheap arse and start spending money on software. :^)

those numbers indeed suck compared to the KR7A @ 200 :)

I know sisoft is not the way to bench/compare motherboards
but reallife gaming software benches weren't possible due to the fact that the R9700Pro failed to work

otherwise I had included:
PCMark2002
3DMark2K1
QuakeIII
Serious Sam
UT2003

I had these done on the KT3, as this was the first motherboard I installed, but.. well.. you know.. a TNT2 doesn't really make it an even challenge :)

They don't 'suck', but it just doesn't reflect the performance gain you get from dcddr in real life apps.


By the way, this epox nforce² at 200FSB is still faster than my ol' KR7A at 205FSB. 3104/2920 <> 3055/2730