Info on KT400A with benchmark tests (faster then nforce2 chipset)

saw this on x labs link

Supports all Socket A Athlon XP/Athlon/Duron processors with 200, 266 and 333MHz EV-6 system bus;
VIA KT400A North Bridge can be accompanied by VT8235 or VT8237 I/O Controller;
8x V-Link architecture to connect North and South Bridges;
Up to 4GB of PC2100, PC2700 or PC3200 (DDR400) DDR SDRAM;
AGP 8x slot;



VIA KT400 is fantastically 0.65% faster than single-channel NVIDIA nForce2 and amazingly 0.43% quicker than dual-channel nForce2 in Winstone Content Creation test.
VIA KT400 outperforms single-channel nForce2 by outrageous 2.93% and is 2.05% faster compared to dual-channel nForce2 in 3DMark 2001SE test.
VIA KT400 is can boast with dazzling 0.82% of extra performance versus single-channel nForce2 and surpasses dual-channel nForce2 by unbelievable 0.25% in Quake III Arena (640x480) test.
Somehow VIA KT400 is 2.20% faster than single-channel nForce2 and outdistances dual-channel NVIDIA nForce2 by formidable 0.48% in Quake III Arena (640x480) test.
VIA KT400 outperforms single-channel nForce2 by marvelous 5.78% and dual-channel nForce2 by incredible 0.68% in Jedi Knight II test.
Due to unknown reason VIA KT400 managed to outdo dual-channel nForce2 in CodeCreatures BenchmarkPro by shiny and astonishing 0.1fps in 1600x1200 resolution. I wonder how did it happen, given that CodeCreatures is a test fully for graphics cards?

i'll believe it when i see it

via also claimed they had a 1/6 divisor in kt400 but that never happened

a PCI lock is 100times better than a 1/6 divisor anyway cause then you never need to worry about currupting your hardrive

Emericana said:

a PCI lock is 100times better than a 1/6 divisor anyway cause then you never need to worry about currupting your hardrive

Click to expand...

i disagree. a PCI lock is infinitely better than a divisor. even if it's a 1/8 divisor locking you're PCI at a stable speed would be better than having it's speed contingent on the FSB.

Where the hell did u get all of those "stupendous" adjectives?

FSB running synchronously (divisors) to the PCI/AGP/HDD speed gives more performance then the pci lock.

i'm just passing on info, no need to bash me

Samoyed said:

Where the hell did u get all of those "stupendous" adjectives?

Click to expand...

http://www.dictionary.com

Hc000 said:

FSB running synchronously (divisors) to the PCI/AGP/HDD speed gives more performance then the pci lock.

i'm just passing on info, no need to bash me

Click to expand...

yeah because the higher you OC the FSB the higher your PCI bus goes. so your overall system speed would be greater but if something on the PCI bus is unstable at higher speeds than it wil hinder your overclock and actaully be slower. in most cases ovreclocking the FSB will be beneficial because the people that do this are using memory intensive applications (games). so locking out the PCI to keep the IDE controllers stable and than upping the memory frequency would be much better.

Of course you have to consider all the crazy mods some people have to do to get their pci-lock mobo to high fsbs. Id be willing to bet that with a 1/6 divisor a via chipset would do around the same max fsb that a good nf2 board would.

Thats just my opinion. No i didnt just offend all you nf2 guys and start some flame war. Just a thought.

if u keep ur fsb @

100 with 1/3 dividor
133 with 1/4 dividor
166 with 1/5 dividor
200 with 1/6 dividor
233 with 1/7 dividor
266 with 1/8 dividor

assuming there is dividers that goes up that high) then everything is fine

Hey NForce2 users - if you look at the system setup that VIA used to prove the KT400A is faster, you'll notice that the NF2 board has the memory set async with the FSB ( 333 FSB and 400DDR) - has anyone tested this configuration to see if you incur a latency penalty when running an NF2 board async ( 333 FSB and 400DDR) as opposed to synchronously ( 333FSB and 333DDR)?

I think both boards are equally fast within a margin of error, but I'm wondering if VIA is trying to make their board look faster by saying it can beat an NF2 board with faster memory.

I hate using dividers, at least with a pci-lock u can be sure anything using the pci bus like harddrives are not causing instabilty when overclocking. Same goes for the AGP bus lock.
I got screwed when my old KT333 mobo only had a 1/4 divider so it couldnt run 166 fsb cpus. Having dividers limits your upgrades down the road. The KT333 mobo didnt even last for a year before i had to change it.

ya i pretty much agree with everyone else about the PCI lock vs divisor... lock is way better... all Nvidia needs to do now is put out better quality NBs so we can get higher... i really think it would be cool to have a mobo iwth a replaceable NB that you just plop in like a CPU... then you wouldnt need to buy a new everything when you upgrade just a new NB and you are set as long asno big new features came out like USB3.0 or some newer storage type... that would be sweet

Hc000 said:

VIA KT400 is fantastically 0.65% faster than single-channel NVIDIA nForce2 and amazingly 0.43% quicker than dual-channel nForce2 in Winstone Content Creation test.
VIA KT400 outperforms single-channel nForce2 by outrageous 2.93% and is 2.05% faster compared to dual-channel nForce2 in 3DMark 2001SE test.
VIA KT400 is can boast with dazzling 0.82% of extra performance versus single-channel nForce2 and surpasses dual-channel nForce2 by unbelievable 0.25% in Quake III Arena (640x480) test.
Somehow VIA KT400 is 2.20% faster than single-channel nForce2 and outdistances dual-channel NVIDIA nForce2 by formidable 0.48% in Quake III Arena (640x480) test.
VIA KT400 outperforms single-channel nForce2 by marvelous 5.78% and dual-channel nForce2 by incredible 0.68% in Jedi Knight II test.
Due to unknown reason VIA KT400 managed to outdo dual-channel nForce2 in CodeCreatures BenchmarkPro by shiny and astonishing 0.1fps in 1600x1200 resolution. I wonder how did it happen, given that CodeCreatures is a test fully for graphics cards?

Click to expand...

Er I think all those superlatives were given tongue in cheek!
Fantastically 0.65% faster indeed

With how much longer the damn KT400a took to come out, I would EXPECT it to be faster! If it wasn't, it would be pretty pathetic! But still, 0.65% faster? What the hell is that? 10MB/s faster bandwidth? Oh yeah, really worth the extra $50-$100 to buy....NOT!!! I'll keep my NForce2 plan for now....

I'd like to have both options, dividers selectable, and a PCI lock if I want to use that.

ya that would be the best but on the amd scene it aint gonna happen because of the way the chipsets are made

Emericana said:

via also claimed they had a 1/6 divisor in kt400 but that never happened

a PCI lock is 100times better than a 1/6 divisor anyway cause then you never need to worry about currupting your hardrive

Click to expand...

VIA never claimed to have a 1/6 divisor. Their earliest roadmaps always showed it to have only a 1/5 divisor. The issue you are confused with is the fact that VIA left the decision with the mobo manufacturers to include the 1/6 divisor if they felt their boards had a clean enough signal to run it.

PCI locks are ok if you like increasing the bandwidth to the CPU only. Others like complete system overclocks & for those divisors are the way to go.

Will VIA release the KT400A even when they just released the KT400 & tried to kill off the KT333 in the process? Well if they do I'm definitely going to buy one from ABIT.

why would you want both, why not just fully user configurable PCI bus with a default of locked at 33?

dividers are worthless, why would you want to deal with VIA and their slow divider addition process? its going to be really pathetic if VIA releases the kt400a with only the 1/6. how hard can it be to include high dividers?

Demont said:

its going to be really pathetic if VIA releases the kt400a with only the 1/6. how hard can it be to include high dividers?

Click to expand...

lets see you make a chipset for athlon that include all the dividers!

Sonny said:

VIA never claimed to have a 1/6 divisor. Their earliest roadmaps always showed it to have only a 1/5 divisor. The issue you are confused with is the fact that VIA left the decision with the mobo manufacturers to include the 1/6 divisor if they felt their boards had a clean enough signal to run it.

Click to expand...

Im still convinced because of via saying that, theres still the probability of bios updates to include the 1/6 divisor to the kt400 boards. Especially due to those 200mhzfsb bartons coming up.

inspectorhammer said:

Im still convinced because of via saying that, theres still the probability of bios updates to include the 1/6 divisor to the kt400 boards. Especially due to those 200mhzfsb bartons coming up.

Click to expand...

you may be right!