VIA KT400A Review

pnabais

pnabais

Power Member
Última edição:
acho que tá porreirinha! :) falta a mãozinha dos fabricantes pa afinar akilo e dps ver o que é que os OC fazem dakilo :cool:
 
A questão será se isso é capaz de manter os bus PCI/AGP locked independente do FSB, tal como a nforce2 faz.

Se não nada feito.
 
Um rumor interessante sobre o KT400A por aqui ...
http://www.sudhian.com/news/index.cfm#3


VIA KT400A Launching With Dubious 400 MHz FSB Support 03/10

Ever since rumors of the KT400A began to surface it was billed as a potential "nForce2 killer" (wasn't that long ago we talked about "KT266A killers--but times change), with a much-improved memory controller, integrated serial ATA, and a level of performance that would put VIA firmly back in pole position ahead of its largest Socket A rival. Initially it was rumored that KT400A would be a dual DDR solution, but even after VIA announced it would offer only a single DDR channel there was reason to be optimistic, as VIA again claimed its incredibly streamlined 64-bit DDR channel could still outperform a dual DDR nForce2 solution. Although the market has grown somewhat caustic over VIA's continual habit of "Here's a Chipset / Here's the one we should've made--we'll call it 'A'" pattern of release, KT400A looked to be a genuine leap in features / performance over its predecessor.

It appears, however, that in their rush to bring KT400A to market VIA has taken a bit of a shortcut. Initial KT400A boards (including those sampling now) are shipping with Revision CD of the chipset, which does not offer optimized support for a 400 MHz front-side bus, or a full set of memory dividers to run at these speeds. Furthermore, the capabilities of these boards to even reach a 400 MHz FSB is somewhat in doubt as they will be running in an unofficial and sub-optimal configuration in order to do so.

The performance impact of an improperly tuned chipset at a given FSB can be significant, and is one reason why the nForce2 swept the original KT400 by as much as 10% in certain FSB / bandwidth oriented tests. While KT400A looks to rectify that problem at a 333 MHz FSB, boards based on the CD version of the chipset will promptly re-introduce it once upgraded to 400 MHz FSB AthlonXP's--assuming, of course, that CD-based boards can reach that speed to begin with. DDR400 memory compatibility could also be a factor, particularly if CD-based boards are sub-optimally timed to begin with. This is precisely why Intel was leery of a DDR400 memory standard to begin with, and why, even after they accepted one, they've spent a great deal of time carefully tuning performance on their Canterwood and Springdale motherboards.

VIA plans to release a new chipset stepping (KT400A_CE) in about six weeks, with the new drop-in design first appearing in boards around late May / early June. CE will introduce a number of improvements, including a full set of memory dividers (allowing for asynchronous memory operation (either lower or higher speed), be optimized for running at a 400 MHz bus, and will likely offer improved DDR400 compatibility as well. Even once CE-based boards are available it'll be difficult to tell which boards use which chipset revision--unless, of course, motherboard manufacturers choose to re-issue new versions of boards barely six weeks old.

When KT400A launches it's possible it will offer the fastest, most stable, and best-priced Socket A chipset available--on a 333 MHz FSB. At this point, however, anything above that could be a crap shoot depending on your motherboard, your manufacturer, and sheer luck. The upcoming launch of the board should tell us more, but for the moment those of you interested in upgrading to a 400 MHz FSB AthlonXP just might want to hold out until summer to see what the verdict is.


 
Inicie sessão ou registe-se para poder participar.
Back
Topo