|Freshly in: first C7-M "review".
||[Dec. 16th, 2005|08:11 pm]
|||||Eels - Electro-shock blues - Last stop, this town||]|
Talk about timing. Thank you VIA and friends, for tackling me as effectively. I never knew i actually had VIA folk as readers, i suddenly feel all important and suchlike.
So, what has happened. VIA handed some kit to the usual suspect again, for a very early review of hardware that will not be reviewed anywhere else for several months at least. In the short lifespan of this blog, two such reviews have already taken place, and only the second one has had other reviews since. And these are reviews by hardware review sites, as in, not by people who have gone out and bought one in a shop or on ebay.
Does this mean that the C7 is not vapourware? Most certainly not. C7s will not actually appear before Q2 2006. It will probably take until Q3 or Q4 before anyone from .eu or .us will be able to actually buy one. So, with a bit of luck, VIA might actually be delivering within 3 years of the initial announcement. Congratulations.
So what can we learn about the C7-M from this review?
More northbridge naming idiocy. Hurrah!
Yes, the never ending story continues (and here), even though it seemed that there was some closure to the last mess. New additions are: Vx700, Vx700M, VN890 and VN900. It really is "another marketing release, another name" with VIA, isn't it.
VN890 seems to be a VN800, which was only announced 3.5months ago, but this time with PCI-E. It's amazing how VIA apparently left out PCI-E in the first place, but then, i still have no certainty as to what unichrome version is included on this.
The VN900 is supposedly a Deltachrome IGP like the K8M890, which, even though it was announced months ago, hasn't materialised either. And no, VIA doesn't provide source (i'd love the ability to merge the XXXchrome modesetting with the unichrome driver btw - silicon image TMDS encoder, wee.), so free software users, steer well clear of these devices when they do appear.
And which unichrome the Vx700 will be i don't know either, chances are that this is VT3157 (as supposedly found on the CN700). Plain Vx700 wasn't good enough, so there also needed to be a miniaturised version with the Vx700m.
What a mess. Well, it would be understandable if most of these devices actually made it to market at some point. But VIA seems to just rename devices more or less at random for the ability to spew out more empty marketing, adding and removing some features to justify themselves.
Another conclusion from the C7-M laptop review is the following: You might remember the pointe of my last blog entry; VIA announced the C7 way before they even had a capable northbridge for it. These were of course assumptions, as with most things coming out of VIA, there is no certainty, but you try to apply logic to it and try to make some sense out of it.
The first Northbridge that was announced to be supporting the C7-M was the VN800, its marketing states the following: "By harnessing VIA's unique Flexi-Bus Technology, the VN800 enables support for both the all new VIA C7-M™ and Intel® Pentium® M processors."
The announcement of the VN800, which still has to be spotted in this dimension, was followed shortly by the announcement of the P4M800Pro. I then assumed that it was the intel only twin of the VN800, as its marketing states: "Processor Support: Intel® Pentium® 4".
Then CN700 came around, and marketing states the following "The first chipset to support the new VIA C7™ processor family".
So... The VN800 plainly implements the pentium-M bus. This explains why the PN800 was named in the early C7-M laptop marketing. Flexi-bus is just VIA marketing at work again and I wonder if they actually bothered with trademarking that. I'm sure that VIA has very close relations with a .tw trademarking office, what with VIAs naming strategy.
And... The CN700 apparently supports the C7 and only the C7, while the C7-M only implements the Pentium-M bus.
I also dug up a storyline on Xbit-labs; a few weeks before the announcement, a russian VIA representative told theinquirer that VIA licensed the Pentium M bus also for cpus, but that story was retracted the very next day. Pretty curious when current marketing evidence seems to point out that the inquirer was informed correctly the first time round.
So what does all this mean?
First of all, VIA announced the C7 on may the 27th, and the C7-M on june the first. Amazingly, the original C7 pressrelease in VIAs archive now surreptitiously points to the C7-M pressrelease, something which i at first discounted to be an archiving accident (giving VIA the benefit of the doubt). What the above means is that VIA announced the CPU six and a half months before a suitable northbridge was announced, and then tried to cover up the mess a few days later.
Secondly, why is there this difference in bus? Could the reason be backward compatibility with the intel bus for both CPU and northbridge? Surely not. Surely the CN700, as it succeeds the CN400 and CLE266, is aimed at the embedded market, where power is at a premium. Surely a device like the CN700 should be the perfect candidate for a laptop when a VIA chip is chosen. So the use of the Pentium-M bus for the C7-M is not borne out of backward compatibility but out of need.
A very likely cause for both of the above is the following:
Back in Q1 or Q2, about 1.5years after the initial Esther rumours, VIA execs were becoming increasingly restless. After their long marketing lull in H2 04 they urgently needed to make some noise again, and they knocked at Centaurs door and demanded that the C7 was to be finished ASAP. It must've been around this time that VIA Chipset engineers and Centaur people sat down and talked for a change. They then quickly realised that they didn't even have a northbridge in the pipeline that supported the much rumoured new bus for the C7. And the above mess ensued.
Once again, it will boil down to a communication issue. I believe that most centaur folk have a decent grasp of the english language, and so do most of VIAs marketing folk (who seem to be hired solely on the following: "Tick box when you're a native english speaker"). But the quality of english of VIAs own engineers and driver writers is downright sickening. Small example: according to VIAs .tw engineers the unichrome has grammar correction and a virtuous gueue.
Cultural differences go a long way in explaining why VIA is such a Frankensteinian company.