My answer to your first question:
Probably the quick-and-dirty way to get that info would be to go to shop.amd.com and search for the respective CPUs; the initial results will be mixed (CPUs and systems), and you can use the refinement tools at the left side of the page to get the specific info you're looking for.
The info will not be the absolute most-up-to-date that you could possibly get, but it will let you know who has had systems available for at least a short while. This isn't an AMD problem, per se: even system manufacturers sometimes seem to not know what they're producing, as information available to potential customers is often at least 1 generation old (and in some instances, can be several generations old). Blame this on capitalism: companies actually save more money (or so goes the argument) by not staffing for output-stream (product info) updates than they would make by presenting the latest information; this reduces both their labor and their bandwidth costs, and discourages bothersome technostalkers from pestering them with questions about systems over which the "potential consumer" is only going to drool (and/or the related plaintiff gripes from the "why can't I get a ___" or "you should do/make ___ instead" crowd).
It's a pain to do it this way (hence the generic nature of my answer), but you can get the latest useful info from actual resellers (CompSource, Newegg, TigerDirect, etc.); sometimes the info online is more up-to-date than what their sales departments seem to have, and vice-versa on other occasions. I've personally discovered searches through sites like Google, PriceSpider and Yahoo sometimes return links to pages inside retailer websites that aren't accessible through the retailer's website's search engine -- so, just in case you don't immediately find what you're looking for doesn't mean it isn't available. It all depends on how badly you want the information.
My answer to your second question:
For the most part, you're stuck with poring over the info at the motherboard manufacturer's respective websites. Asus, Supermicro and Tyan definitely make G34 boards that support the 6200-series Opterons; there may be others besides these. Note: I think the manufacturers have been slow to update their websites to show Interlagos support; I've seen at least a few pages that showed support only for the 6100-series (Magny-Cours) Opterons despite the actual product supporting the Interlagos CPUs -- so if you call the company and get someone in the sales department who only knows how to read what is on the company's web page ....
My answer to your third question:
I think the Supermicro H8DG6-F is also marketed as a workstation board; there may be others, but all -- including the H8DG6-F -- are primarily marketed as server boards. OS compatibility varies by manufacturer, model and revision -- and occasionally, by serial number group; moreover, because these boards are marketed for servers, OS-compatibility testing and certification is far more rigorous than what might be required from a consumer-oriented perspective -- meaning that you might be able to successfully use an OS that either hasn't been tested or that hasn't been certified on the particular board you're considering, as long as you don't need everything that such certification indicates.
The respective board manufacturers have OS-compatibility lists for their boards; they're not sexy, but -- with few exceptions -- they're available. Note: it's conceivable that all the G34 boards have by now OS-compatibility lists available online; however, a couple of months ago, that wasn't the case (I got something akin to the "Under Construction" warning a few times).
I believe that I have seen comments in other fora from users claiming to be running VMs on Interlagos hardware; however, I am not involved with VMs in my current project, so I can't personally say for sure. On the other hand, and for whatever it's worth, the such comments of others seemed credible to me.
As to your final comment and question:
Intel is a behemoth having vastly greater resources than AMD; this doesn't just apply in terms of R&D funds and access to intellectual capital, but rather extends to every aspect of marketing: it gives Intel the flexibility to reallocate profits to "compete more aggressively" against Intel offerings, either directly (as by price) or indirectly (as by advertising) or both.
And no one can deny the incestuous relationship between Windows and Intel (sometimes derisively, "Wintel"); since Windows is ubiquitous, and since almost every consumer thinks first of a "Windows computer" as an "Intel Inside" product, the bandwagon mentality drives the market -- especially in depressing economic times: mainstream consumers are risk-averse; therefore, they're unwilling to change -- or to explore whether their "open-minded" prejudice has blinded them to a greater truth. That, of course, is in part proof of the propaganda dominance of Intel and Microsoft, and proof of the efficacy of their methods to stifle their competitors. Sadly, hardware manufacturers are slaves to Wintel dominance, held hostage by market forces.
The foregoing is not to indicate that Intel CPUs are entirely junk: we have seen with the Core i7 CPUs that Intel can deliver good performance in a package only slightly overpriced. For instance: the 4-core i7-2600K is currently about $320 with free shipping to the contiguous 48 US states; for the same money (+/- $10), you can get a pair of 6-core Opteron 4234 CPUs that, together, deliver competitive performance even under Windows (which loads the cores in a way that favors Intel's architecture and simultaneously disfavors AMD's architecture).
I'm not terribly familiar with the present Xeon CPUs, but know that earlier generations were in their day formidably powerful; the modern versions seem to have capitalized on Intel's "GPCPU" strengths. The net result is a system that delivers excellent performance at competitive (and sometimes better) power-consumption rates when contrasted against application-equivalent Opteron-based systems. AMD products presently hold a slight edge in terms of capital cost, but Intel has shown signs that it may be preparing to dominate the low-cost market as well: if it does, AMD could disappear -- and with it, historic definitions of what used to be the low-cost market. This is consistent with the strategic aims of Intel and Microsoft (and other companies driving the rush to cloud architecture):
Mainstream consumers ultimately end up with (comparatively) dumb terminals; their intellectual property is uploaded to the cloud, where it is scrutinized by various agents of security -- from anti-virus and anti-spam filters to bots run "by or on behalf of" credit agencies, the FBI and "Homeland Security." Privacy as a right enjoyed by businesses and/or real persons will cease to be anything but a delusion shared among those unwilling to acknowledge reality; information will be traded as a commodity among "anonymous" qualified brokers who manage their sharing of information as wisely as, for instance, the Allies after they broke Enigma.
In this way, victimization is stratified and results are output-driven; from a QC perspective, it's the ultimate feedback loop: resources are allocated to the persons or classes, enterprises or affiliations the ultimate controller-overseers wish to promote, and withheld from those whose aims or ideology is hostile to the authoritarian agenda. Such resources may be of different, similar or identical nature and quality, and may be applied simultaneously or sequentially --and even randomly across selected (or not-deselected) groups. In essence, the cloud becomes a sort of graphic omniscient device, having the power of its acronym over those subject to it, and itself being subject to the will of the titans that control it.
Alas, the full manifestation of such a monstrosity is yet days away: there are now more pressing issues to consider. Unfortunately for AMD, the battlefront isn't a single one: until recently, ATI was a separate enterprise; however, AMD and ATI have long been associated in the minds of consumers, just as have Intel and nVidia. In the recent past, nVidia spent enormous effort optimizing drivers for its products, and it pretty much blitzed the market with its Quadro 6000. Today, AMD's top pro-grade graphics card is the FirePro V9800, which is in most markets owned by the Quadro 4000-and-better cards from nVidia, and in some cases is regarded as a step backwards from the V8800 (probably because of drivers).
We see something similar happening with the Radeon HD7970 vs the GTX680, and in perhaps less than a year, it will be clear which card offered the greatest ultimate potential, but reviews in which these cards are contrasted presently indicate the GTX680 has measurably -- sometimes, significantly -- better performance than the Radeon HD7970, in terms not only of frame rates and computational output, but also in terms of power consumed and temperature. Some sources indicate AMD currently holds an edge in terms of image quality, but nVidia isn't resting on its laurels, and AMD hasn't indicated a path to better results.
I have opinions about how AMD should develop its products, but I'm not privy to AMD's strategy: it is possible that, if I was working with the same set of information AMD has, I would arrive at the same conclusions as its leadership. When you attempt to divine the causes for the behaviors of consumers, just remember that a representative fraction of those are the people who vote: that should tell you everything you need to know about them.