There's no price stated yet - MSI is remaining tight lipped - however given the components used we expect the Fuzion to be expensive. Very, expensive.
A whole boards worth of high-c caps and DrMOS MOSFETs are a significant investment unto themselves, let alone the FAE and QA time invested in this one off solution and the additional cost of the Hydra chip. Unless Lucid has struck some sort of deal with MSI, we expect the company wants some preliminary cash flow pretty sharpish as well.
I wouldn't be surprised if the Fuzion easily costs as much, if not more than premium X58 boards, unless MSI intended this to be some sort of Bugatti Veyron of the motherboard world, selling it at a loss in order to gain some engineering and marketing kudos. The difference is, the Veyron is an exceptional machine. The Fuzion is not.
Addendum 10:30am 8th Jan 2010:
We've been informed by eagle eyed community member, M7ck, that it's currently in stock for £268.55
It works! Well, to some extent. We didn't get a blue screen - not a crash not a driver issue or conflict with the Hydra software. It installs flawlessly, it's simple to enable and a little logo shows up in games to let you know it's all fine and dandy. In fact, there was less headaches than we're used to with SLI or CrossFire. Usually new or early CrossFire or SLI drivers/games will simply bluescreen or crash out, or simply refuse to run altogether. Hydra didn't fuss like this and was ultimately quite stress free.
The motherboard is very tasty - it's kitted out to the max and takes on the likes of Asus' Maximum III Formula on the spec sheet, although we've yet to test its overclocking potential, SATA and general productivity performance.
Firstly, see the anticipated price above. We're not going to be jumping over ourselves to get one since it will be a serious investment. I'm sure there is a premium that would be paid for the privilege if it had worked, but it has to be a worthy investment that generates a considerable performance improvement, which just isn't the case.
Next: it's fussy. Really fussy. There's no Crysis
support and no Modern Warfare 2
support for crying out loud! The former is still the most requested benchmark we run and the latter is the biggest PC game of not just the last year - the last few years. Why did we use Call of Duty 4
? It's the closest thing we have to it, and still played online by a great many people. In fact, the games we can choose that people actually play
, let alone test reliably, are slim pickings. The list may read long but when you see Lego Indiana Jones
, X-Men Origin: Wolverine
- are people going to be buying an expensive Hydra board and multi-GPU solution to run these games? I seriously doubt it.
That said, in the last two driver releases there's been an explosion of games support across the board. While we still don't have a blanket coverage or the ability to make our own. We'd still hold off and keep an eye on the update cycle.
supported across every mode and works really very well then?
3DMark: Vantage and 06.
Yes, if you spend lots of money on an expensive machine to all day run 3DMark - you're in luck, because apparently it works better than SLI and CrossFire. MSI's own numbers shown to us confirm this, but 3DMark is not a game and we don't waste our time with it here.
The Damnright Ugly (and the bottom line)
The Lucid Hydra is too complicated and too dependent on driver support, you can't force it or make your own profiles like "AFR" in normal multi-GPU setups. We tried making our own profile for Crysis
and it just outright refused to work. So we have to wait entirely on Lucid for any kind of multi-GPU support? No thanks. At least if I run SLI or CrossFire the day a new game drops the chances are I'll get some
kind of improvement by running two cards, even if its in the single percent figures, I'll have a small benefit until optimised drivers arrive.
But, the biggest, biggest kicker is that Lucid Hydra just simply doesn't provide a performance boost.
in our testing since we started did we find a game or situation where we went "Wow, that's a real improvement". Why would anyone ever buy this?
We had rumblings from others last year - Asus and Gigabyte - both of which have been hands off the whole Hydra setup. That and the continual delays rang alarm bells for us and it maybe the case that MSI put too much commitment to it in the first place and had to go through to ship something.
Privately I was told that these other companies had looked into it and already come to the same conclusions we have here.
We must point out that to some extent MSI disputes our finding and claims there is an advantage to its Big Bang Fuzion board: we've seen some scores for Batman: Arkham Asylum
run with the identical driver versions, hardware and two GeForce GTX 260 cards that look somewhat promising under a very specific 16xAA game mode, but we cannot ascertain how this was tested and we strongly doubt it was not a manual run-through like we run our tests. MSI also used a 32-bit OS (and therefore drivers) where we run 64-bit, and if the setup is that fussy, how can we ever accept a consistent performance pattern for any game? While we have shared our findings continually with MSI, including what games we've used, like always we have refused to share specific manual benchmarks as a measure of independence: the game should work reliably wherever anyone tests.
While it was worth getting so worked up about mixed graphics cards the bottom line is that it was and still is a pipe dream. There's no magical solution to make things work better and that's why both Nvidia and AMD commit considerably more manpower to writing drivers than they do developing new hardware.
As a final note, bit-tech
and MSI are committed to continually testing its Big Bang Fuzion through future software releases to see if things do improve, and we'll keep you informed of advances in the platform. If the code is cracked and it works wonders - we'll let you know - however, I wouldn't hold my breath, or wallet, for this one.