No announcement yet.

Fermi: Big, hot, slightly faster & uneconomical...

  • Filter
  • Time
  • Show
Clear All
new posts

  • Fermi: Big, hot, slightly faster & uneconomical...

    So Fermi was rumoured to be big and to be running hot... confirmed by reports which outlined the neutering of the original design to fit thermal envelops... Now that they've taken care of "hotness", it is still big and quite uneconomical and it makes sense why Nvidia might be re-spinning this as soon as possible with a shrink and design improvements and well, it was confirmed by a story about FERMI II... For what its worth, i really hope Nvidia has some sense and it sells it at a reasonable price, so we users get some respite from excessive prices on the latest hardware...

    Nevermind me and my rant, read the story...


    THE GRAPHICS CARD INDUSTRY is an odd one, as counter to other parts of the computing hardware market individual products aren't always judged on their own merit. In fact for the most part, when weighing up which of the big two companies to pick from when choosing a GPU, most will go with the current market leader. However, the market leader isn't determined by which company makes the best card for your price point, but the one that produces the most powerful high-end card.

    This creates a bit of a strange paradigm, as graphics card manufacturers have to push their top of the line performance products to be as fast as possible, while concentrating on the far more lucrative budget to mid-range cards. The vast majority of consumers can't, and won't fork out the several hundreds of pounds required for the absolute best card. This means that getting your mainstream GPU right while doing well at the top end is a strange balancing act that plays out with each new generational release.

    Normally the new range launches from either camp take place around the same time, with a little bit of leap-frogging thrown in for good measure. However, with the current 'new' generation of GPUs, that isn't the case at all. ATI has a firm grasp on the market, beating out rival Nvidia to the release date by several months. Will Nvidia's long in-development Fermi hardware allow it to spring back and retake the market now dominated by ATI hardware, or is it going to be too little too late?

    Performance numbers from the Green Goblin's camp have started to emerge, and without doubt Fermi appears to be rather quick on its feet, pushing maybe 50 per cent faster than Nvidia's current big boy, the GTX 285. Pretty impressive, and even compared to ATI's top of the line single card it fares pretty well until it comes face to face with the current monster from the boys in red, the 5970, which seemingly will walk all over it. Of course these performance numbers are all conjecture for the moment, but as more start to be released, it's expected that the results will be similar to those already unveiled.

    Speculation aside, what we do know is that ATI has a very strong hold on the market place currently. Not only does it have the top performing single GPU card available, but the top dual chip card too. Not to mention mid-range performance parts by the truckload, including the 57XX range and the recently released 56XX and 55XX cards. Since we've already established that having a powerful top end combined with a solid mid-range is the key to money making in the GPU world, ATI appears to be in the perfect position to gain significant market share. Combine this with a release date from Fermi that's still at least a month away and poor availability on the current Nvidia top end cards, and its no wonder ATI proudly announced at CES this year that it had already sold over 2 million of its 5 series GPUs.

    Martin Sawyer, sales manager at performance PC specialists, has some harrowing sales figures for Nvidia. In the pre-Christmas rush the firm sold more than 20 ATi performance video cards for every GeForce, and with Nvidia stock now hard to come by the gulf between the two has only widened.

    Despite all this, the green team might surprise us all and come out swinging with hardware that captures the market so well that those with new 5 series cards jump ship and go 4 series, and those waiting for the upgrade pay top dollar for the higher end cards; but it is unlikely. However Nvidia has made it clear that it'll be opening with the mid-range cards in an attempt to capture whatever is left of the soon-to-upgrade market, which could pay off. Though, if initial perceptions of the card aren't that it beats ATI's mid range, but are in-fact that the card "isn't as good as the top end", whether fair or not, it could mean that Nvidia will need to start thinking about Fermi v2 before long.

    All this is nothing particularly new in the graphics game though, as the back and forth struggle between the big two has been ongoing since the early 2000s, which saw each company vying for position and desperately hoping that the other would drop the ball; as Nvidia seem to have done here. The earliest instance of the all important round object slipping through one of the player's grip was with Nvidia's ill fated 5 FX series which saw ATI walk all over it with the range of 9XXX cards; most favourite of all the 9800 pro.

    Things evened out between the pair during the 6800/7800 vs X800/X1900 days, but ATI's popularity and performance crown took a nose dive when it came to facing off against Nvidia's monstrous 8800GTX card. It trounced the then redder than ever company for a whopping two years before it could counter that effectively.

    Still, this sort of back and forth behaviour is what you and I as Joe consumer want. Competition is good, because it invokes product progression as well as price wars which mean cheaper products across the board. When Nvidia ruled the roost for a solid couple of years, prices didn't fluctuate that much as it had no need to sell itself to you on value; performance alone was enough. Hell, it took them over a year to release an affordable version of the GTX, the 8800GT.

    From then on ATI clawed its way back slowly with the 3 series and eventually retook pole position for a while with the 4870s and 50s, leaving Nvidia playing catch up despite not being too far behind. We're now seeing something similar though a little more extreme, as Fermi has had its launch pushed back from November 2009 all the way to March, showing us the real effects of early market saturation leading to a dominant position in "next gen" trading. ATI clearly learned from their partnership with Microsoft when creating the 360 that getting your goods to market first pays off. The Xbox with its ATI powered graphics trounced the PS3 mostly by being available first. This allowed Microsoft to keep its hardware competitive price wise, completely undercutting the Nvidia powered PS3 at launch.

    DirectX saw a resurgence as a viable selling point with the most recent ATI hardware, as it managed to coincide the unveiling of DX11 nicely with the release of its compatible hardware; this being compounded by the near release date of the supporting OS, Windows 7. Some enterprising Inq readers pointed out to me that Vista also supports DX11, but this doesn't take away from the fact that most of those moving to the new operating system will be doing so from Windows XP, meaning new PC builds bundled with a nicely compatible ATI DX11 capable GPU is going to sound much more attractive than one sporting an obsolete standard Nvidia card.

    Time will of course tell how this quickly becoming current-gen battle turns out, but either way, Nvidia has to hope that Fermi provides a solid base to springboard to the next generation, as you know that ATI will already be more than hard at work developing its next Nvidia killer.