Connect with us


GeForce GTX 400 series available in higher volume than Radeon HD 5800?




One of the most common rumors about the availability of GeForce GTX 480 series was that the problematic yield would only result in 9,000 boards available worldwide. Sadly for some, truth is completely different.

During the past few weeks, we tried to find out the launch volume of GeForce GTX 470 and GTX 480. Bear in mind that our conversations with nVidia’s own partners didn’t end well for the Graphzilla of Silicon Valley. In our talks with various company owners and high-level executives it was well, obvious that the situation is just not good for nVidia’s ecosystem as a whole.

The default line given to us by our nVidia contacts usually resulted in a comment such as “they’re our customers, why they don’t ask us directly?” However, miss-information that nVidia gives out to partners was explained to us in great detail, including the ways how company tracks for leaks. Looking into the future, that is also the biggest danger nVidia faces. For instance, nVidia officially announced that they’re launching the part this coming Friday, yet most of the partners told us that they only have the product boxes ready – and no actual products. Everyone is waiting for nVidia to ship them the parts properly labeled, so that they can pack them in the boxes, add appropriate bundle and ship them by air to their respective distributing chains. Then again, AMD does the same thing – this is just a reality in life of an AIB [Add-In-Board, AMD nomenclature]/ AIC [Add-In-Card, nVidia nomenclature].

Having witnessed AMD launching Radeon HD 5800 series with insufficient number of boards, this would be nothing weird or unusual – according to our information at the time, ATI only had 20,000 Radeon HD 5800 boards worldwide. Our sources at the time were dismayed at low availability of Cypress GPUs. Unfortunately, the levels of Cypress availability aren’t great even today, given the amount of limited-availability products [Radeon HD 5970, for instance]. In the past two weeks, though, there has been a large shipment of Cypress GPUs to counter nVidia’s launch of GeForce GTX 400 Series. In any case, when the products finally debut later this week and come into retail over the course of this and next one, you’ll see two things – larger availability than the previous DirectX 11 GPU launch and a shortage due to the fact that some AIC’s already told us that their allocation is sold out completely.

For instance, we learned that more than one AIC vendor ordered 10,000 boards all on their own, and that was the volume higher than some of malicious rumors that were spread around by ill-informed individuals. While we do not have the complete number for “Day 1” of sales, we know that nVidia sales team approached all sizes of AIC manufacturers and even accepted the terms from vendors that only wanted to see the GPUs. Yes, we should even see a custom design GTX 400 at launch. According to leaked images online, don’t be surprised if we see more than one board design carrying DisplayPort connectors.

Getting back to the subject of nVidia’s allocation and how come it has more than 2:1 ratio with Cypress GPU at launch, we learned that there is quite a good reason why nVidia’s GeForce GTX 480 will come with 480 cores, as we first reported here. According to the sources at hand, nVidia was only able to produce a lowly board figure with 512 cores enabled [low 10s of thousands], thus there was a situation with potential shortage and a media and market backlash. The company acted upon the situation at hand and limited the number of cores, radically increasing the amount of functional GF100 dies. As a result, there should be more than 50,000 Fermi boards [this may not be the final figure, as we could not contact all AIC vendors and OEMs] available in the first 10 days of sales. According to our sources close to heart of the company, nVidia wants to overtake the number of Cypress GPUs [both Hemlock and Cypress, i.e. 5800 and 5900 Series] by the end of this summer. This does not include the mainstream parts of Evergreen family, to which AMD is steadily approaching the number of three million shipped dies [the three million mark should be hit very soon, if not hit already].

We also learned that nVidia shifted allocation of its 40nm wafers from lowly DirectX 10.1 parts into big-die Fermis – those numerous 40nm DX10.1 GPUs weren’t only nVidia’s plan to capture those 80% of Arrandale designs, but also to keep the highest possible allocation of 40nm process node for itself, and appropriately shifting the allocation from cheap DX10.1 GPUs to high-end Fermi dies. In any case, a clever strategy.

All in all, regardless of you being Green or Red, this will be a good two weeks for graphics. If you chose to acquire Red cards, there will be a series of “adjusted pricing” across the board, while Green cards will show their teeth and show themselves among numerous e-tailers and retailers.

Original Author: Theo Valich

Webmaster’s note: This news article is part of our Archive, if you are looking for up-to date articles we would recommend a visit to our technology news section on the frontpage. Additionally, we take great pride in our Home Office section, as well as our VPN Reviews, so be sure to check them out as well.

Continue Reading