Nvidia Corp.’s chief financial officer said that lower-cost Fermi graphics chips would emerge sometime in the middle of the year and around the same time the company would release its next-generation Quadro professional graphics accelerator based on the Fermi micro-architecture.
“Fermi will show up as the GeForce and Tesla first and it will definitely show up first in the highest-performance configuration. Sometime in the mid-part of the year we will see a lower-cost version of that come out both for Quadro and Tesla,” said David White, chief financial officer of Nvidia, at Goldman Sachs Technology Conference.
Traditionally Nvidia released its Quadro graphics accelerators considerably later than the GeForce graphics cards for consumers, hence, there is no surprise that there will be no new Quadro at the launch of the GeForce GTX 400-series on the 26th of March, 2010. Still it is a surprise that Nvidia, perhaps, even decided not use its flagship GF100 graphics processor for the next-generation Quadro graphics boards. Of course, Nvidia does have very good positions on the market of professional graphics accelerators and can allow itself to postpone the launch of a new Quadro in case it has issues with software or thermals. However, Nvidia has had the GF100 (NV60, G300, GT300) chip for over six month now and the drivers should be more or less polished off.
There will be one chip – GF100 – powering high-end GeForce GTX 470 and GeForce GTX 480 high-end graphics cards as well as Tesla C2050 and C2070 add-in computing boards, it will only have different configuration. It is rather clear that GF100 is an expensive chip and it will not be able to power mainstream solutions. As a result, it is completely expected that Nvidia will tape out and release a new graphics processing unit (GPU) – some call it GF104 – for more affordable accelerators. It is a surprise that Quadro FX 6000-series will not feature the GF100, at least, based on the information provided by the CFO of Nvidia. Nvidia also confirmed that in Q2 – Q3 2010 graphics processors powered by Fermi architecture would become mainstream for the company.
Nvidia’s financial chief reaffirmed that the company’s Fermi architecture is equally good for computing, graphics processing or professional usage thanks to “modular design”.
“We did something different when we announced Fermi: we announced it at a compute conference, not a graphics conference. A lot of people interpreted that announcement to mean that this is compute device and we are giving up gaming. That is not the case. Fermi is an entire architecture and it is a modular architecture that allows us to scale the feature-set depending on the market need. So, we will taping out and announcing a whole stack of Fermi products that will be very specific for computing at the very high-end of the stack and others will be targeted at desktops and personal computing,” added Mr. White.
The first graphics cards based on Fermi architecture are projected to become available in late March, 2010.