When NVIDIA VP Jen-Hsun Huang presented the company’s last fiscal report he also participated in some Q&A and told participants about the Fermi graphics circuit architecture. Even if NVIDIA has refused to admit it, the market has been set for Fermi architecture availability next year for a long time. Something Huang finally confirmed at the conference.
Back in October NVIDIA presented the Fermi architecture but most of the focus was on how the circuit could be used in NVIDIA’s Tesla platform for GPGPU calculations. Jen-Hsun Huang said this was planned and that it will do a similar presentation for its GeForce and Quadro versions of Fermi.
“Next year is going to be an interesting first quarter because in fact, we will need more than ever dies in Q1. And the reason for that is because – and I mean more 40-nanometer wafers than ever because we are ramping in Q1 – we are obviously fully ramping Fermi for three different product lines, GeForce, Quadro, and Tesla. ”
Through its statements Huang makes it pretty clear that we will not be seeing much of Fermi this year, at least not on the market to any major extent. We’ve had this confirmed much earlier and now we just need a specific launch date from NVIDIA. There may be a paper launch before Christmas but much points to AMD playing all alone on the DirectX 11 market in 2009.
NVIDIA VP Jen-Hsung holding up the “fake” Fermi card at GPU Technology Conference
No active posts found.