It’s a constant struggle between AMD and NVIDIA for the performance leadership position. For the last few years NVIDIA with their 8800GTX through the GTX 480 has had the single-chip leadership position. AMD countered the GTX 480 with their dual-chip HD 5970, which was slightly more expensive but outperformed the single chip GeForce card, as two of the AMD chips were a bit much for the card to handle.
NVIDIA announced the GeForce GTX 480 card late last year as the GF100 Fermi architecture. Due to the size of the chip and other problems the first card based upon the new architecture did not release until early this year. The 480 is a good card but was not the card they announced with 480 SPs instead of 512 and one less Tessellation unit. It’s been about six months since the release of the GTX480 card and NVIDIA has launched a top-to bottom lineup of cards based upon that architecture. Generally NVIDIA releases a ‘kicker’ in the middle of their release cycles of a new architecture.
With that in mind NVIDIA had a real decision to make on their naming scheme with the newest card in their lineup being announced today, the GTX 580. The GTX 480 name and all of its variants does not make sense for the new card in some ways, as we’ll find out in the Features section of the review. I suppose NVIDIA could have called it the GTX 490, but that would mean that they would have no room for derivatives to replace the GTX 470 or 460 cards if that were the case. One could also be mindful that AMD just released the HD 6870 which is not as high performing a part as the last series of the HD 5870 but at least NV can keep the 80 suffix for the high-end, and lower numbers in the 5 series depending on the part. In any event, today’s launch of the GeForce GTX 580 is in essence a fixed fully functioning GTX 480 card that was announced 7 months ago. This is an article on the GeForce GTX 580 reference card it should be interesting.