Nvidia native PCI Express comes in Q3
Shorty
Manchester, UK Icrontian
GRAPHICS FIRM Nvidia has already introduced its PCI Express.
In the second quarter, Nvidia plans to invade the market with its PCI Express PCX series which are used with the NV19 to NV39 old chips. But you actually won’t see any PCI Express platforms until very late Q2 with serious availability until Q3.
Nvidia has four cards that will be native PCI Express starting with NV45 PCI Express version of NV40. There is an interesting thought about this chip anyway. Someone suggested that NV40 might be a native PCI Express chip with inside bridge to make it AGP capable. We don’t have any proof for this but it sounds very interesting.
NV41 is a performance card that should cost $200 to $300 and then Nvidia will have an NV43 mainstream part. Also the PCI Express card should replace PCX 5750, NV36+bridge. The PCX 5300 NV34+bridge card that replaced FX5200 will be replaced with NV44 card named PCX 6xxx.
oops, but you can still not do anything with those cards since you can not plug it anywhere.
Source: The Inquirer
In the second quarter, Nvidia plans to invade the market with its PCI Express PCX series which are used with the NV19 to NV39 old chips. But you actually won’t see any PCI Express platforms until very late Q2 with serious availability until Q3.
Nvidia has four cards that will be native PCI Express starting with NV45 PCI Express version of NV40. There is an interesting thought about this chip anyway. Someone suggested that NV40 might be a native PCI Express chip with inside bridge to make it AGP capable. We don’t have any proof for this but it sounds very interesting.
NV41 is a performance card that should cost $200 to $300 and then Nvidia will have an NV43 mainstream part. Also the PCI Express card should replace PCX 5750, NV36+bridge. The PCX 5300 NV34+bridge card that replaced FX5200 will be replaced with NV44 card named PCX 6xxx.
oops, but you can still not do anything with those cards since you can not plug it anywhere.
Source: The Inquirer
0
Comments
BTW, they did not have to really bus-bridge for GPU itself per se, just slow it down (clock sync chip changeout, not GPU alteration in GPU IC) and use slower RAM, Shorty. I would expect something on the lines of DDR2 dual channel bandwidthing on the PCI-Express NVidia cards for the fastest Nvidia PCI-Express cards.
THEY should, in Q3 or Q4 or Q1-Q2 of '05 when bugs are removed, be blazing fast in high-end cards. Murphy and heating issues say there will probably be some very much needed hardware and BIOS timing detail fixes and also some driver patching needed once these cards hit the market. I would guess in a year or so, higher-end folks will be using Opteron duallies and PCI-Express video cards to massively power-dev graphics and multimedia, though.
John D.