|
|
 |
|
|
|
Nvidia has officially launched a series of chipsets: the nForce 780a SLI, 750a SLI, 730a, 720a, and 710a as well as the GeForce 8300, 8200 and 8100 which are all destined for the AMD platform and with HyperTransport 3.0 support. They all ended up with a DirectX 10 GPU and the difference between the GeForce and nForce is the possibility of adding an HDMI output for the former and support of "classic" SLI for the latter.
The big innovation of course is the slightly delayed arrival of Hybrid SLI which is no longer presented. It would seem that perfecting the bios and sufficient drivers was long and fastidious, which doesn’t mean this new flagship technology is immune to certain limitations – at least for the moment. For example, the activation of Hybrid Power is still manual. Finally, you will obviously need a compatible graphic card. Up until now, only the GeForce 8400 GS and 8500 GT are GeForce Boost-capable and for HybridPower-capable cards, you will have to look towards the GeForce 9800 GX2 and 9800 GTX.
 nForce 780aThe nForce 700a reminds us of the 700i and with good reason as it also integrates the nForce 200. Connected in PCI-Express to the Northbridge, this chip enables managing 32 PCI-Express 2.0 lanes that can be devoted to 2 to 4 graphic ports. Consequently, we may fear that available bandwidth between the nForce 200 and graphic cards is much larger than the amount between this bridge and the Northridge. For triple SLI, there are three x16 ports, two of which are cabled in x8. For the rest, we find up to 5 PCI ports, 12 USB 2.0 ports, 6 SATA-II and 2 PATA connections, while there is RAID 0, 1, 0+1 and 5 support. Of course, there is Native Gigabit Ethernet as well as HDA (Azalia) for sound. A slight detour to this rather exhaustive Nvidia PDF, document on the subject could be useful in making a proper choice. |
| |
|
 Intel and Samsung, which are respectively numbers 1 & 2 in the semi-conductor industry, as well as TSMC, much in demand from its clients such as ATI, Nvidia, Broadcom, Conexant, Marvell and even VIA, have decided to work conjointly in the gradual transition to the fabrication of chips made from 450mm wafers. They cite the need for industry wide collaboration and have set a target date of 2012.
It’s interesting to note that R&D costs have become so high that even Intel, which generally prefers to fend for itself in this domain, is now more or less obliged to join others in order to maintain a roughly 10 year cycle. You may recall that engraving on 300mm wafers was put into place in 2001-2002 and production of wafers measuring 200mm started in 1991. The main advantage of going with larger sizes is to lower per unit production costs.
As environmental concerns should now be a bigger consideration, the press release specifies that through more efficient use of energy, water and other resources, bigger wafers can help diminish overall use of resources per chip. For example, the conversion from 200mm wafers to 300mm wafers helped reduce aggregate emissions per chip of air pollution, global warming gasses and water, and further reduction is expected with a transition to 450mm wafers.
In theory, this is something good for consumers, because it should enable reducing per unit production costs and to justify R&D expenditure of the three companies. Moreover, it can generally mean slightly more aggressive prices or for the same price the arrival of better products. Otherwise, we can hope that other large manufacturers also manage to quickly come to agreement on 450mm as competition is often the first factor in reducing prices. |
| |
|
|
Copyright © 1997- Hardware.fr SARL. All rights reserved.
Read our privacy guidelines.
|
|
