<font color=red>NVIDIA Everywhere
AMD and NVIDIA are getting very tight, and it shows. The GeForce 3 was designed to work well with the AMD Athlon, and NVIDIA also made sure that the chip/card will work on all the chipsets that AMD supports (AMD 760, KT133, KT133A, KT266, SiS 750, etc.). At the Tech Tour NVIDIA is one of four Platinum members, and it is only one upped by Microsoft in that small group. While not much was said about future products, some hints were dropped here and there that really sheds light on what NVIDIA will be doing very soon. The NVIDIA people didn't really tell me much, but after the show I had a two hour drive back to home to think about all of what I did see and hear. With what little information I was given, I came up with a bit of speculation.
It is well known that NVIDIA hired quite a few Aureal engineers and support people, and much of that technology (technology not owned by Creative) is now being applied to the MCPX (media and communications processor for the X-Box). NVIDIA has already confirmed that it has completed the design of the MCPX, but it looks as if this product will not only be used in the X-Box.
NVIDIA is also developing the Crush chipset for the AMD platform, and it will have the most advanced capabilities of any integrated chipset. The Northbridge will support DDR memory, utilize the graphics core of the GeForce 2 MX 100, and it looks as if the Southbridge will actually be the MCPX. It seems that Microsoft really won’t care about NVIDIA selling the MCPX as a separate product, as Microsoft is mainly concerned with the software (which is where it makes its money). Microsoft wants great support for the X-Box, and with products such as the GeForce 3 on the market, software will easily be ported to the X-Box and PC. Utilizing the MCPX for integrated chipsets for motherboards fits nicely into Microsoft’s plans for having great support for its software. This also saves NVIDIA time and money, since they only have to develop one southbridge that is packed with features that no other chipset manufacturer has (such as broadband support, high quality sound, and most interestingly the Hypertransport technology from AMD). Initial numbers from the Crush chipset are VERY good, and according to sources we should see the first version of Crush released by the beginning of Summer. The second version will be very similar, but will have support for an external AGP controller.
Taking this into consideration, it is also very feasible that NVIDIA will release the audio portion of the MCPX as a separate product. This again enhances the software support for the X-Box as developers have yet another reason to develop their software to run spectacularly on X-Box components. NVIDIA is also in a position financially to take on Creative Labs, as many, many board level manufacturers are screaming for a high quality sound chip (manufacturers such as Turtle Beach, Hercules, and many others). The loss of Aureal was devastating to the sound card market, as no one company has stepped forward to compete with Creative. Now that NVIDIA looks to have revenues of around $1.1 billion US this year, it has the muscle to take on Creative and stay competitive. In the past year and a half since Aureal went belly up, the sound card market has faltered, and many major OEM’s have gone back to Creative products exclusively (since there is no competition). This field has stagnated, and the market is ready for a new influx of products and technology, and NVIDIA is poised to deliver.
AMD Tidbits
The processor roadmap is essentially unchanged. Mobile and 2P Palominos are already in production and are being delivered to the major OEMs supporting notebooks, laptops, and 2P systems. AMD is able to rely upon the Thunderbird until the end of Summer due to modifications in the fabrication process that allows cooler running and higher speeds. At that point desktop Palominos will be released, and it will of course have many architectural changes that will make it a more "efficient" processor in terms of power consumption, and more interestingly performance. On a clock for clock basis, a Palomino will outperform a T-Bird. It was not released what these changes were, but we will find out about them within the next 2 months.
Hypertransport will be a very good technology for many companies and their products. AMD showed a diagram that was somewhat interesting, and perhaps sheds a little light on future products. This diagram showed a dual processor setup with Hypertransport connecting the Southbridge to the Northbridge, and had a few things like dual ethernet in between the two, as well as SCSI and other products (does this by chance sound like the Tyan/Guiness board?). The 760MP may in fact support Hypertransport, but it is not official yet. AMD publically stated that it is going to release a new Southbridge for use with the 760MP (as the 766 is not well supported or feature rich). Further speculation may in fact lead to the Southbridge being the MCPX, but that is a bit of a stretch (but again it would make sense as the design is already there and it utilizes Hypertransport, and is again feature rich for a good price).
The 760MPX (Paulaner based boards) will be the next AMD MP design, and it will include more features such as the 64bit/66MHz PCI interface. This will most definitely utilize Hypertransport on the motherboard for all communications from the memory controller and the peripherals. Hypertransport is a very powerful technology, and it is essentially available today for manufacturers to implement.
Conclusion
AMD is a very smart company, and it is making sure to work with as many partners as possible to help everyone succeed. AMD needs these partners to broaden its technology base, and when AMD succeeds, it brings along these other partners, so they all end up profiting through the experience. AMD is willing to share its technology with anyone that wants to support AMD processors (this technology includes DDR tech, Hypertransport, and many other emerging technologies that have not been announced as of yet).
AMD is not trying to be an Intel, and this looks to be where AMD will succeed. AMD is content to let its partners do much of the groundwork based on AMD technology, as everyone profits from this philosophy, and when everyone profits, they will be more willing to work with AMD in the future. Intel tried to buy everything and do everything, and that alienated some of Intel’s staunchest allies in the technology field. AMD is hoping not to make that mistake. AMD is a company to keep close watch on.
penstarsys.com
Sorry if duplicate, but I enjoyed this page of article.
M. |
