Good analysis, for sure.

IMHO AMD should not BUY ATI, however they should merge and make ATI an independent division od AMD.

The other options for AMD is to wait and see how intel pressure affects GPU bussines and act the moment ATI has problems.

IMHO AMD cannot do without ATI(or nVidia for that aspect) in the log run. In '08-'09 timeframe intel will have system on a chip designs and AMD will need them to to compete successfully in the Corporate market.

Since when are Intel and AMD considered NVIDIA's "rivals"???

I think it would be more interesting if Intel and ATI merged. They're already like best friends.

This is never going to happen because AMD needs all the cash they've got to be able to fight the upcoming price war with Intel... No cash = no competitive price cuts possible = less CPUs sold = less revenue

They need every single $$ they've got to get 65nm ready, get their new architecture ready and do that while facing *very* tough competition on both price and technology.

If you were AMD and in the situation above would you think it's the right moment to spend all your money and buy that shiny new sports car? :) Well, you get the idea...

Based on your comments it is a complete disaster for Nvidia.

As ATI number 1 enemy is Nvidia the will of ATI actually wanting this to happen it will be very large.
Cheers

Bush is a fucking mass murderer on the loose. New age Hitler. I hope he rots in hell!

On the topic, I like ATi for their superior image quality and tech innovations and I love AMD for their A64 :) and how they proved that Intels race through megaherts was nothing but marketing gimmick.

I hope the possible merger wont have any negative effects on their positive mentioned qualities. If so I myself have no problems with the meger.


-----Bush deserves to be shot as many times as he had taken the lives of innocent people. ------

Very good article!

I don't believe in this specific merger. However, I'm sure that with current market cap ATi will be taken over.

I really hope this doesn't happen, It will be really bad for the entire CPU/GPU industry.

I am not sure about the possibility of a complete acquisition, but I do believe is almost necessary for AMD’s future.

Look back at most major system architecture changes in PC history. Unless it was supported by a multiple big contender (Intel, ATI, NVIDIA, IBM, and others), it failed. AMD does not hold enough market control to make a “Standards” change. The up coming 4x4 technology is going to need a big contender and branding. The only way I see AMD getting NVIDIA/ATI to use their R&D & manufacturing resources to develop a completely proprietary GPU interface for use in the 4x4 system; and take a change on loosing market shares in the mainstream system market, is to take one of them over(at least in part). NVIDIA (My preferred GPU) is a little too big for AMD to try and acquire. Even though they do have a long standing relationship, I don’t believe NVIDIA would be willing to go out on a limb for the number 2 man in the x86 market.

Now the 4x4 technology requires an unparalleled interface between the CPU and GPU. The two basically have to act as one (or 2 if 2 CPUs and 2 GPUs). The interface has to be flawless in design because there are no quick release driver fixes due to the fact that the interface has no driver. The possibilities of the architecture is mostly unknown to AMD; due to the fact that they have no prior experience/knowledge in the GPU market. AMD needs the experience/knowledge base of a high-end GPU manufacturer just to build a suitable interface.

Even if the acquisition does happen, I do not believe that ATI would quit supporting Intel.
1. That would be a large revenue lose that they wouldn’t be able to spend on R&D
2. Major system builders would be forced to take sides ATI for one chip of NVIDIA for both (DELL would be forced to drop ATI product line if it quit supporting Intel).
3. ATI is the key to a few back doors where AMD still isn’t really welcome (DELL).

If all of what AMD say about the 4x4 Technology is true, then this is going to be a great venture for both companies (together or separate). And we shall reap the reward, and Intel will finally bow to the underdog.

atlast ....

with this agreement i hope, i can get more power on my pc.......

i don't believe it until i hear it from the horses mouth.

There are a few quotes to be made....

For one, maybe AMD has quitted the chipset manufacturing but it has virtually transfered essential parts of this business to its CPUs. Memory controller and HyperTransport links to be more specific. Therefore, AMD controlling such technologies, enables it to have a more specific and important role in platform definition than Intel. Intel's own platforms are nothing more than a product tie-in rather than actual platform.

On the other hand, if AMD is sharply increasing its production capacity and making standards as HyperTransport and Direct Connect freely accessible then one may has to resort to IBM business practices before thinking about merges with ATi or other similar companies. IBM with the Power free licencing scheme, is trying to attract h/w companies to create either compliments of its architecture or system on chips (like PA Semi for example) or other Power-based products, that in the end are going to promote its architecture (and therefore products in terms of soft/hard cores, licences, s/w tools etc) and strengthen its chip manufacturing business. Do now all these moves AMD is making have a better sense? I believe they do.

On the chipset front, if AMD and ATi merge, then this would pose a serious disturbance to the tight relations of AMD and nVidia. nVidia has long been the best supplier for AMD chipset logic and this is not true only for the desktop space but for server/workstation systems as well. This is crucial to AMD, as it heavily depends on the Opterons and ATi does not offer a chipset for servers.

Moving engineering intelligence between the two companies is also a debatable argument. For one thing, engineering teams work on very different environments. CPU and GPU designs are very different and that is not only based on the differences of the actual target applications. GPUs are almost entirely design with the use of standard EDA tools and HDL languages while CPUs are based on a much more custom design process (for example look at the giant differences in MHz and power consumption to see this). Moving engineers from one project to the other is not that simple.

Going further, ATi does not have in its technology portfolio, technologies that AMD needs for its current line of products and the same applies for ATi. If AMD had a really convincing presence in embedded cores (it does not anymore), it would be really interesting to see some system-on-chips with AMD CPU cores, ATi graphics and rest core logic and probably memory from Spansion (did I write it properly?). With the current of moves of AMD to sell off Alchemy and Geode processors I don't see this making any sense.

nvidia doesnt seem to be halting the AMDz chipset production.infact they are really happy.

do i have to be a member to post comments and view others?

AMD Products are always Better as we know .
but this time Advance Micro Device inc.. going to complete hold in high end system with complete chip solution as intigrated Gpu + Igp
. ATI also got realy good advantage and sure they enjoy what they going to do Next .

Best of luck
AMD + ATI

Top Class Analysis!

DAAMIT!!!!