How will AMD Ati merger benefit either company?

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Just a few thoughts:

What I'd hope to see out of it is a GPU that decreases its power consumption. Hopefully AMD can apply the same ideas to teh GPU space and save some wattage. Its annoying to see GPU's skyrocketing power consumption... kind of starts to remind me of prescott levels.

Plus I think it makes sense for AMD to acquire Ati. Their processor lineup is going to take a beating and lost sales because of conroe, so they need another 'product line' to bring in some $.
Ati looks like its on the losing side since its doubtful Intel will renew their license (besides, it didn't look good anyways since Intel had ramped up their production so they wouldn't have to buy as many chipsets from Ati).

Shrug, I guess AMD made an offer Ati couldn't refuse.
 
I have always used an ATI/AMD config with all 7 of my builds.
Its great to know that over all of the years they have finally merged.
 
will this is kind of good news. gives us a run for our money.
all in wonder x1900 + 3800+ windsor = im all happy. but gotta wait til the video card goes on sale.

and both companies have a very great customer service department ( for me at least)
 
This should help AMD concerning chipsets. AMD has always made it a point to rely on others to develop chipsets for their motherboards. AMD has released their own chipsets, but that was mainly to get their CPUs onto motherboards for distribution. Once other companies stepped in like VIA, nVidia, and SiS, AMD backed off on their own chipset production.

I think someone stated this acquistion no longer makes them "a slave to chipset makers." That's quite true, but that was AMD's own doing in trying to cut down on costs. While I think aquiring the ATI to make chipset for AMD is great, I'm not that thrilled about the GPU division being acquired. For one thing, it does limit choices for people when they decide to build a PC.

The licensing of Crossfire technology for Intel chipsets has been rescinded. nVidia's decision to form a closer alliance with Intel. Both situations doesn't really seem to give the consumer the best selection of choices. In the worse case scenario, Crossfire technology will only be for those people owning AMD CPUs, while SLI will only be for those pople owning Intel CPUs. Not much of a wide range of choices for the gaming enthusiast who wants to build a premium gaming machine with the best components.

As an example in this worse case scenarios, what happens if AMD produces a hypethetical killer CPU, but their next generation Crossfire Technology sucks compared to nVidia's SLI next generation solution? You'll have an awesome CPU, but the GPU will drag you down. The opposite would be true for Intel; a hypethetical crappy next generation CPU (compared to AMD's) that is paired with an awesome nVidia SLI solution.

Only time will tell how well this merge will be. It is entirely possible that the production of both CPUs and GPUs can make AMD somewhat of a lumbering giant with two primary products to develop. On the one hand will be the CPU that takes lots of money for research and production, but have a relatively long life cycle. On the other hand is the GPU side of the business that AMD will also need a lot of money to research and produce the next generation GPUs. But GPUs have a relatively short lifespan. Many people expect a "refresh" every 6 - 9 month, then an entirely new GPU architecture, again, every 6 - 9 months. In between the CPU and GPU will be the chipsets; the business they really should have focused on years ago, not now.

I would prefer if AMD were to spinoff the GPU side of ATI in the future, and retain the chipset side for themselves. Nothing like a CPU production company that has the important bases covered (CPUs & chipsets) and be nimble enough to change to market demands. They could also develop their own IGPs as well for the not so enthusiastic gamers market.

These are just my initial thoughts. There are plenty of things to think about and consider before I really am able to form a real good opinion on this issue.
 
Will the company be able to use the same manufacturing process for both processors and graphics cards? Will we see 65nm at the same time for graphics and processors?

Is there some part of the arcitecture that is common between graphic cards and processors?

If not. How will this merger save any costs?

The manufacturing process is different. IBM had a difficult time when they first tried manufacturing gpus for Nvidia. I think ATI and AMD will need a period of time to form a symbiotic relationship - AMD engineers will need to learn what ATI tech requires, and ATI engineers will learn what AMD can fab. There will be some growing pains, but they are surmountable.

I think the biggest problem in both these companies is that neither one knows how to market their product. They've both always had the technology (Well at least with k7 & up on the AMD side), but both companies seem to rely more on word-of-mouth by enthusiasts.

Intel and Nvidia, on the other hand, are masters of marketing and branding. I think more people (non-enthusiasts) think of AMD as "compatible with Intel" rather than as a progressive company.
Same with Nvidia. I'll bet if you surveyed the average people, more would have heard of SLI than Crossfire.

I hope AMD/ATI recognize this weakness and enlist some outside marketing help.
 
How about a 32 core chip - 16 dedicated to "CPU" processes, 8 to Physics, 8 to "GPU".

It all becomes more likely as the number of cores on the die go up. We're already seeing engineering samples of 4 core products. How long before 8 and 16? 1 to 2 years at most? At the same time most software can barely utilize 2 cores, and it'll take years until that changes. Even then, most daily activities don't need more than 2 cores - It's only once you start gaming, video rendering or doing other serious number crunching that we're going to need the extra horsepower.

Both Intel and AMD have been talking about multi-core processing with many non-specific taskable core on each die. Think Sun or Cell and being able to specify what cores should do.

dude, cores are like pipelines. You got logical ones, physical ones. you only need 1 physics core since it is utterly useless... how about a 32 core cpu where whatever gets whatever? i think that is reasonable, i mean, there are game sand there are games. some of these games are video intensive and others are cpu intensive. i reiterate: cores are like pipelines (more elocuently "shaders") since companies cant go faster they go wider (kinda like a river). yep htats it.
 
I randomly chose a name to reply to cause i didnt want to fnd a post of the author.

AMD and Ati is prolly a sick blow to both intel and nvidia. I mean: imagine a platform where the GPU and CPU co esxist, potentially sharin video ram effortlessly and vice versa. with Ati, AMD could develop not only amazing video cpu bridges/interfaces but also platform specific chipsets that up until now only intel had. Im a gamer, and if AMD and Ati can pull off a good enough gamer platform, which they prolly will, im buying. Conroe might have won this round, but with this merger it looks like AMD might have almost just taken the gaming crown... sure theyve got to make the platform first. I like AMD, and im prolly gonna buy AMD next if the prices they set stay..... oh and yeah, dudes, AMD is prolly going to develop Ati's GPU n such a way that they will be built on the 65nm process soon enough (finally, a video card that doest explode with stock clocks/cooling) even if they only put it into the Ati flagships. I suppose AMD will prolly also start to expand their Fabs even further to acomodate the new addition to the family.
 
Personally I am not excited at all... You see we have had a lot of mergers in the graphics industry with many promises about revolutionary features, but not one of them succeded.

1. Rendition+Micron... How many of us remember the advantages advertised of having a foundry in house and then embedded-RAM that allowed for 1Mb memory connections on chip? Rendition vanished slowely after the acquisition.

2. Intel themselves bought Chips and Technologies and tried to make a Processor with graphics that worked at 1GHz but have since failed and cancelled the processor. Not only that, they also failed miserably in their attempts to make a graphics chip of their own.

3. Via bought S3 Graphics and tried many times to launch a chip... They have finally made it, after almost 5 years.

Now I really don't want to sound pessimistic, but there is a reason behind all these failures: Lose focus! As interesting and exciting new ideas may be, if you invest everything there then you are gampling... I really hope they know what they are doing!
 
Now I really don't want to sound pessimistic, but there is a reason behind all these failures: Lose focus! As interesting and exciting new ideas may be, if you invest everything there then you are gampling... I really hope they know what they are doing!

I think AMD is smart enough to let ATI do its own thing. Sure, they will combine techologies for an IGP (by combine, I mean let ATI do it 😛 ), and they will collaborate in order to fab, but AMD should know its limits by now. The problem with the mergers you listed, were that the absorbing company tried to dictate the other with the result of, as you say, they lost focus.

I doubt that on the graphic card end, AMD will even change ATI's name on the boxes. You can't risk a newbie customer seeing AMD's name on a x1900xt, and thinking it will only work on an AMD system. (This has been going on for years - Magnavox once made a game system, and showed people playing it on a Magnavox TV in their advertising - The whole line failed because people thought you could only use it on a Magnavox TV.)

Technology is advancing, but people aren't getting any smarter. :lol:

AMD has put their whole company on the line with this. If they lose focus, they will destroy ATI too.