TSMC is having yield troubles AGAIN.

http://www.digitimes.com/news/a20110707PD211.html

"Meanwhile, AMD will followits original schedule and enter the 28nm era in the first half of 2012.The company's next-generation graphics chips Southern Island as well as Krishna and Wichita processors, which will replace the existing Ontraio and Zacate processors, and will all adopt a 28nm process from TSMC.

This is not a surprise and if we got the 28nm GPUs this year, they would be early

It seems as though with AMD they completely changed gears once nVidia announced Project Denver. They're probably figuring out how to implement ARM CPUs into their lineup, and my bet is that it's their #1 priority right now. They're also coming out with some very interesting low power "fusion" chips. I just don't think they see the x86 desktop and server markets as being as important as they once did.

Perhaps I am wrong and there were simply technical issues in the creation of Bulldozer, however I have a feeling that there is a lot more going on. Have they even found a new CEO yet, or are they still using that interim financial person?

It will be nice to see 8 core CPUs. The thing is, x86 may well be dead.

Linux is the future. It is metaphysical.

It is impossible to change gears that quickly. AMD's next gen GPU was developed over 3 years ago.

They can still dedicate their top engineers to the projects headed in the new direction.

I'm talking about the Bulldozer delay. I'm thinking that had they kept their nose to the grindstone and stayed focused on it, it probably would have come out on time, performing as it should.

Their GPUs are going to be useful for years to come and it's an enormously valuable part of their company. From a hardware standpoint, "ATi"/AMD has always been a phenominal company.

They have brilliant people making their CPUs as well. They need to pull off another K8. It's just a bit of a gamble at this point going forward with x86. It seems as though it might turn out to be a market with diminishing returns.

i believe that ever since the acquisition, AMD has been *focused* on their Fusion and it is a remarkable thing that they have done with their Lano APUs - they totally blew Sandy Bridge graphics away in just a few years - something a much bigger company cannot dream of doing for a couple more.

Intel will pull it off with the graphics. They've done it before and came up with a very potent solution during the 3dfx days.

It won't take them long, either. They say that GPUs are easier to design than CPUs. If they have to, they will license someone else's technology, buy out a small company, or even buy out nVidia completely (if the government will allow it).

Intel is clearly on top in the semiconductor industry right now. IBM does some interesting stuff, and AMD is a great company for the end user (along with nVidia in many respects).

After the recent "battle" between nVidia and AMD (which was really just a marketing gong show), I don't see them doing too much in the near future unless these marketing people have nothing better to do.

It reminds me of General Motors. They focused on creating a ton of vehicles, and the end result was a bunch of ugly vehicles that no one wanted along with poor reliability. I don't know why these guys don't focus on creating a few quality parts rather than constantly creating this plethora of normalcy cloaked by a marketing department.

In other words they solved a problem CPU's didn't really need to solve while leaving the actual CPU portion outdated and very underperforming. Or in other words they "solved" the wrong problem, especially for the people who have traditionally purchased machines with integrated graphics. Those people demonstrably could care less about graphics performance, but DO care about CPU performance.

i don't think people care about CPU performance. They just want to do their PC to work and they don't care how.

Intel will *never* catch up to Nvidia and AMD in graphics. Never ... Ever.
-impossible.



Lano blows core i3 out of the water in any gaming and in multimedia - what people consider important; with Intel you are forced to buy a discreet card for anything beyond light browsing/office work. If Intel didn't need to solve it, then they can stop wasting their time and hundreds of millions of dollars with putting graphics on their CPU die.

I suspect the corporate world and 99% of all laptop purchasers would disagree with all 3 points you raised.

People who don't game on their computer purchase for cpu performance first and foremost and have demonstrably done so since Pentium 3 and the 810 chipset that often had no APG port soldered onto the motherboards. If you want a more recent example compare intels laptop marketshare to AMD's.

The IGP in intels core3 i* series CPU's is enought to accelerate flash and let bored school students and housewives play flash games on facebook. If you want to "game" more seriously than that then you add in a dedicated video card, something that won't happen for either class since the students game seriously on consoles and the housewives only play facebook games (or The Sims).