Fusion of ATI and AMD complete?
By Alex Zaharov-Reutt
Despite leading the race against Intel with the Opteron and the Athlon 64 over the past couple of years, AMD needed a partner with vast chip manufacturing experience of its own to strike back at Intel’s recent resurgence. What does the future hold for the newly merged company?
The rise of the Opteron processor from AMD was the start of a recent golden era for AMD. Coupled with the success of the 64-bit extensions to an x86 processor, and AMD’s later march toward dual core processors that properly integrated two cores onto a single die.
AMD claimed that Intel’s initial dual-core chip, the Pentium D, was just a kludge, jamming two chips together, instead of a true dual-core solution. The market agreed, and AMD held the crown for technological superiority, even though Intel was far more profitable and remains so to this day.
What ‘saved’ Intel during this period was the incredible success of the Centrino platform for notebook and portable computers, the incredible momentum of sales from the decade old Pentium brand name and Intel’s preferred supplier status among most of the computer companies they supplied, including, of course, the then Intel-only shop that was Dell. Oh, and all those profits still coming through the door.
But Intel knew that it couldn’t survive on past glories or wringing out further performance from existing product lines. After a false start with Intel’s Indian R&D labs failing to produce a next-generation processor, it was up to Intel’s Israel R&D labs to come to the rescue.
This was done by re-engineering the Pentium M, born of the Pentium III and crafting it into the Core Duo, and then various versions of the Core 2 Duo for desktop, mobile and ultramobile uses. While it took Intel some time to get going… there’s no question the Core 2 Duo range is a stunning success that has placed AMD on the back foot.
AMD was clearly watching this activity closely, wondering when Intel would get its act together, and calculating when its own processor dominance would end. Now the roles are reversed – Intel is once again dominant, with AMD-beating processors, and an upcoming Core 2 Quad processor that joins two dual-core processors together to create a quad core chip, of sorts.
AMD claims that this isn’t a true quad-core processor, but simply two dual-core processors stuck together. While this is true, it enables Intel to deliver a quad core processor in mid November – in a scant two or so weeks time, while AMD’s own ‘true’ quad-core processor is not due to arrive in the market place until the second half of 2007.
And while AMD managed to finally crack the nut that was Dell Computers, with Dell finally agreeing to sell Opteron and Athlon based systems, it needed something more. A way to grow the company and both add and develop new technologies that would enable it to either level up with Intel, or better still, overtake it again in the processor performance stakes.
But the success of Dell has landed AMD with another problem – that of capacity. Needing to supply Dell with so many processors has caused a bit of a shortage in white-box land, the land of the white-box builder and lower tier computer companies that simply won’t sell anywhere near as many computers as Dell. AMD needed more factories, new technologies, new opportunities and a new story to sell to the world.
The solution came in the form of ATI. A fierce competitor of NVIDIA, ATI not only produces excellent graphics cards, but produces chipsets for a range of other devices including mobile phones, an area that AMD was not really working in at all, unlike Intel.
Interestingly, Intel only recently sold off its mobile phone processor unit to a company called Maxwell. But as these processors were ARM based, instead of x86 based, Intel could afford to let it go. Intel was going back to its core – developing microprocessors and platforms for consumers, business and enterprise use. And getting x86 processors into mobile devices.
So what will AMD do with ATI? AMD plans to use ATI’s multi-faceted experience across all areas of its business, while keeping the two companies mostly separate, at least at this early stage. AMD will continue selling ATI’s powerful discrete graphics cards for the time being, but has a plan for the future called Fusion.
While we won’t see anything until late 2008 or early 2009, which seems a while away here in late 2006, AMD plans to create a new class of x86 processor. It will be one that integrates the central processing unit (CPU) and graphics processing unit (GPU) at the silicon level, combining the best of AMD and ATI in a single chip and no doubt inspiring the Fusion name.
AMD says this is necessary to “provide the best customer experience in a world increasingly reliant upon 3D graphics, digital media and high-performance computing. With Fusion processors, AMD will continue to promote an open platform and encourage companies throughout the ecosystem to create innovative new co-processing solutions aimed at further optimizing specific workloads”.
However taking into account the fact that ATI still sells a great deal of discrete graphics cards, and are likely to continue evolving the technology far into the future, the “AMD-powered Fusion platforms will continue to fully support high-end discrete graphics, physics accelerators, and other PCI Express-based solutions to meet the ever-increasing needs of the most demanding enthusiast end-users”, according to their press release.
Phil Hester, AMD senior vice president and chief technology officer took the opportunity to briefly explain why a CPU and a GPU are needed on the one chip. “With the anticipated launch of Windows Vista, robust 3D graphics, digital media and device convergence are driving the need for greater performance, graphics capabilities, and battery life,” said Phil.
The thing is, by 2008 or 2009, modern computers of that time will be much faster than they are today, with processors sporting multiple cores, very powerful graphics cards from the ATI division and NVIDIA and lots of memory. AMD’s new Fusion will have to ensure high-performance 3D graphics can be pumped out. And naturally, Intel will be closely watching AMD and will be doing everything in its power to out-do them.
If Fusion is an integrated solution that has produced the middling performance of previous integrated graphics systems of the past, few will go for it, unless it is specifically powering simple devices. So there’s a lot riding on Fusion in the power computing stakes, but really, it won’t see the real light of day for at least a couple of years. But it’s a plan for the future, and AMD has already begun the march.
To counter criticism from Intel that chips with dozens of cores is instead the answer, Hester says that: “In this increasingly diverse x86 computing environment, simply adding more CPU cores to a baseline architecture will not be enough. As x86 scales from palmtops to petaFLOPS, modular processor designs leveraging both CPU and GPU compute capabilities will be essential in meeting the requirements of computing in 2008 and beyond.”
That’s because Intel plans an 80-core chip by 2009 or 2010. AMD says that its processor cores will be fine tuned to perform different tasks, and to perform them very well, whereas the inference is that Intel will just be pumping out processors that have very powerful sets of cores, but will expend too much energy and processing power to get all those cores to focus on specific tasks.
Why not just build each core in a processor to do one thing, and do it well? At least, that’s what AMD wants us to think about.
Intel’s Core 2 Quad launch is on in the first week of November here in Australia, where we’ll no doubt hear more about Intel’s plans to douse any excitement about AMD’s current and future plans, and we’ll be reporting on that next week.
Of course, Fusion isn’t just meant to be AMD’s version of the desktop processing platform of the future. AMD wants to use it everywhere, in “laptops, desktops, workstations and servers, as well as in consumer electronics and solutions tailored for the unique needs of emerging markets”.
Sounds just like Intel – they want their processors in all of those devices too, and indeed that already is the case for them. But it’s all about future orders. Whose processors sell in the largest quantities, whose processors are the most powerful for consumers and whose processors deliver the best profits.
It’s good news for we consumers. Competition almost always fosters a better performance from all concerned. The computing world does not need a complacent Intel, AMD or anyone else. The efforts of these companies are the future of computing as we currently see it. They can’t advance it fast enough!
So, while the ATI and AMD merger has been approved by the shareholders, and the ATI logo removed from the (now) AMD website, the future is a strong company with a real ability to continue competing against Intel and delivering new technology innovations, just as it did with the Opteron and Athlon 64.
The fusion of both companies is certainly not complete. It has only just begun.
Related:
Entry was posted
on Saturday, October 28th, 2006 at 7:05 am
under
Read more entries by Alex Zaharov-Reutt.
Stumble It!
