 |
||
|
||
 |
The Technological Future of AMD: An Interview with Giuseppe Amato
Paolo Corsini, Gabriel Ikram - 16 Jul 2007
|
|
|
||
|
||
Hardware Upgrade: Good morning Giuseppe, could you tell us what your position at AMD is?
Giuseppe Amato: I am AMD's EMEA Sales & Marketing Technical Director; I'm superintendent of all the technical specifications of AMD's x86 family of products and of AMD embedded solutions that are distributed on the European, Middle Eastern and African markets. Consequently, I’m also involved in the technological matters related to AMD GPUs and chipsets.
Hardware Upgrade: In recent weeks a rumor has spread that AMD, starting next year, will completely leave the manufacturing scene and instead rely on partners for chip fabrication. We have been rather skeptical of this rumor, so could you give us your point of view about it?
Giuseppe Amato: That rumor is the result of a misinterpretation of an announcement that our CEO, Hector Ruiz, made. He was saying that AMD will leave the care of manufacturing and production to partner foundries only for parts that are not an intellectual property, IP. This does not imply that all AMD manufacturing will be decentralized. Only standard technology parts will be outsourced for manufacturing to our partner factories, which will result in an economic benefit for us.
Hardware Upgrade: For us, the rumor came across as rather odd after seeing the large investments AMD has made in its two Dresden factories and the increase in productivity that has been a result of the two factories.
Giuseppe Amato: Dresden plays a vital role for our development. Inside these factories we have reached the technological level that we need to produce our most complex architectures. Just think of the SOI technology, Silicon On Insulator, as a model of technology that justifies the existence and exclusiveness of our factories.
Hardware Upgrade: On the other hand, AMD acquired ATI, which has been a fabless corporation for a long time. Your thoughts on this?
Giuseppe Amato: "Fabless" is in fact the manufacturing model that we will employ for the production of our GPU and chipset products. The fabless model is very widespread: just think of Nokia. It doesn't have any of its own foundries and so appoints the manufacturing of its architectures to external partners. However, the fabless model is also hard to handle. It is imperative to understand which parts will be outsourced to partners for production and, on the opposite end, which parts will be produced in our own manufacturing centers.
Hardware Upgrade: What do you think is the largest benefit AMD has received by its acquisition of ATI?
Giuseppe Amato: There are serveral benefits that AMD received from the ATI acquisition, but the most important is probably the following: AMD changed from a processor company to a platform company. We are now able to provide our customers with complete solutions, balancing the CPU, GPU, and logic management of the motherboard, while also still leaving full flexibility with our partners in case they intend to use other brands with our CPU or GPU products. AMD is an open corporation and believes in platform development. However, AMD platforms are not to be restricted to AMD products only.
Giuseppe Amato: As for Opteron solutions, we have already made public the release date which is in September. At that time we will launch the new processors of the Barcelona family. As for the Phenom processors for desktops, the first samples of Phenom FX CPUs will be available by the end of the year, while the remaining versions of the family will arrive by March 2008.
Hardware Upgrade: So the Barcelona processor will be launched in around two months. Can you tell us some more information about the features of the samples that are currently available?
Giuseppe Amato: At the present moment our main customers have Opteron Barcelona CPU samples, which are begin used for the validation process and internal research for compatibility with existing systems. These processors are not available for the press at this moment, but they will be by the end of August. We prefer not to spread information on operating clock speeds and performance as of right now, but I will say that the early versions of Barcelona processors will reach 2 GHz.
Hardware Upgrade: A very interesting topic in the past few months has been the employment of AMD’s latest generation of GPUs for parallel processing. NVIDIA is particularly active in this market sector thanks to its new TESLA family of products (we have already taken a brief look at GPGPU processing here), but we still fail to clearly see AMD’s approach to this new market.
Giuseppe Amato: Last fall we introduced a product family specifically designed for GPGPU processing. We are calling GPGPU processing on AMD products "stream computing.” In “stream computing" we use the R580 family of GPUs and adjust the boards for a GPGPU environment, making them different from traditional graphics cards typically used for 3D applications.
Hardware Upgrade: Can you tell us something about how AMD intends to make its moves in this sector with the new R600 architecture? Actually, so far no recent announcements have been made concerning GPGPU solutions that are specifically based on the R600 GPU.
Giuseppe Amato: We intend to keep offering GPU solutions that are specifically designed for parallel processing to be placed side by side with our processors. GPGPU products based on the R600 GPU, which uses a unified shader model, are obviously under production, but I cannot tell you more about them at the moment. The only thing that I can tell in advance is that AMD will soon announce more GPGPU news.
Hardware Upgrade: What do you think the future of the GPGPU processing field will be? If correctly used GPUs may add powerful processing power to existing platforms, but so far their use is limited by the intrinsic nature of the GPU which creates large programming hurdles.
Giuseppe Amato: At the moment we are really only at the dawn of GPGPU programming. Because of this, developers are facing extremely hard and complex work programming for the GPU. On the other hand, when correctly and efficiently used this new technology allows significant increases in performance. An example that we have shown that uses our previous video architectures is Tarari. By recompiling its antivirus scanner and using an AMD GPU, Tarari was able to reach significantly higher performance compared to what would have been obtainable using only a CPU.
In the future, GPUs will become more easily usable for general purpose processing, thanks to the increasing availability of compilers designed for this environment. After 2010 we can assume that there will be a real time compiler available that will choose what code to execute by checking which hardware resource is available, choosing from a CPU, GPU or specific accelerators. A dispatcher will check which resources are available and will compile the code in real time to be used with the processor that at that moment is able to provide the most efficiency.
ATI GPU programming for general purpose computing relies on Close To Metal, which is AMD’s first step in the direction of GPGPU programming. Eventually we will witness evolutions in our approach to GPGPU which will allow software developers to more easily use the full capacity of the GPU. Thanks to the availablilty of a higher number of registries General purpose GPU computing will also be made much easier in 2009 when Microsoft releases the DirectX 11 API.
Giuseppe Amato: During an Analyst Meeting in 2006 we pointed out some of the future trends of our products, one of which was the use of parallel GPUs. In fact, there will be new architectures that use both a CPU and GPU. In this architecture, processes that are more suited to the GPU will be assigned to the GPU. A large error that has been made regarding Fusion is that people are thinking that this type of architecture will only be a single chip package architecture, meaning both the GPU and GPU are to be integrated on the same die.
With Fusion we’ll be able to offer single chip and multichip products, which will also make it possible to have parallel GPUs in the same architecture. So far AMD has not made a final decision solidifying the future of Fusion, but this is a possible route we might take. If we consider existing low consumption CPUs and GPUs, like our Radeon HD 2400 GPU that has a TDP of about 30 Watts, it is likely that some future Fusion products may eventually integrate a CPU and GPU, connected through Crossfire technology, and still have a total TDP within 120 Watts.
Hardware Upgrade: What kinds of technologies do you expect AMD to use in the manufacturing of complex architectures like Fusion, where the GPU and CPU are going to be integrated on the same die?
Giuseppe Amato: After developing a native quad-core processor, AMD now possesses all of the technologies it needs to develop Fusion architectures. Whether it is a native solution with serveral cores integrated in the same die, similar to what we are using for Barcelona, or a multi-die package (author’s note: the same architecture used by Intel for its Core 2 Quad chips) composed of two separate silicon die installed on the same package, AMD is open to all technological evolutions that the market requires.
Hardware Upgrade: What type of role will Hypertransport technology play in these future architectures?
Giuseppe Amato: Hypertransport is only one element that will guarantee remarkable flexibility in the composition of these architectures. Our plans of Fusion will allow us to integrate a specific number of GPU and CPU cores depending on the customer and the uses for which they will use the chip. CPUs and GPUs are common IP Blocks that we can integrate and install on the products depending on the customer's needs.
AMD's goal is to provide products that specifically suit the needs of its customers. This is the reason why AMD should not simply be regarded as a conventional processor manufacturer, but as a system company that is able to produce and supply products for its customers that are complete solutions.
Hardware Upgrade: GPU manufacturers have been talking a lot about how other corporations can use GPGPU, but not much of how they use it themselves. In which ways has AMD used GPGPU to increase productivity? Can you give us any specific examples of how AMD is or will be using GPUs in its factories?
Giuseppe Amato: An actual use will be in wafer image recognition, a process that is usually carried out in the clean rooms of internal foundries. Once a processor wafer has been manufactured, it undergoes a check that uses extremely expensive and complex probing machinery to perform a visual check of the wafer. The test tells exactly which dies are correctly working and which are not. In the future, it is likely that GPUs will be used for these kinds of operations, which are extremely demanding in terms of processing power. The increased processing power of the GPU will allow more wafers to be produced each day, thus helping to reduce the bottleneck in chip production.
Copyright 1997 - 2012 Hardware Upgrade S.r.l