For PS3's creator, this isn't a game
By Yoshiko Hara
Ken Kutaragi, president and CEO of Sony Computer Entertainment Inc., proposed the underpinnings of the Cell processor to IBM engineers as part of a next-generation gaming platform that would fully realize the "computer" portion of computer entertainment. That system, the Cell-powered Playstation 3, was unveiled at the recent Electronic Entertainment Expo. In the wake of a management restructuring that may indicate Sony brass was not of one mind on the best path for the company, Kutaragi told EE Times that the PS3 will kick off the new "computer revolution" he has dreamed about since he proposed the first Playstation console.
EET: When did you first get the idea for the Cell processor?
Ken Kutaragi: It was immediately after we completed the development of the PS2. [The Playstation2 was introduced in 2000.] After we presented the papers on the Emotion Engine processor at the International Solid-State Circuits Conference and the main work proceeded to the building of the PS2 console, I started thinking about what would be next.
EET: Then the Cell processor was conceived for gaming?
Kutaragi: No, it is for computing-but I wanted to change the concept of computers. The name of our company is Sony Computer Entertainment; I wanted to merge computer technology and entertainment. It may be regarded as game applications for the time being, but I wanted to realize the day when "computer entertainment" would mean all such entertainment applications, including games.
From the engineers' side, what we are working on is computer technology. When we designed the Emotion Engine, for example, we strongly based it on computer technology and deployed a computer architecture for it.
The PS1 and PS2 were based on computer technology, but they were still transitional. At the next stage, I hoped to tackle computer technology from the start. I have harbored that hope since the 1980s.
Now, in the third phase, I have aimed at bringing about a computer revolution, teaming with the world's largest computer company and with Toshiba, a longtime partner since the days of the Playstation 2.
EET: IBM was in your mind from the beginning?
Kutaragi: Yes. From Eniac and the System/360 and 370, to the servers that followed them and the IBM PC-and even to HAL [the computer in the film 2001: A Space Odyssey]-all came from IBM.
But I couldn't just suddenly show up [at IBM's T.J. Watson Research Center], so first I knocked on the door of IBM Japan, asking whether there were any engineers who shared my dream. I got a warm, enthusiastic reaction from eminent researchers and architects at IBM, such as James Kahle. Through discussions with these people, the dreams and ambitions of each side became a common dream and ambition.
The drive to revolutionize computer technology is in IBM's DNA. So when I showed up, they probably thought, "This guy may come from a different field, but maybe his idea's not so bad . . ." something like that.
IBM was a teacher for us about what a "true" computer should be. But I believe there was some common ground about the future of computers -a vision that we pictured and that IBM drew.
We finally formed the development team in Austin [Texas], gathering top engineers from all the companies. IBM delegated fellow-class engineers to the project. Sony sent engineers who were involved in Playstation development and also delegated young engineers who would lead the next generation. Toshiba sent top engineers who were specialists in processor development. . . . Several tens of engineers gathered.
EET: Presentations on the Cell processor were made at this year's ISSCC. Were there any challenges to implementing the Cell in a game console?
Kutaragi: There have been plenty of challenges, everywhere. First was the Cell processor itself. We want the Playstation 3's arrival to be as explosive as the PSP [Playstation Portable] launch, so we have to establish a volume production system for the processor. How can we fabricate a large-scale, high-performance processor like the Cell at a pace of a million units a month in a quite-short time frame? It is madness to take that challenge, but we expect to pull it off.
[Pilot production of the Cell processor has already begun with an eye toward providing development kit chips for the Playstation 3, which is expected to hit the market next spring. Volume production details for the processor and the system, have not yet been divulged, however.-Ed.]
The Cell processor presentation at ISSCC was an academic treatise. The volume product will be different from the thesis, since we have to consider the various pros and cons in practical implementation.
The package warrants special consideration. The power consumption of the Cell processor is quite low for its high performance, but it consumes a large amount of power on an absolute scale. Handling large current, though the supply voltage is low, is a challenge for the power-circuit design.
Beyond the Cell processor, we have to consider the graphics processor that we are jointly developing with Nvidia, and the bridge chips as well. The Playstation 3 will use Blu-ray Disc media, but related technology and components are not available off the shelf. Everything is a challenge.
EET: Will game-title development also be a challenge for such a high-performance processor?
Kutaragi: It should be the reverse. I want it to be the reverse. The greater the limitations of the hardware, the more labor is needed for software development. With earlier hardware, direct access [from a game program to machine-language-level] hardware or some tricky programming was required to pull off the full performance.
The PS2, for example, made full use of the semiconductor technology available at the time, but only a software effort on the part of the game developers enabled high performance titles like Gran Turismo [a realistic car racing game that has sold more than 43 million copies worldwide]. The more closely they accessed the hardware, the higher the performance they achieved.
Cell is not like that. Application programs can no longer directly access the hardware; instead they will have to be written in high-level, object-oriented language. That was done for security reasons: If processors of high performance and wide bandwidth like the Cell were linked together without sufficient security, a worldwide system crash could occur with one attack.
The big feature of the processor is that multiple operating systems run on it. From the beginning, I wanted multiple operating systems to run on the processor simultaneously.
The Cell processor has a kernel called Level 0 at the bottom. This level is not disclosed and is kept secure. Level 1 handles operations close to the kernel, such as scheduling, the real-time kernel and device drivers. Level 2, which we call the guest OS layer, is for general-purpose operating systems such as Linux and PC OSes and operating systems for the Playstation. All operating systems and applications run on Level 2 or higher. Programmers can concentrate on their targeted area of concern without worrying about other operating systems.
EET: Game applications require real-time processing. Might the multilayered structure cause a problem for games?
Kutaragi: We are real-time game developers, so all necessary schemes-such as resource allocation, priority setting and preemptive control in a multithread environment-have been taken into consideration.
EET: The Cell processor targets general-purpose applications. Does it satisfy all the requirements for a game processor?
Kutaragi: All of those requirements have been incorporated; we did not compromise the processor.
The Cell processor will completely change the concept of programming. Conventionally, programming has sought to pull out the best performance from a closed processor and a certain capacity memory. The Cell processor assumes the existence of multiple Cells.
The model image for the Cell-based network may be the Internet: Servers around the world form one virtual "computer," and each PC accesses it. One Cell or cluster of Cells can also function as a server; but whereas the present Internet mainly handles characters, applications on the Cell network will also handle semantics and reasoning.
Though sold as a game console, what will in fact enter the home is a Cell-based computer.
People now have various media in their homes, but those media are not yet organized. The Cell handles those media intelligently. For example, it can compile images and video footage of you by recognizing your images from a vast accumulation [of images]. When used for security, it could use image recognition to identify the presence of a stranger in a monitored area.
The Cell enables enormous applications that require image processing and large-volume computing. Using the Cell, it would be easy to form a secure home network that would connect home audio and video devices.
EET: The Cell processor has been described as realizing the various possibilities of the electronic home. Does all of Sony accept your vision?
Kutaragi: We can pursue a lot of dreams based on the Cell, but it will take time. My mission [at Sony] is to concentrate on implementing the Cell in practical applications and propagating it.
I am sure that a technology revolution is about to occur, not only within Sony but throughout the digital consumer electronics industry.
An enormous demand for computing will be born in homes. A part of the Cell's computation power could be sold to, for example, information service providers and could be purchased when large processing power was necessary. This is already realized as a part of grid computing, but the Cell could realize it in much faster and more effective way. This would encourage the emergence of a new business.
EET: Many people in the industry had expected that you would lead Sony's semiconductor strategy.
Kutaragi: I am now in a position to use semiconductors. I made most of the decisions on investments, and developed the midterm strategy, based on semiconductors. Sony will surely continue heading in a semiconductor-focused direction.
If the semiconductor business is important [to a company], if there is not an immediate, large demand, the company cannot invest in semiconductors with confidence. And in most cases, there is no such demand. But we have the game business, and I will concentrate on that at Sony Computer Entertainment with confidence.
EET: Some people say that no next-generation game console can create as big a market as the Playstation 2 did.
Kutaragi: We heard the same doubts and criticism when we launched the PS1 and PS2. This time there will again be a tough battle.
But as you can see, Microsoft and Nintendo are also grappling with the game business. The direction is the same, and it's a good sign that we're not the only one showing interest in this direction.
http://www.eetasia.com/ARTICLES/2005JUN/C/2005JUN_INT_WK2.HTM
By Yoshiko Hara
Ken Kutaragi, president and CEO of Sony Computer Entertainment Inc., proposed the underpinnings of the Cell processor to IBM engineers as part of a next-generation gaming platform that would fully realize the "computer" portion of computer entertainment. That system, the Cell-powered Playstation 3, was unveiled at the recent Electronic Entertainment Expo. In the wake of a management restructuring that may indicate Sony brass was not of one mind on the best path for the company, Kutaragi told EE Times that the PS3 will kick off the new "computer revolution" he has dreamed about since he proposed the first Playstation console.
EET: When did you first get the idea for the Cell processor?
Ken Kutaragi: It was immediately after we completed the development of the PS2. [The Playstation2 was introduced in 2000.] After we presented the papers on the Emotion Engine processor at the International Solid-State Circuits Conference and the main work proceeded to the building of the PS2 console, I started thinking about what would be next.
EET: Then the Cell processor was conceived for gaming?
Kutaragi: No, it is for computing-but I wanted to change the concept of computers. The name of our company is Sony Computer Entertainment; I wanted to merge computer technology and entertainment. It may be regarded as game applications for the time being, but I wanted to realize the day when "computer entertainment" would mean all such entertainment applications, including games.
From the engineers' side, what we are working on is computer technology. When we designed the Emotion Engine, for example, we strongly based it on computer technology and deployed a computer architecture for it.
The PS1 and PS2 were based on computer technology, but they were still transitional. At the next stage, I hoped to tackle computer technology from the start. I have harbored that hope since the 1980s.
Now, in the third phase, I have aimed at bringing about a computer revolution, teaming with the world's largest computer company and with Toshiba, a longtime partner since the days of the Playstation 2.
EET: IBM was in your mind from the beginning?
Kutaragi: Yes. From Eniac and the System/360 and 370, to the servers that followed them and the IBM PC-and even to HAL [the computer in the film 2001: A Space Odyssey]-all came from IBM.
But I couldn't just suddenly show up [at IBM's T.J. Watson Research Center], so first I knocked on the door of IBM Japan, asking whether there were any engineers who shared my dream. I got a warm, enthusiastic reaction from eminent researchers and architects at IBM, such as James Kahle. Through discussions with these people, the dreams and ambitions of each side became a common dream and ambition.
The drive to revolutionize computer technology is in IBM's DNA. So when I showed up, they probably thought, "This guy may come from a different field, but maybe his idea's not so bad . . ." something like that.
IBM was a teacher for us about what a "true" computer should be. But I believe there was some common ground about the future of computers -a vision that we pictured and that IBM drew.
We finally formed the development team in Austin [Texas], gathering top engineers from all the companies. IBM delegated fellow-class engineers to the project. Sony sent engineers who were involved in Playstation development and also delegated young engineers who would lead the next generation. Toshiba sent top engineers who were specialists in processor development. . . . Several tens of engineers gathered.
EET: Presentations on the Cell processor were made at this year's ISSCC. Were there any challenges to implementing the Cell in a game console?
Kutaragi: There have been plenty of challenges, everywhere. First was the Cell processor itself. We want the Playstation 3's arrival to be as explosive as the PSP [Playstation Portable] launch, so we have to establish a volume production system for the processor. How can we fabricate a large-scale, high-performance processor like the Cell at a pace of a million units a month in a quite-short time frame? It is madness to take that challenge, but we expect to pull it off.
[Pilot production of the Cell processor has already begun with an eye toward providing development kit chips for the Playstation 3, which is expected to hit the market next spring. Volume production details for the processor and the system, have not yet been divulged, however.-Ed.]
The Cell processor presentation at ISSCC was an academic treatise. The volume product will be different from the thesis, since we have to consider the various pros and cons in practical implementation.
The package warrants special consideration. The power consumption of the Cell processor is quite low for its high performance, but it consumes a large amount of power on an absolute scale. Handling large current, though the supply voltage is low, is a challenge for the power-circuit design.
Beyond the Cell processor, we have to consider the graphics processor that we are jointly developing with Nvidia, and the bridge chips as well. The Playstation 3 will use Blu-ray Disc media, but related technology and components are not available off the shelf. Everything is a challenge.
EET: Will game-title development also be a challenge for such a high-performance processor?
Kutaragi: It should be the reverse. I want it to be the reverse. The greater the limitations of the hardware, the more labor is needed for software development. With earlier hardware, direct access [from a game program to machine-language-level] hardware or some tricky programming was required to pull off the full performance.
The PS2, for example, made full use of the semiconductor technology available at the time, but only a software effort on the part of the game developers enabled high performance titles like Gran Turismo [a realistic car racing game that has sold more than 43 million copies worldwide]. The more closely they accessed the hardware, the higher the performance they achieved.
Cell is not like that. Application programs can no longer directly access the hardware; instead they will have to be written in high-level, object-oriented language. That was done for security reasons: If processors of high performance and wide bandwidth like the Cell were linked together without sufficient security, a worldwide system crash could occur with one attack.
The big feature of the processor is that multiple operating systems run on it. From the beginning, I wanted multiple operating systems to run on the processor simultaneously.
The Cell processor has a kernel called Level 0 at the bottom. This level is not disclosed and is kept secure. Level 1 handles operations close to the kernel, such as scheduling, the real-time kernel and device drivers. Level 2, which we call the guest OS layer, is for general-purpose operating systems such as Linux and PC OSes and operating systems for the Playstation. All operating systems and applications run on Level 2 or higher. Programmers can concentrate on their targeted area of concern without worrying about other operating systems.
EET: Game applications require real-time processing. Might the multilayered structure cause a problem for games?
Kutaragi: We are real-time game developers, so all necessary schemes-such as resource allocation, priority setting and preemptive control in a multithread environment-have been taken into consideration.
EET: The Cell processor targets general-purpose applications. Does it satisfy all the requirements for a game processor?
Kutaragi: All of those requirements have been incorporated; we did not compromise the processor.
The Cell processor will completely change the concept of programming. Conventionally, programming has sought to pull out the best performance from a closed processor and a certain capacity memory. The Cell processor assumes the existence of multiple Cells.
The model image for the Cell-based network may be the Internet: Servers around the world form one virtual "computer," and each PC accesses it. One Cell or cluster of Cells can also function as a server; but whereas the present Internet mainly handles characters, applications on the Cell network will also handle semantics and reasoning.
Though sold as a game console, what will in fact enter the home is a Cell-based computer.
People now have various media in their homes, but those media are not yet organized. The Cell handles those media intelligently. For example, it can compile images and video footage of you by recognizing your images from a vast accumulation [of images]. When used for security, it could use image recognition to identify the presence of a stranger in a monitored area.
The Cell enables enormous applications that require image processing and large-volume computing. Using the Cell, it would be easy to form a secure home network that would connect home audio and video devices.
EET: The Cell processor has been described as realizing the various possibilities of the electronic home. Does all of Sony accept your vision?
Kutaragi: We can pursue a lot of dreams based on the Cell, but it will take time. My mission [at Sony] is to concentrate on implementing the Cell in practical applications and propagating it.
I am sure that a technology revolution is about to occur, not only within Sony but throughout the digital consumer electronics industry.
An enormous demand for computing will be born in homes. A part of the Cell's computation power could be sold to, for example, information service providers and could be purchased when large processing power was necessary. This is already realized as a part of grid computing, but the Cell could realize it in much faster and more effective way. This would encourage the emergence of a new business.
EET: Many people in the industry had expected that you would lead Sony's semiconductor strategy.
Kutaragi: I am now in a position to use semiconductors. I made most of the decisions on investments, and developed the midterm strategy, based on semiconductors. Sony will surely continue heading in a semiconductor-focused direction.
If the semiconductor business is important [to a company], if there is not an immediate, large demand, the company cannot invest in semiconductors with confidence. And in most cases, there is no such demand. But we have the game business, and I will concentrate on that at Sony Computer Entertainment with confidence.
EET: Some people say that no next-generation game console can create as big a market as the Playstation 2 did.
Kutaragi: We heard the same doubts and criticism when we launched the PS1 and PS2. This time there will again be a tough battle.
But as you can see, Microsoft and Nintendo are also grappling with the game business. The direction is the same, and it's a good sign that we're not the only one showing interest in this direction.
http://www.eetasia.com/ARTICLES/2005JUN/C/2005JUN_INT_WK2.HTM
