In the part 1 of this story 3 engineers appear, Masakazu Suzuoki, the chief at Sony Computer Entertainment microprocessor division and the architect of PS2, Takeshi Yamazaki, an ex-NEC supercomputer architect specialized in distributed shared memory and an architect at SCE microprocessor division, and Mitsuo Saito, the head of the Toshiba LSI development center and the project manager of Emotion Engine.
Here's my translation of the pages above.
The Engineers Who Created Cell: Part 1 Float the World on the Network
(c) Nikkei Electronics Magazine
The sunshine of a few days before almost like high summer gave way to an autumn breeze that comes from the nearby sea with a subtly salty flavor. September 18th, 1999. At the road from Kaihin Makuhari Station at JR Keiyo Line to Makuhari Messe, a current of people spill over. They are non-business-day visitors to Tokyo Game Show 99 Fall, rushing for the venue.
Most of them go to the SCE booth without thinking twice when they arrive at the show. Their target is the next-generation game console PlayStation 2 of which specification was announced just 5 days ago. The prototypes of PS2 SCE prepared for the show are 17 units. Most of them are used for the demonstration of new games in development for the PS2 release set for only half a year later.
The microprocessor 'Emotion Engine' runs at the 10 times faster clockspeed than the first PlayStation and has 6.2GLOPS floating-point operation performance that surpasses the latest PC of the day, while the graphics LSI 'Graphics Synthesizer' has as much as 4 MB DRAM embedded. What are those innovative games enabled by those 2 LSIs? Though those games are prototypes, you will be able to peek into the real performance of them. The SCE booth is packed with visitors who want to see it by their own eyes. The booth is as if it were covered locally by summer heat.
The Chip That Realizes A Dream
While the world is captivated by the finally unveiled PS2, Sony Computer Entertainment HQ at Aoyama, Tokyo has been ringing with a film that was just released in Japan at the same month. The source is, none other than the father of the PlayStation and PS2, SCE CEO, Ken Kutaragi.
"Hey Suzuoki, that movie is awesome. I was impressed."
"What?"
"Matrix, I mean."
Kutaragi calls after Masakazu Suzuoki who's come across him in the company, abruptly spouts about his feelings on the film. For Kutaragi, Suzuoki is a subordinate with the companionship of many years since the days they'd designed things like recording devices at Sony HQ. He could entrust Suzuoki with the development of PS2 since they are close to each other.
For Suzuoki, it is right after he could be relieved to see years of PS2 development had been able to reach the stage of unveiling the prototype. He nods at spouting Kutaragi as usual, with occasionally replies.
Suzuoki has been able to understand almost intuitively why Kutaragi is interested in this movie. It's been long since Kutaragi started to express his strong interest toward a virtual world formed on a network. In the story of Matrix where heroes fight each other for human dignity traversing computer generated worlds and reality, he must have witnessed a part of the near future computer society he draws in his mind. Actually, Kutaragi even preached this film not only those in SCE but also to people outside of SCE.
"The network itself will be a computer soon."
Kutaragi's world view, which he fluently preachs as an engineer rather than as a manager, may sound like just another romantic utopian story to those who are not in the know. But, in any event, Suzuoki is different. It's because he is already on the mission to research the specification of the microprocessor for the succesor of PS2 even though the release of PS2 is half a year away. Apparently Kutaragi's message is that he wants to incorporate a mechanism to build a certain "world" on a network into the microprocessor for the next-gen console. Since it's Kutaragi's pressing request, he has to do it. However...
"Hmm."
Suzuoki sighs when he's been back alone. As it's a too grand and intangible goal, he can't think of any guidance. It's too demanding to tackle this problem and complete a specification all alone.
"I have to make him involve with this, after all."
The face of a man who just joined SCE a half year ago has come across Suzuoki's mind. He is a unique engineer who after engaging in the design of a cutting-edge supercomputer at NEC went back to the graduate school saying he had something left to do, to immerse himself in a study of computer architecture and got a PhD just recently - his name is Takeshi Yamazaki.
The Man Led By An Invisible Thread
Yamazaki, who is a specialist in computers and at the same time a "gamer" everyone including himself recognizes, had been known in the gaming business since before he joined SCE. It was because he'd published his speculations about game console architectures, that amazed even engineers at game console manufacturers, on his web site in his grad school days. Among such speculations, there were paragraphs that looked like the prophecy of Emotion Engine and Graphics Synthesizer. Yamazaki was even the target of respect by the readers of the web site who were interested in state-of-the-art game consoles.
Just before the end of the doctoral course, Yamazaki applied for SCE thinking that he wanted to develop a microprocessor with a new architecture and SCE was the only place he'd be able to do it. It was in the fall of 1998.
He went to SCE at the appointed date to face the test and opened a heavy door of the interview room with anticipation and anxiety in his heart. In the room, Ken Kutaragi, the very man awaited Yamazaki with a smile on his face.
"I browse your website often "
Yamazaki was shocked by the very first words thrown from Kutaragi's mouth. He had had no idea at all that his hobby WWW site was well-know in the business. The interview went on smoothly and some days later Yamazaki could get the acceptance letter.
"It's nice if this guy comes to our company."
After joining SCE, Yamazaki knew that Kutaragi and Suzuoki had talked like that pointing at his web site just before the recruitment. For SCE, Yamazaki was a long-awaited talent of destiny.
It's in no way that Yamazaki doesn't accept Suzuoki's offer. As the result, the research about the microprocessor for the successor of PS2 has been secretly started with those two at the center, in the fall of 1999 before PS2 was released. At first it was non-periodical brainstorming held every week by 5-10 volunteers.
"What is the microprocessor of Network Age? How much performance does it need? What process technology can be used?"
Each debates about a theme set for every meeting. Through such discussions, the work to form the factors required for the microprocessor put on the successor of PS2 continued for a while.
Dream Again
In parallel with those at SCE, there was another group that was seeking the architecture of the microprocessor for the successor of PS2. They are 1 hour away from the SCE HQ by train, and grounded in Semiconductor Technology Center located at the place 5 minutes away by walk from the north gate of JR Kawasaki Station - engineers of Toshiba. Most of them are those who've been in the co-development of Emotion Engine until recently.
Through the development of Emotion Engine which runs at 300Mhz that was outstanding back then for a consumer-product microprocessor, engineers at Toshiba had realized they could accumulate many design know-hows required for bigger LSI development projects of the near future. "To solidify this confidence, we want to join the development for the successor of the PS2 again" - voices wishing like that had risen in Toshiba.
Mitsuo Saito, who had headed such engineers as the chief of the Toshiba System-LSI Technology Laboratory, had ordered some engineers to research the architecture of the microprocessor for the PS2 successor since the spring of 1999 when the development of Emotion Engine ended.
Saito and Kutaragi had been old acquaintances who met at computer-architecture conferences since they were engineers in the front line. The trust relationship built there evolved into the co-development of Emotion Engine. However, it doesn't ensure that Toshiba can get the appointment from SCE again for the PS2 successor. Saito has recognized very well the height of the hurdle he faces as he knows Kutaragi's persistence that rejects things which doesn't meet his standard as an engineer.
To work with SCE again, he has to present a convincing idea proactively. For that purpose Toshiba had been prepared the architecture called "Force System" in which many simple RISC cores are integrated as one chip. Its selling point is complement the clockspeed improvement that gets difficult as the semiconductor technology shrinks with parallel processing by multiple RISC cores.
The 3D graphics generated by Emotion Engine in PS2 amazed viewers with detailed images far beyond game consoles by then. For the successor of such a microprocessor, a mediocre thing is inexcusable. The Saito group sounded out about the adoption by visiting SCE with a 3D graphics movie generated by a simulation to show the performance of a Force System microprocessor.
Not only the architecture but the Saito group had envisaged also the organization of developers for the next-gen microprocessor. Toshiba put over 200 engineers in the Emotion Engine development. It's almost confirmed that they have to put more people in the next-gen microprocessor. It's not easy to gather such a large number of talented human resources.
What the Saito group considered is to build a development center at Silicon Valley which is the center of high-tech industries in the U.S. There's no other place in the world where you can secure as many first-string engineers. By sending engineers there also from Japan, it will be a perfect development team.
Actually the vaildity of this method was verified at the development of Emotion Engine. They designed the MIPS CPU core which is the heart of EE at Toshiba America Electronic Components, Inc. in Silicon Valley. Of course Kutaragi knows too the track record that the Japan-US mixed team could complete the 64-bit CPU core with 2 ALU to the schedule.
We did everything we could. The rest is only to wait for a reply from SCE. The Saito group has waited around nervously to see the spring of 2000.
Behind The Release Of PS2
March 4, 2000. The special day for SCE developers led by Kutaragi and Suzuoki, and for engineers at Toshiba like Saito, has finally come. The sales of PS2 has started at last.
"We ship 1 million units in 3 days"
The world was astonished by the hype Kutaragi laid out. But the reaction of consumers didn't betray the expectation. The reservation application that started 2 weeks ago of the release stopped 1 minute after the start. Hundreds of thousands of accesses brought down the server. Thousands of people lined up before big electoronics stores in the morning of March 4th. PS2 was sold like hotcakes at internet, at shops. 0.98 million units SCE prepared were sold out in 3 days. Even fuss about troubles of memory cards and DVD playback occurred.
When the PS2 fuss got settled, Saito got a call from an SCE representative.
"Hello, Mr. Saito? Sorry for our haste, but could you attend a meeting of tomorrow? Actually we'd like to have a talk, with IBM"
"What?"
Saito wondered about what he heard. There were talks with SCE about the microprocessor for the PS2 successor so far. But why does IBM show up now? For Saito who knew an anecdote of the previous year it felt very strange.
-- to be continued --
Well, so the above is the part 1.
The excerpt of the part 2 is like this.
The Engineers Who Created Cell (Part 2): IBMs Indignation
+ Why Saito wondered about IBM is because in 1999 he reported to SCE that making a partnership with IBM would be difficult as it seemed IBM as developers of server CPU and Toshiba as developer of consumer CPU were too different in their styles, after he visited the IBM HQ in the U.S.
+ Saito, Kutaragi and IBM representative met at a hotel in Roppongi near IBM Japan. It was Kutaragi's wish to invite IBM after all so Saito accepted it.
+ Later 3 companies had meetings to discuss the architecture of CELL. The target performance of the project was 1 TFLOPS. Toshiba proposed Force System that has many simple RISC cores and a main core as the controller. Jim Kahle, the POWER4 architect, from IBM proposed an architecture which has just multiple identical POWER4 cores. When a Toshiba engineer said maybe Force System doesn't need a main core, Kahle was greatly pissed off (thus the title of this chapter) as without a main core POWER has no role in the new architecture.
+ Meetings continued several months and Yamazaki of SCE was inlined toward the IBM plan and voted for it. But Kutaragi turned down it. Eventually Yamazaki and Kahle talked about the new architecture and agreed to coalesce the Toshiba plan and the IBM plan. Finally IBM proposed the new plan where a Power core is surrounded by multiple APUs. The backer of APU at IBM Austin was Peter Hofstee, one of the architects of the 1Ghz Power processor. It was adopted as the CELL architecture.
-- to be continued --
