Sunday, November 22, 2009

Xlr 125 Maximum Speed

When three's a crowd: Sony, Microsoft, IBM ... GO

Although I was accused (and rightly) to focus too much on Apple, Microsoft and IBM, today I will tell a story that encompasses the three companies (along with Sony) and the truth is that I have found curious.

Much has been written and commented on the movement of Apple's developer conference in 2005 to switch to Intel and to abandon the PowerPC platform from which so proud they have always felt maqueros (I had a few months be, my first Mac was an iMac G5 of the former). The fact is that reading the book The Race For A New Game Machine going basic processor design that governs the PS3 Cell processor and PB that governs the Xbox 360, also discussed relations with Apple.

Actually, the design of this process could take a pill or two himself, but this time I will focus on a specific event that caught my eye at the same time I read it.

Anyway, I do a quick summary of the background story to get in and then write what I wanted to say. In 2001, Sony swept with the PS2, and then Ken Kuratagi, the father of the PSX and PS2, began working on the successor of the latter, which would the PS3 and had to be so superior to anything rejodidamente seen that effectively ending the competition.

Ken talked to people from IBM, which showed him the different families of processors in the PowerPC family had, but did not fit what I wanted for new Sony console. Basically, what I wanted Ken was all a supercomputer on a single chip, with a surprising vector processing capability and therefore capable of making the PS3 in the center of the digital home with not only games, but all kinds of content and services multimedia unsuspected until now. So IBM offered to build it from scratch but based on POWER4 processor (a version these were the G4 processor that Apple and some mounted on IBM minicomputers).

The grace of all that, but Toshiba and Sony were to finance with IBM throughout the development process, IBM was left with rights to trade freely with the resulting process.

In 2003, new customers knock on the door of IBM seeking a new processor for their new machines. Although there were several, the two most prominent and important were Apple and especially Microsoft, which was already thinking about their new console (which had not yet placed or name).

Although the project was top secret Cell processor, such as Bill Gates rejected all designed based on the POWER4 processor (for the same reasons that two years ago were rejected by Sony), IBM decided to try his luck and he spoke confidently of a new processor that was under development and prototypes that had not even yet, but milk was fucked up shit and piss and liked it so much that Gates said he wanted one of those to rule his new XBox.

Of course, Apple was very concerned, because a processor with a vector processing unit built-in series that the Macintosh would be the ideal computer for mathematical calculations, and especially image processing and video, as the chips which produced both Intel and AMD does not implement anything that could compete with this technology.

So IBM began working with the top three customers for the new processor separately and "secretly" trying to not become an inferno that the final processor take most of the different components that all three needed not have to test three different implementations of the processor but one, thus saving resources, by the way, IBM was not because barely managed to form an expert group capable of developing chips to form three different.

The fact is that, of course, the new additions to Sony that were made to the processor is not convinced, for I did not need your PS3 at all and they were financing the processor design. So did Microsoft and Apple, just that they were not only clients and co-designers.

So we got to where I wanted to go. The processor had to be ready, tested and manufactured in sufficient quantities to Sony and Microsoft (and presumably Apple) could come to the Christmas season of 2005 with its new products, but eventually threw them over and not enough progress made . Then you

transcribe and a half pages of the book which explains the contents of a meeting to discuss this problem. I was thinking about translating it but I prefer to copy as is to avoid my mistakes. The names I will leave, but I think they are important and what matters is what is being discussed and the conclusion we arrive and the implications these have. This is a follow up meeting between the main directors of Project:

Why was it so dificult for the three of us to walk in there and speak with one voice? Our core Had just too many cooks in the kitchen.

I Took my turn at the podium Then And Present the Expected performance of the PowerPC core. I based my conclusions on the performance software That the Rochester team maintained.

It didn't take Kahle long to show his han. "Schedule is oour top priority," he said, pounding a fist on the table for emphasis. "Not performance or frequency or power or area. What are the tradeoffs than can help reduce the schedule risk?" He walked to the whiteboard and with big bold strokes wrote: Remove Out-of-Order Processing.

I was still standing at the podium when Kahle turned away from the whiteboard and quickly scanned the room for support. He was obviously waiting for me to make an argument for keepong his high-value feature in the design. Mayve I should have, but I was so tired of fighting. I just wanted to move on.

I sighed and addressed the audience. I didn't see too many happy faces out there. "We've been trying to satisfy too many customers with one product, and now I worry that they will all be unhappy with the final product. I agreee that removing out-of-order processing will reduce the level of complexity -and therefore, risk- in the design, but it will also likely Apple to abandon the project. If you can live with that consequence, then go ahead."

Surprisingly, Akrout allowed Kahle to have this way. Maybe he thought he could still persuade his friends to Apple to accept a compromise. But as I predicted, Apple executives Jon Rubenstein and Bob Mansfield were furious with the design changes and the proposed revamping of their roadmap, claiming the lower PowerPC performance planned. They blamed Akrout for not protecting their interests. To IBM vice president John Kelly, it was all about economics. Though Apple was a treasured customer, the volume of chip business they pushed through the IBM foundry was not big in comparison to the volumes of game chips expected from Sony and Microsoft. Apple simply didn't have the clout to make Kelly reverse this decision. Besides, Sony and Microsoft were both contributing enormous amounts of money up front to the cost of the development of their chips, whereas IBM always footed the bill for the development of Apple's chip. The only money IBM got from Apple was for chip production.

Akrout was fiercely loyal to Apple. and it distressed him greatly to have fallen out their favor. He remained in his office with the door shut for long periods of time, and when he did come out, his characteristic dimpled smile was absent, his shoulders slumped, he was late to meetings, and he was noticeably distracted. He worked hard to quickly find and present other acceptable options to Apple's future desktop and laptop lines. With a great deal of angry opposition, Apple finally agreed to accept a different core designed by IBM's Server Group. In reality, they were at least temporarily backed into a corner and had no other choice, but I'm sure they inmediatly started pouring energy and money into scouting for a replacement solution. Akrout still retained ownership of IBM's business with Apple, but his relationship with his favourite customer was forever changed.

El resto, como suele decirse, es historia. NEXTSTEP, el sistema operativo Apple bought and developed into what is now known as MacOS X, it was multiplatform, and of course had an implementation for Intel processors, and Apple instead of eliminating this feature in the room kept just in case. Thus, to the contempt of IBM, Apple began looking for alternatives. Motorola discarded, the other manufacturer of PowerPC processors, this combination was obvious: Intel.

Thus, after a suspicious meeting with people from Intel Jobs for this to become a technological partner Pixar (I was surprised when I read a lot and even got angry. "Not even Apple itself is believed his own publicity their PowerMacs?), And a few months after Apple announced in a keynote Macs actually changed the PowerPC architecture to Intel.

0 comments:

Post a Comment