[SOUND] I'd like to introduce Professor Uri Weiser, who is now a professor here at Technion. And Uri, together with my colleague Dadi Perlmutter pioneered the Pentium microprocessor, which was an innovation revolution for Intel. And the story about it is fascinating. Uri recently received a coveted award, the Eckert–Mauchly prize for computer architecture. So I'm going to leave the stage and let Dadi and Uri together chat, because together, these two revolutionaries as junior engineers at Intel managed to change the course of a large company and the course of computing and the world. So I'm going to turn the stage over to Dadi and Uri. >> One of the great examples of taking a completely different route than the rest of the company wants to go is the invention of the Pentium architecture back in the late 80s. I am having here with me Professor Uri Weiser, whom we hired from National Semiconductors in the late 80s with the mission that we received at the time to make a change in the 486 for the sake of IBM requirements. So tell us about your background from National, and the experience of moving from National Semiconductors to Intel. >> Yeah, I was hired at National in 1984. Before that, I was in RAFAEL and I was hired to do the architecture of the National 32000 family. After about six months, they gave me the responsibility for designing the processor NS32532. And I was the project manager for three years until the tapeout. It was the best microprocessor that was ever designed at that time. Much better than the 386. And it was sold in less than 100,000 units, in about ten years. Then I got the offer to move to Intel, in which, at that time, I read an article from Nick. And Nick was writing at the time that between breakfast and coffee break, Intel will produce more 386s than all the RISC vendor together in a year. I took the numbers from the Intel number that I knew, and I found that it's correct. Comparing to what happened in National, I understood that compatibility is the most important thing. >> I remember, at the time when I hired you, we were given the assignment to make a small change to the 486, while the rest of the company were moving very much after where the industry was going at that time, to what was called the RISC processor. Could you give us some, in the layman, what was the situation, what was happening in the industry, and what was your reaction to the, at that time, Intel strategy? >> Yeah, this period was called the CISC versus RISC war. The concept was, CISC was complex instruction set computing. And this is, the computer is built out of instruction set architecture. The instructions were pretty complex. While in IBM, as well as in Berkeley, they designed a processor that was based on reduced instruction set computer. That means that the instruction are very simple, and by combining several instruction you can get any function that you think about. So it was new. The x86 was already old. And everybody was looking for the new to save the world. So there was a huge war between CISC vendor and RISC. While everybody who was new in the market decided to go to RISC. Intel management at that time decided to check the RISC. We had at that time, Intel had at that time a RISC processor, the i860, the N10. It was a RISC processor. And they decided to provide this processor to Microsoft. And Microsoft, at that time, ported OS/2, the operating system OS/2, on the i860. And the image was that, let's go and build personal PCs, personal computers, to build out of the RISC processors. And that was the situation when I came. The order from Santa Clara, the request from Santa Clara, was to build what we called at that time 486.5. They called it P16 at the time. And the basic idea was to do a evolutionary step on x86. So that was, at that point, meaning Intel is moving away from x86 while starting to design a line of processor towards RISC computers. >> What amazes me in this story, that the big Intel with all these big marketing people understanding, did not understand what we, a team of three people in Intel Israel far away, understood. Which we talk about in the chapter of talking about the importance of the ecosystem. That if Intel to abandon x86, they would abandon the whole billions of dollars software compatibility that was already developed for x86. And Intel would have started even behind the rest of the RISC processors.