1 00:00:04,519 --> 00:00:12,840 Speaker 1: Technology with tech Stuff from stuff works dot com. Hey there, 2 00:00:12,880 --> 00:00:17,000 Speaker 1: and welcome to text Stuff. I am your host, Jonathan Strickland. 3 00:00:17,200 --> 00:00:21,000 Speaker 1: I'm a senior writer here with how stuff works dot com. 4 00:00:21,040 --> 00:00:24,760 Speaker 1: And today we're going to continue our story about Intel, 5 00:00:25,440 --> 00:00:29,880 Speaker 1: a very important company in the history of Silicon Valley 6 00:00:29,920 --> 00:00:33,760 Speaker 1: and electronics and computing in general. Now, in our last episode, 7 00:00:34,080 --> 00:00:37,600 Speaker 1: we really focused on the origins of Intel and traced 8 00:00:37,640 --> 00:00:40,080 Speaker 1: it all the way back to the birth of the 9 00:00:40,159 --> 00:00:45,720 Speaker 1: semiconductor industry itself. Two of the co founders of Intel, 10 00:00:45,920 --> 00:00:50,280 Speaker 1: Gordon Moore and uh and Rob Noyce. They had come 11 00:00:50,360 --> 00:00:56,480 Speaker 1: from a a an organization that didn't treat them very well, 12 00:00:56,640 --> 00:00:59,840 Speaker 1: and so they, along with six other people, left that organization. 13 00:00:59,880 --> 00:01:03,640 Speaker 1: They became known as the Traitorous Eight and founded the 14 00:01:03,840 --> 00:01:07,959 Speaker 1: fair Child Semiconductor business under the umbrella of a larger 15 00:01:08,040 --> 00:01:13,240 Speaker 1: company that did camera and instrumentation work. Worked there for 16 00:01:13,240 --> 00:01:17,160 Speaker 1: about eleven years before they left that company to co 17 00:01:17,360 --> 00:01:21,399 Speaker 1: found Intel. And in our last episode, I ended right 18 00:01:21,440 --> 00:01:25,520 Speaker 1: around the time that they had introduced their first micro processor, 19 00:01:25,680 --> 00:01:28,960 Speaker 1: the four zero zero four. Now we're gonna pick up 20 00:01:29,000 --> 00:01:31,039 Speaker 1: from that point forward. We're also going to backtrack a 21 00:01:31,040 --> 00:01:34,720 Speaker 1: little bit just to explain some other details, and we're 22 00:01:34,720 --> 00:01:38,320 Speaker 1: gonna explore what Intel was all about since its founding 23 00:01:38,440 --> 00:01:43,200 Speaker 1: up until present day. Now, the four zero zero four 24 00:01:43,280 --> 00:01:46,760 Speaker 1: started off as one of a set of four chips. 25 00:01:47,200 --> 00:01:50,400 Speaker 1: It's like a package of four chips that were designed 26 00:01:50,440 --> 00:01:53,880 Speaker 1: specifically for another company. That company was the Nipon Calculating 27 00:01:53,920 --> 00:01:58,320 Speaker 1: Machine Corporation. Now, the original designation for those four chips 28 00:01:58,480 --> 00:02:04,880 Speaker 1: was the MCS DASH four. Along with that four oh 29 00:02:04,960 --> 00:02:08,400 Speaker 1: four or four zero zero four, I should say chip, 30 00:02:08,840 --> 00:02:12,520 Speaker 1: were three others. Right. You had a chip that was 31 00:02:12,600 --> 00:02:17,480 Speaker 1: a read only memory chip or ROM chip that was 32 00:02:17,639 --> 00:02:21,800 Speaker 1: responsible for storing the custom applications for calculating machines. So 33 00:02:21,919 --> 00:02:24,720 Speaker 1: essentially the programs of these calculating machines were stored and 34 00:02:24,760 --> 00:02:26,919 Speaker 1: read only memory. That meant that you could not write 35 00:02:27,000 --> 00:02:30,359 Speaker 1: over that information. It was always going to be there. 36 00:02:30,360 --> 00:02:33,480 Speaker 1: It was hard coded onto the chip itself, and that 37 00:02:33,600 --> 00:02:36,000 Speaker 1: way it made it efficient and reliable and you didn't 38 00:02:36,000 --> 00:02:39,640 Speaker 1: have to worry about accidentally erasing it. There also was 39 00:02:39,720 --> 00:02:44,000 Speaker 1: a random access memory chip or RAM chip. Random access 40 00:02:44,040 --> 00:02:48,760 Speaker 1: memory allows you to store data in a temporary storage space, 41 00:02:49,040 --> 00:02:52,280 Speaker 1: so that way the processor can call upon that data 42 00:02:52,360 --> 00:02:57,080 Speaker 1: quickly without having to search a deeper storage UH solution. 43 00:02:57,880 --> 00:03:01,560 Speaker 1: The fourth chip was a registered chip, and that was 44 00:03:01,600 --> 00:03:06,080 Speaker 1: handling the input output port, so taking in the information 45 00:03:06,160 --> 00:03:09,639 Speaker 1: from the input devices and being able to translate that 46 00:03:09,720 --> 00:03:13,240 Speaker 1: into whatever the commands were for the CPU to process. Now, 47 00:03:13,240 --> 00:03:16,960 Speaker 1: it was the CPU that really ended up changing everything. 48 00:03:17,760 --> 00:03:21,480 Speaker 1: So Intel designed the CPU as sort of a general 49 00:03:21,520 --> 00:03:26,720 Speaker 1: purpose programmable processor. Now it wasn't a true general purpose 50 00:03:26,760 --> 00:03:29,160 Speaker 1: processor yet. That wouldn't come until a little bit later, 51 00:03:29,560 --> 00:03:34,920 Speaker 1: but it showed the promise of this strategy. It wasn't 52 00:03:35,000 --> 00:03:39,320 Speaker 1: locked into a single practical application. So up until the 53 00:03:39,360 --> 00:03:45,080 Speaker 1: early nineteen seventies, hardware for computing machines was heavily customized 54 00:03:45,160 --> 00:03:48,800 Speaker 1: for each specific machine. In other words, the chips you 55 00:03:48,800 --> 00:03:52,880 Speaker 1: would find in one type of computer, we're completely incompatible 56 00:03:53,200 --> 00:03:56,200 Speaker 1: with every other type of computer. To design the processor 57 00:03:56,320 --> 00:03:59,240 Speaker 1: for these machines, you essentially had to go back to 58 00:03:59,600 --> 00:04:03,160 Speaker 1: the big inning and rebuild the wheel every single time 59 00:04:03,200 --> 00:04:05,960 Speaker 1: you wanted to do it. Uh. There wasn't really a 60 00:04:06,000 --> 00:04:09,480 Speaker 1: concept of a general purpose processor the way the four 61 00:04:09,600 --> 00:04:12,840 Speaker 1: zero zero four was, and the concept of plug in 62 00:04:12,920 --> 00:04:17,160 Speaker 1: play would be decades further into the future. Intel's move 63 00:04:17,240 --> 00:04:21,159 Speaker 1: to create this programmable processor allowed for an explosion in 64 00:04:21,279 --> 00:04:25,799 Speaker 1: various uses. So the four zero zero four would provide 65 00:04:25,800 --> 00:04:28,960 Speaker 1: the foundation for Intel's future chips, which then would become 66 00:04:30,000 --> 00:04:34,280 Speaker 1: an instrumental component in countless types of technology, not just 67 00:04:34,400 --> 00:04:40,640 Speaker 1: computers or calculators. It virtually guaranteed Intel's success in the market. 68 00:04:41,560 --> 00:04:44,400 Speaker 1: Beyond that, the processor was able to take advantage of 69 00:04:44,400 --> 00:04:48,040 Speaker 1: the advances and manturization that had followed the invention of 70 00:04:48,080 --> 00:04:51,360 Speaker 1: the transistor. Now you'll remember from part one of this 71 00:04:51,520 --> 00:04:55,280 Speaker 1: series that Gordon Moore the one of the co founders 72 00:04:55,320 --> 00:04:58,360 Speaker 1: of Intel. He was the guy who came up with 73 00:04:58,440 --> 00:05:01,559 Speaker 1: what we now call Moore's law. All he had made 74 00:05:01,720 --> 00:05:05,159 Speaker 1: an observation that observation would become Moore's laud But the 75 00:05:05,200 --> 00:05:10,719 Speaker 1: observation said that the balance of technology, economics and manufacturing 76 00:05:10,760 --> 00:05:15,679 Speaker 1: processes would mean that for the foreseeable future. And remember 77 00:05:15,680 --> 00:05:18,479 Speaker 1: he made this production back in nineteen sixty five, the 78 00:05:18,560 --> 00:05:23,919 Speaker 1: number of discrete elements on microprocessors, on semiconductor chips would 79 00:05:23,960 --> 00:05:27,440 Speaker 1: double every two years or so. So if you were 80 00:05:27,480 --> 00:05:30,520 Speaker 1: to get a semiconductor chip in nineteen sixty seven, it 81 00:05:30,560 --> 00:05:33,000 Speaker 1: would have twice the number of transistors that you would 82 00:05:33,040 --> 00:05:36,039 Speaker 1: have found on a on a semiconductor in nineteen sixty five. 83 00:05:36,440 --> 00:05:39,000 Speaker 1: The ones in nineteen sixty nine would have twice as 84 00:05:39,040 --> 00:05:43,360 Speaker 1: many as the ones from nineteen sixty seven, and so on. Now, 85 00:05:43,400 --> 00:05:45,719 Speaker 1: by the time Intel was ready to produce the four 86 00:05:45,839 --> 00:05:49,000 Speaker 1: zero zero four chip, they could make a single CPU 87 00:05:49,120 --> 00:05:55,120 Speaker 1: microprocessor that was as powerful as the famous Eniac computer. Now, 88 00:05:55,120 --> 00:05:58,280 Speaker 1: the Eniac computer was one of the first electronic computers 89 00:05:58,640 --> 00:06:03,240 Speaker 1: ever in the history of mankind. Uh. It was constructed 90 00:06:03,520 --> 00:06:08,760 Speaker 1: in the forties, came online more or less in and 91 00:06:09,640 --> 00:06:12,679 Speaker 1: at the time it was one of it was really 92 00:06:12,720 --> 00:06:16,120 Speaker 1: the most powerful electronic computer in existence. It was one 93 00:06:16,160 --> 00:06:18,800 Speaker 1: of the first ones, so obviously it didn't have a 94 00:06:18,800 --> 00:06:23,400 Speaker 1: whole lot of competition, but it took up an entire room. 95 00:06:23,440 --> 00:06:26,160 Speaker 1: It had a lot of very large components that generated 96 00:06:26,160 --> 00:06:29,200 Speaker 1: a ton of heat. It used vacuum tubes instead of transistors. 97 00:06:29,839 --> 00:06:34,680 Speaker 1: But for this supercomputer of the early days, it would 98 00:06:34,720 --> 00:06:38,680 Speaker 1: take up an entire room. Well until's first microprocessor chip 99 00:06:38,760 --> 00:06:43,880 Speaker 1: had the equivalent amount of power stored on a chip 100 00:06:43,920 --> 00:06:46,880 Speaker 1: of semiconductor material that was the size of your fingernail. 101 00:06:47,520 --> 00:06:50,000 Speaker 1: So you went from a device that took up the 102 00:06:50,200 --> 00:06:52,720 Speaker 1: entire room in a building, and it was a big room. 103 00:06:52,760 --> 00:06:55,599 Speaker 1: It wasn't like a little office to something that could 104 00:06:55,640 --> 00:06:59,839 Speaker 1: fit on your fingernail with the equivalent amount of processing power. 105 00:07:00,080 --> 00:07:06,839 Speaker 1: This was a transformational moment in computer history, and one 106 00:07:06,839 --> 00:07:09,039 Speaker 1: of these days I'm doing I'll do a full episode 107 00:07:09,440 --> 00:07:12,520 Speaker 1: about the Eniac computer and also how it came to 108 00:07:12,600 --> 00:07:15,000 Speaker 1: be and the people who worked on it. It was 109 00:07:15,040 --> 00:07:18,520 Speaker 1: a fascinating device all on its own. In the decade 110 00:07:18,800 --> 00:07:22,560 Speaker 1: that it was in operation from around nineteen to nineteen 111 00:07:22,600 --> 00:07:27,920 Speaker 1: fifty five, researchers estimate that it ran more calculations during 112 00:07:27,960 --> 00:07:31,680 Speaker 1: that ten year period than all the calculations that had 113 00:07:31,720 --> 00:07:36,760 Speaker 1: been performed by humans leading up to that moment. So 114 00:07:36,880 --> 00:07:40,960 Speaker 1: in ten years, it did more calculations than all of 115 00:07:41,080 --> 00:07:45,600 Speaker 1: humans had done throughout history leading up to the creation 116 00:07:45,640 --> 00:07:52,840 Speaker 1: of the Eniac. That is incredible. But then it ended 117 00:07:52,920 --> 00:07:57,280 Speaker 1: up dying after it was struck by lightning. So it 118 00:07:57,320 --> 00:07:59,480 Speaker 1: has a tragic end to that story, which also means 119 00:07:59,480 --> 00:08:02,120 Speaker 1: it would make a great subject for a podcast in 120 00:08:02,160 --> 00:08:06,239 Speaker 1: the future. I think that maybe thour got a little 121 00:08:06,280 --> 00:08:09,400 Speaker 1: miffed that we humans that we're getting really confident, and 122 00:08:09,440 --> 00:08:11,200 Speaker 1: so he can kind of nipped in the bud. But 123 00:08:11,440 --> 00:08:15,360 Speaker 1: that's another story. Let's get back to Intel Now, if 124 00:08:15,400 --> 00:08:18,680 Speaker 1: you want to know what the clock speed was of 125 00:08:18,720 --> 00:08:21,960 Speaker 1: the four zero zero four, it's initial speed was one 126 00:08:22,200 --> 00:08:26,640 Speaker 1: d eight killer hurts. Now, clock speeds, in case you 127 00:08:26,640 --> 00:08:30,239 Speaker 1: aren't aware, they're measured in hurts, and it's a measure 128 00:08:30,280 --> 00:08:33,520 Speaker 1: of how many clock cycles a CPU can perform in 129 00:08:33,559 --> 00:08:37,360 Speaker 1: a second. So a hundred eight killer hurts clock speed 130 00:08:37,400 --> 00:08:42,160 Speaker 1: means that the processor can perform one eight thousand clock 131 00:08:42,280 --> 00:08:48,040 Speaker 1: cycles every second. In an ideal world, every single individual 132 00:08:48,160 --> 00:08:52,240 Speaker 1: instruction sent to a processor takes up one clock cycle. 133 00:08:52,760 --> 00:08:56,440 Speaker 1: So you could translate this as saying this processor was 134 00:08:56,480 --> 00:09:01,600 Speaker 1: capable of completing one hundred eight thousand inst ductions every second. 135 00:09:02,400 --> 00:09:05,920 Speaker 1: That's not exactly true because not all instructions take a 136 00:09:06,000 --> 00:09:10,720 Speaker 1: single clock cycle, and efficiency makes a big difference, but 137 00:09:11,120 --> 00:09:14,240 Speaker 1: it gives you a general idea of the capabilities of 138 00:09:14,240 --> 00:09:19,520 Speaker 1: this microprocessor. Now, really, when you get down to it, 139 00:09:19,600 --> 00:09:23,040 Speaker 1: everything that a processor does takes up a certain number 140 00:09:23,040 --> 00:09:26,439 Speaker 1: of clock cycles, and it depends on what the processor 141 00:09:26,559 --> 00:09:29,360 Speaker 1: needs to do, but it tells you there's a physical 142 00:09:29,520 --> 00:09:33,760 Speaker 1: limit to the number of tasks a processor can complete 143 00:09:34,080 --> 00:09:37,200 Speaker 1: within a given amount of time. So a second in 144 00:09:37,200 --> 00:09:42,079 Speaker 1: this case, if you are throwing stuff at the processor, 145 00:09:42,160 --> 00:09:45,720 Speaker 1: that takes fewer clock cycles than what it is capable 146 00:09:45,760 --> 00:09:48,880 Speaker 1: of doing. Things should run pretty smoothly, right. If the 147 00:09:48,960 --> 00:09:51,880 Speaker 1: number of instructions you're sending to the processor is less 148 00:09:51,880 --> 00:09:56,400 Speaker 1: than the processors clock speed, it should be a pretty 149 00:09:56,440 --> 00:10:00,720 Speaker 1: smooth experience for you on the user side of the computer. 150 00:10:01,160 --> 00:10:03,960 Speaker 1: But as you start to throw more processor hungry tasks 151 00:10:04,080 --> 00:10:06,800 Speaker 1: at the chip, you can start to slow down because 152 00:10:06,920 --> 00:10:11,120 Speaker 1: you might be throwing more directions per second than it 153 00:10:11,200 --> 00:10:13,840 Speaker 1: can handle, and then you start getting lag and things 154 00:10:14,160 --> 00:10:17,200 Speaker 1: get jittery and slow. There are a lot of other 155 00:10:17,200 --> 00:10:19,760 Speaker 1: factors that affect us as well, including how efficient that 156 00:10:19,840 --> 00:10:24,280 Speaker 1: processor is. Some processors are more efficient and can do 157 00:10:24,840 --> 00:10:30,280 Speaker 1: the same task with fewer clock cycles than a similar processor. 158 00:10:30,640 --> 00:10:34,240 Speaker 1: So if you get two different processors and they have 159 00:10:34,320 --> 00:10:37,640 Speaker 1: similar clock speeds, but one of them is designed and 160 00:10:37,760 --> 00:10:42,000 Speaker 1: optimized so that it's much more efficient, you can do 161 00:10:42,240 --> 00:10:45,040 Speaker 1: more with that processors than you can with the other one, 162 00:10:45,120 --> 00:10:48,280 Speaker 1: even if they're both rated at the same clock speed, 163 00:10:48,600 --> 00:10:52,920 Speaker 1: just because the other one handles tasks more efficiently than 164 00:10:53,040 --> 00:10:56,440 Speaker 1: it's uh than it's peer. This is true even if 165 00:10:56,480 --> 00:11:00,920 Speaker 1: you have two processors that have different clock speeds. If 166 00:11:00,920 --> 00:11:04,959 Speaker 1: you've got one processor that has a measurably faster clock 167 00:11:05,040 --> 00:11:08,640 Speaker 1: speed than your first one, but the first one is 168 00:11:08,679 --> 00:11:12,960 Speaker 1: still more efficient. You can still end up having a 169 00:11:12,960 --> 00:11:19,600 Speaker 1: better experience using the quote unquote slower microprocessor because it 170 00:11:19,720 --> 00:11:23,000 Speaker 1: is more efficient in its design. So that's something to 171 00:11:23,080 --> 00:11:27,200 Speaker 1: keep in mind if you're ever shopping for microprocessors. Just 172 00:11:27,360 --> 00:11:30,679 Speaker 1: looking for that clock speed is not really an indicator 173 00:11:30,760 --> 00:11:35,199 Speaker 1: of how good that microprocessor is. It is an indicator, 174 00:11:35,720 --> 00:11:37,640 Speaker 1: but it is not the only one you should pay 175 00:11:37,679 --> 00:11:40,199 Speaker 1: attention to. It's kind of like when you go shopping 176 00:11:40,240 --> 00:11:43,480 Speaker 1: for digital cameras and you start looking at megapixel numbers. 177 00:11:44,200 --> 00:11:49,440 Speaker 1: Larger megapixel numbers doesn't necessarily equate with better pictures. There 178 00:11:49,480 --> 00:11:52,880 Speaker 1: are a lot of other factors that are very important 179 00:11:52,920 --> 00:11:57,200 Speaker 1: when it comes to color, representation, contrast, all of these 180 00:11:57,240 --> 00:12:01,640 Speaker 1: other elements that go into creating digital images. So I 181 00:12:01,679 --> 00:12:03,559 Speaker 1: just want to bring that out here in this part 182 00:12:03,600 --> 00:12:06,040 Speaker 1: of the podcast, just in case you are looking at 183 00:12:06,440 --> 00:12:08,480 Speaker 1: building a computer and you want to find a really 184 00:12:08,480 --> 00:12:12,040 Speaker 1: good micro processor. Just going for that high number is 185 00:12:12,040 --> 00:12:14,280 Speaker 1: not a guarantee that you're getting the absolute best for 186 00:12:14,320 --> 00:12:17,080 Speaker 1: your money. You have to take these other elements into consideration. 187 00:12:18,000 --> 00:12:20,880 Speaker 1: Uh So, I know it's a bit of a tangent. 188 00:12:21,400 --> 00:12:26,080 Speaker 1: But it's important to remember just because otherwise we oversimplify 189 00:12:26,200 --> 00:12:31,239 Speaker 1: and say a higher clock speed equals a faster, better processor. 190 00:12:31,400 --> 00:12:38,040 Speaker 1: That's not always the case. Uh. Anyway, back in Killer 191 00:12:38,120 --> 00:12:43,520 Speaker 1: Hurts was wicked fast. That was a really fast processing time. Today, 192 00:12:43,640 --> 00:12:47,520 Speaker 1: something like Intel's Core I seven KB Lake processor can 193 00:12:47,600 --> 00:12:51,160 Speaker 1: run at four and a half giga Hurts. That means 194 00:12:51,240 --> 00:12:55,120 Speaker 1: it can run four and a half billion clock cycles 195 00:12:55,160 --> 00:12:57,600 Speaker 1: per second. It can handle four and a half billion 196 00:12:57,640 --> 00:13:00,840 Speaker 1: instructions per second. Compare that to the four zero zero 197 00:13:00,880 --> 00:13:03,520 Speaker 1: four that was a hundred eight thousand, four and a 198 00:13:03,520 --> 00:13:07,559 Speaker 1: half billion. You start to see the power of Moore's 199 00:13:07,640 --> 00:13:11,400 Speaker 1: law how that plays out over time, and being able 200 00:13:11,440 --> 00:13:16,760 Speaker 1: to go into the billions of instructions per second would 201 00:13:16,760 --> 00:13:19,840 Speaker 1: probably have amazed even Gordon Moore back in the day, 202 00:13:19,920 --> 00:13:23,360 Speaker 1: because when he was making this observation he didn't necessarily 203 00:13:23,440 --> 00:13:28,239 Speaker 1: think this would be sustainable indefinitely. He thought that eventually 204 00:13:28,240 --> 00:13:32,080 Speaker 1: we would run into some sort of fundamental limit as 205 00:13:32,120 --> 00:13:34,840 Speaker 1: to what we are able to accomplish, and at that 206 00:13:34,920 --> 00:13:38,680 Speaker 1: point the actual progression would break down and you would 207 00:13:38,720 --> 00:13:43,280 Speaker 1: no longer be able to have a processor that's effectively 208 00:13:43,320 --> 00:13:47,640 Speaker 1: twice as powerful two years into the future because we 209 00:13:47,679 --> 00:13:51,360 Speaker 1: would have bumped up against some sort of fundamental limit 210 00:13:51,520 --> 00:13:55,120 Speaker 1: that would prevent us from doing that. Now, for the 211 00:13:55,160 --> 00:13:58,560 Speaker 1: four zero zero four, Intel was using two inch wafers, 212 00:13:59,120 --> 00:14:04,079 Speaker 1: which is interesting because typically semiconductor companies were using twelve 213 00:14:04,120 --> 00:14:08,559 Speaker 1: inch wafers to design microchips, but Intel was using two 214 00:14:08,559 --> 00:14:11,439 Speaker 1: inch wafers for this one. And that might sound delicious. 215 00:14:11,480 --> 00:14:15,240 Speaker 1: You're hearing wafer, you might be thinking cookies. It's not. 216 00:14:15,600 --> 00:14:18,360 Speaker 1: The wafer is called that because it's a big circular disk, 217 00:14:18,760 --> 00:14:21,040 Speaker 1: or in the case of the four zero zero four, 218 00:14:21,080 --> 00:14:25,400 Speaker 1: they're using small circular discs. But a wafer is a substrate. 219 00:14:25,960 --> 00:14:30,760 Speaker 1: It's a foundation. It is the ground upon which you 220 00:14:30,800 --> 00:14:34,920 Speaker 1: build a processor or you build a circuit. It's made 221 00:14:34,920 --> 00:14:39,600 Speaker 1: of semiconductor material It's frequently silicon these days, it's more 222 00:14:39,640 --> 00:14:42,280 Speaker 1: often silicon than anything else, but there are other types 223 00:14:42,320 --> 00:14:45,560 Speaker 1: of semiconductor materials, so I don't want you thinking silicon 224 00:14:45,760 --> 00:14:50,160 Speaker 1: is the one and only type. There are others. Silicon 225 00:14:50,640 --> 00:14:55,760 Speaker 1: wafers are usually twelve inches in diameter, not two inches 226 00:14:55,800 --> 00:14:58,479 Speaker 1: in diameter, so this was a bit of a departure. 227 00:14:58,720 --> 00:15:01,640 Speaker 1: But you have to imagine in a really shiny disk 228 00:15:01,800 --> 00:15:04,440 Speaker 1: that has like a grid like pattern across it that's 229 00:15:04,440 --> 00:15:08,840 Speaker 1: repeated over and over and over again. Uh, the patterns 230 00:15:08,880 --> 00:15:11,800 Speaker 1: that you see, that's actually the patterns of where circuits 231 00:15:12,040 --> 00:15:15,120 Speaker 1: will be laid. So it's not just a design, that's 232 00:15:15,160 --> 00:15:19,080 Speaker 1: actually the physical architecture of where the circuits are going 233 00:15:19,120 --> 00:15:22,200 Speaker 1: to be. They don't just come out that way. They 234 00:15:22,240 --> 00:15:25,080 Speaker 1: have to be etched that way. When you first create 235 00:15:25,320 --> 00:15:28,320 Speaker 1: these silicon wafers and you and you get them polished, 236 00:15:28,600 --> 00:15:32,520 Speaker 1: they are just reflective. They don't have any grid like 237 00:15:32,600 --> 00:15:36,240 Speaker 1: patterns on them. Now, for the geometrically minded out there, 238 00:15:36,280 --> 00:15:38,840 Speaker 1: you might wonder, why the heck would you produce round 239 00:15:39,040 --> 00:15:44,640 Speaker 1: silicon wafers when microprocessors are rectangular. Now, obviously you can 240 00:15:44,800 --> 00:15:48,080 Speaker 1: build a bunch of microprocessors on top of a wafer 241 00:15:48,080 --> 00:15:51,040 Speaker 1: and then you cut them out. So you use your 242 00:15:51,080 --> 00:15:55,720 Speaker 1: wafer to be the the foundation for all of your microprocessors. 243 00:15:55,760 --> 00:15:58,800 Speaker 1: You you etch it out using a repeated pattern over 244 00:15:58,840 --> 00:16:02,640 Speaker 1: and over again. You build out the microprocessors, then you 245 00:16:02,680 --> 00:16:04,480 Speaker 1: cut them all out, and you get your little dies 246 00:16:04,840 --> 00:16:10,200 Speaker 1: of of CPU chips or other microchips, doesn't have to 247 00:16:10,240 --> 00:16:13,320 Speaker 1: be just ae CPU. Other microchips also followed this pattern, 248 00:16:14,040 --> 00:16:16,680 Speaker 1: but you're thinking, well, if it's round and ultimately the 249 00:16:16,760 --> 00:16:19,360 Speaker 1: chips are square, that means there's a lot of waste, right, Like, 250 00:16:19,440 --> 00:16:23,960 Speaker 1: you start to get where the edges of the the 251 00:16:24,040 --> 00:16:28,400 Speaker 1: individual microprocessors are coming up against the edge of the disk. Well, 252 00:16:28,440 --> 00:16:30,960 Speaker 1: square and circle, they're not going to match up perfectly, 253 00:16:30,960 --> 00:16:33,280 Speaker 1: so you're gonna have some wasted material. Why would you 254 00:16:33,320 --> 00:16:37,360 Speaker 1: go with a disc in the first place. Well, in 255 00:16:37,440 --> 00:16:39,160 Speaker 1: order to answer this question, I'm gonna have to talk 256 00:16:39,200 --> 00:16:42,000 Speaker 1: about how these chips are made today for the most part. 257 00:16:42,080 --> 00:16:45,200 Speaker 1: But remember that the process back when Intel made it's 258 00:16:45,240 --> 00:16:48,560 Speaker 1: four zero zero four chip was similar but not nearly 259 00:16:48,600 --> 00:16:53,160 Speaker 1: as sophisticated as the way we do it today. Um, 260 00:16:53,200 --> 00:16:58,600 Speaker 1: the microprocessor manufacturers cut wafers up to make several microprocessors 261 00:16:58,600 --> 00:17:03,640 Speaker 1: per wafer, but that does create some way. So here's 262 00:17:03,720 --> 00:17:09,560 Speaker 1: why we have round wafers instead of say, square wafers. 263 00:17:09,600 --> 00:17:15,480 Speaker 1: It's because we grow the wafers and it's done in 264 00:17:15,480 --> 00:17:18,240 Speaker 1: a very interesting way. So imagine that your goal is 265 00:17:18,280 --> 00:17:22,800 Speaker 1: to create a sheet of silicon that's perfect from a microprocessor. 266 00:17:23,040 --> 00:17:25,720 Speaker 1: That means you have to have an incredible amount of 267 00:17:26,200 --> 00:17:29,000 Speaker 1: pure silicon, or as close to pure silicon as you 268 00:17:29,040 --> 00:17:34,680 Speaker 1: can possibly manage. Any sort of impurity will introduce electrical 269 00:17:34,800 --> 00:17:39,080 Speaker 1: elements into your substrate that you don't want because that's 270 00:17:39,119 --> 00:17:41,200 Speaker 1: going to create errors. So you need it to be 271 00:17:41,240 --> 00:17:44,040 Speaker 1: as pure as you possibly can make it. You then 272 00:17:44,080 --> 00:17:47,040 Speaker 1: treat it to allow transistors and other components to transmit 273 00:17:47,080 --> 00:17:51,800 Speaker 1: electricity across the circuit without having it bleed through or 274 00:17:51,880 --> 00:17:55,960 Speaker 1: leach away the substrate. Because silicon is a semiconductor and 275 00:17:55,960 --> 00:17:58,760 Speaker 1: can either conduct or insulate based on properties, it is 276 00:17:58,840 --> 00:18:03,639 Speaker 1: ideal for this. And we start with sand. Sand has 277 00:18:03,680 --> 00:18:07,760 Speaker 1: a very high percentage of silicon in it, and again 278 00:18:07,840 --> 00:18:11,720 Speaker 1: silicon is our semiconductor material of choice. The sand gets 279 00:18:11,760 --> 00:18:14,800 Speaker 1: melted down and then you eventually end up with a 280 00:18:14,840 --> 00:18:19,360 Speaker 1: material called polly or poly silicon, and you use this 281 00:18:19,480 --> 00:18:22,960 Speaker 1: to um you separate out as much of the impurities 282 00:18:23,000 --> 00:18:27,800 Speaker 1: as you possibly can to you get to pure silicon. 283 00:18:28,480 --> 00:18:32,240 Speaker 1: You melt down ingots of this stuff into a purged 284 00:18:32,480 --> 00:18:35,760 Speaker 1: furnace at a temperature of more than two thousand, five 285 00:18:35,840 --> 00:18:39,640 Speaker 1: hundred degrees fahrenheit, which is more than one thousand, three 286 00:18:39,880 --> 00:18:45,760 Speaker 1: d seventy degrees c intigrade. The furnace must be purged 287 00:18:45,880 --> 00:18:50,480 Speaker 1: with oargon gas to make certain there's no unintended impurities present. 288 00:18:50,600 --> 00:18:52,720 Speaker 1: You want to make sure that you've gotten rid of 289 00:18:52,760 --> 00:18:56,919 Speaker 1: anything that could end up affecting the silicon as you 290 00:18:56,960 --> 00:18:59,920 Speaker 1: try to create a wafer. Now, at this point there's 291 00:19:00,000 --> 00:19:07,159 Speaker 1: about one non silicon atom per billion silicon atoms, so 292 00:19:07,240 --> 00:19:10,480 Speaker 1: that's incredibly pure. You get a billion silicon atoms and 293 00:19:10,480 --> 00:19:15,240 Speaker 1: then one thing that is not silicon. That's pretty pure. 294 00:19:17,119 --> 00:19:21,920 Speaker 1: You then have all of this molten silicon inside a crucible, 295 00:19:22,000 --> 00:19:27,399 Speaker 1: which is like a giant, uh cylindrical column, and it 296 00:19:27,520 --> 00:19:31,919 Speaker 1: keeps that molten silicon nice and hot. The crucible starts 297 00:19:31,960 --> 00:19:36,359 Speaker 1: to spin in a given direction. For the purposes of 298 00:19:36,400 --> 00:19:41,080 Speaker 1: this discussion, will say it's spinning witter shans, also known 299 00:19:41,119 --> 00:19:45,800 Speaker 1: as counter or anti clockwise, so it's spinning anti clockwise. 300 00:19:46,760 --> 00:19:52,199 Speaker 1: You then insert a seed crystal of silicon. It's about 301 00:19:52,280 --> 00:19:55,160 Speaker 1: the size and shape of a pencil, and you put 302 00:19:55,200 --> 00:19:59,880 Speaker 1: this it's lowered down by machine, into the molten sil 303 00:20:00,080 --> 00:20:03,879 Speaker 1: con and it acts as sort of the nucleaic spot 304 00:20:04,080 --> 00:20:06,679 Speaker 1: for other silicon crystals to form. And this way you 305 00:20:06,720 --> 00:20:14,240 Speaker 1: get a a very regular crystalline structure, a mono crystalline 306 00:20:14,280 --> 00:20:19,400 Speaker 1: structure of silicon, and this seed will turn in the 307 00:20:19,400 --> 00:20:22,480 Speaker 1: opposite direction of the crucible, So in our example, it 308 00:20:22,520 --> 00:20:28,400 Speaker 1: would turn clockwise while the crucible turns counterclockwise. Now, as 309 00:20:28,440 --> 00:20:35,320 Speaker 1: this happens, you end up creating this column, this cylindrical 310 00:20:36,040 --> 00:20:42,520 Speaker 1: column of pure silicon or mostly pure silicon. Uh, and 311 00:20:42,560 --> 00:20:45,679 Speaker 1: you get that monocrystalline structure all the way throughout the 312 00:20:45,880 --> 00:20:51,960 Speaker 1: entire thing. When you withdraw the cooled silicon from the crucible, 313 00:20:52,440 --> 00:20:56,120 Speaker 1: it looks like a giant silicon log that tapers at 314 00:20:56,119 --> 00:20:59,840 Speaker 1: an end because that's where you've been pulling the sea 315 00:21:00,160 --> 00:21:03,679 Speaker 1: crystal out, where you've been very slowly drawing it upwards. 316 00:21:03,720 --> 00:21:08,320 Speaker 1: You you draw it incredibly slowly, like a millimeter and 317 00:21:08,320 --> 00:21:12,399 Speaker 1: a half per minute, so it's a very very slow process. 318 00:21:12,480 --> 00:21:16,200 Speaker 1: You're not just whipping the crystal out of the molten silicon. 319 00:21:17,320 --> 00:21:20,400 Speaker 1: Doing this ends up creating, like I said, a tapered column. 320 00:21:20,960 --> 00:21:24,399 Speaker 1: And the silicon has great tent sile strength, meaning you 321 00:21:24,440 --> 00:21:29,600 Speaker 1: can suspend it from a thread like amount of silicon, 322 00:21:29,760 --> 00:21:35,119 Speaker 1: and even though it weighs between two and four pounds 323 00:21:35,240 --> 00:21:40,080 Speaker 1: or between a hundred and two hundreds, it'll hold it. However, 324 00:21:40,280 --> 00:21:43,080 Speaker 1: is very brittle, so while it has great tent sile strength, 325 00:21:43,160 --> 00:21:44,919 Speaker 1: if you were to just try and cut it. It 326 00:21:44,960 --> 00:21:50,120 Speaker 1: would cut very easily. The column or pole of silicon 327 00:21:50,560 --> 00:21:54,520 Speaker 1: would be about twelve inches or so in diameter, and 328 00:21:54,560 --> 00:21:57,040 Speaker 1: you would then use that to slice it into wafers. 329 00:21:57,080 --> 00:22:01,320 Speaker 1: So imagine this log of silicon and you put it 330 00:22:01,359 --> 00:22:06,280 Speaker 1: through a device that uses very thin wire cutters or 331 00:22:06,440 --> 00:22:09,480 Speaker 1: wire blades. I guess I should say they're not cutters, 332 00:22:09,480 --> 00:22:12,639 Speaker 1: they're they're blades made a very very thin wire that 333 00:22:12,800 --> 00:22:17,760 Speaker 1: then just zoom through this column and chop it up 334 00:22:17,800 --> 00:22:20,800 Speaker 1: into these very thin wafers. They're about a millimeter thick, 335 00:22:22,200 --> 00:22:25,840 Speaker 1: and you can get quite a few out of one 336 00:22:25,880 --> 00:22:30,400 Speaker 1: column of this silicon material. At that point, you send 337 00:22:30,440 --> 00:22:33,199 Speaker 1: them to another device to polish them out, because the 338 00:22:33,359 --> 00:22:38,919 Speaker 1: cutting creates uneven set sections on the surface of the wafers, 339 00:22:38,960 --> 00:22:41,120 Speaker 1: so you have to polish it to remove as much 340 00:22:41,160 --> 00:22:43,320 Speaker 1: of that unevenness as you can. But even that's not 341 00:22:43,400 --> 00:22:48,520 Speaker 1: good enough, because when you're building microprocessors, you're working on 342 00:22:48,600 --> 00:22:52,160 Speaker 1: the microns scale or smaller. These days, you're working on 343 00:22:52,560 --> 00:22:58,080 Speaker 1: the nano scale. At that scale, even the tiniest of 344 00:22:58,480 --> 00:23:02,040 Speaker 1: bumps or grooves is going to look like an enormous 345 00:23:02,040 --> 00:23:05,240 Speaker 1: mountain range or incredible valley. You have to buff it 346 00:23:05,280 --> 00:23:08,000 Speaker 1: all out and get it as level as you possibly can. 347 00:23:08,400 --> 00:23:11,560 Speaker 1: So after putting it through the polishing machine, you typically 348 00:23:11,560 --> 00:23:15,880 Speaker 1: would need to chemically treat the wafers to get them 349 00:23:15,920 --> 00:23:20,400 Speaker 1: as smooth as they possibly can be. Now, the next 350 00:23:20,440 --> 00:23:24,040 Speaker 1: step would involve etching the wafers to create the pattern 351 00:23:24,480 --> 00:23:29,320 Speaker 1: for your circuits. So this is sort of like building 352 00:23:29,359 --> 00:23:34,840 Speaker 1: the blueprint for the circuits directly onto the substrate itself. 353 00:23:35,640 --> 00:23:39,040 Speaker 1: Uh So, obviously, any sort of flaw in either the 354 00:23:39,119 --> 00:23:45,640 Speaker 1: silicon material or it's physical properties, whether it's the unevenness 355 00:23:45,880 --> 00:23:49,399 Speaker 1: or maybe there's an impurity that has fallen down and 356 00:23:49,480 --> 00:23:52,760 Speaker 1: touched the wafer, it's enormous when you get down to 357 00:23:52,800 --> 00:23:57,200 Speaker 1: that scale, so the tiniest of imperfections can completely ruin 358 00:23:57,640 --> 00:24:06,960 Speaker 1: a chip. For etching, Intel uses a light sensitive layer 359 00:24:07,400 --> 00:24:11,280 Speaker 1: called photo resist and coats the surface of the wafer. 360 00:24:12,240 --> 00:24:17,520 Speaker 1: That's essentially an etch resistant material. Anything that UH encounters 361 00:24:17,520 --> 00:24:21,200 Speaker 1: it is going to resist being etched. And they let 362 00:24:21,280 --> 00:24:25,040 Speaker 1: this layer harden and then they use other little stencil 363 00:24:25,240 --> 00:24:29,840 Speaker 1: like devices called masks. The masks cover parts of the 364 00:24:29,920 --> 00:24:32,920 Speaker 1: chip while allowing other parts of the chip, or rather 365 00:24:33,000 --> 00:24:35,760 Speaker 1: I should say cover parts of the wafer and allow 366 00:24:35,840 --> 00:24:40,000 Speaker 1: other parts of the wafer uh to be exposed. You 367 00:24:40,080 --> 00:24:44,960 Speaker 1: then use ultra violet light, which turns the photo resist 368 00:24:45,040 --> 00:24:50,720 Speaker 1: material it encounters soluble, and through the process of building 369 00:24:50,720 --> 00:24:52,760 Speaker 1: a circuit, you have to use lots and lots of 370 00:24:52,760 --> 00:24:56,320 Speaker 1: different masks, essentially, lots of different stencils, and lots of 371 00:24:56,359 --> 00:24:59,880 Speaker 1: applications of this ultra violet light. Because circuits are really 372 00:25:00,040 --> 00:25:03,760 Speaker 1: three dimensional creations. They're not just two dimensional. They're not 373 00:25:03,840 --> 00:25:08,000 Speaker 1: just width and and height. There's also depth to them. 374 00:25:08,080 --> 00:25:11,080 Speaker 1: So you use a sequence of these masks and you 375 00:25:11,720 --> 00:25:15,240 Speaker 1: expose the wafer to ultraviolet light. Each time the ultra 376 00:25:15,320 --> 00:25:19,840 Speaker 1: violet light hits through the mask and contacts the photo 377 00:25:19,960 --> 00:25:23,119 Speaker 1: resistive layer below, it makes it soluble. You then treat 378 00:25:23,520 --> 00:25:27,480 Speaker 1: the wafer with a chemical that removes all the soluble material. 379 00:25:28,000 --> 00:25:32,080 Speaker 1: So that's you're left with everything else. You've etched away, 380 00:25:32,119 --> 00:25:34,520 Speaker 1: all the stuff you do not want. All the stuff 381 00:25:34,560 --> 00:25:38,199 Speaker 1: you do want remains on the wafer. That is the 382 00:25:38,200 --> 00:25:43,960 Speaker 1: blueprint for your circuit. At that point, then you would 383 00:25:44,520 --> 00:25:47,639 Speaker 1: implant some ions, which are charged particles. They can have 384 00:25:47,680 --> 00:25:50,000 Speaker 1: either a positive or a negative charge, depending upon the 385 00:25:50,040 --> 00:25:52,320 Speaker 1: application you need them to be. You would either put 386 00:25:52,320 --> 00:25:55,240 Speaker 1: in positive ions or negative ions, but you use that 387 00:25:55,640 --> 00:25:58,359 Speaker 1: in the silicon itself. This is called doping. This is 388 00:25:58,480 --> 00:26:02,600 Speaker 1: used in semiconductors all the time to specifically dictate how 389 00:26:02,640 --> 00:26:08,640 Speaker 1: the semiconductor performs under specific circumstances. While you're etching, you're 390 00:26:08,640 --> 00:26:13,400 Speaker 1: creating channels for what will become transistors. The transistors themselves 391 00:26:13,520 --> 00:26:17,080 Speaker 1: must be deposited into the channels, and today Intel uses 392 00:26:17,160 --> 00:26:21,879 Speaker 1: a method called atomic layer deposition to apply materials to 393 00:26:21,920 --> 00:26:24,600 Speaker 1: the wafer surface at a level of precision necessary for 394 00:26:24,640 --> 00:26:28,800 Speaker 1: components on the nanoscale, because at that scale you're talking 395 00:26:28,880 --> 00:26:31,199 Speaker 1: about something so small you can't even see it with 396 00:26:31,280 --> 00:26:34,959 Speaker 1: an optical microscope. You would need a scanning electron microscope 397 00:26:35,000 --> 00:26:36,880 Speaker 1: or something similar in order to even get a look 398 00:26:36,920 --> 00:26:40,000 Speaker 1: at it, so obviously you have to have atomic precision 399 00:26:40,080 --> 00:26:45,520 Speaker 1: with this. Intel also uses electroplating to deposit copper ions 400 00:26:45,640 --> 00:26:48,320 Speaker 1: onto the transistor, and if you listen to my History 401 00:26:48,359 --> 00:26:52,040 Speaker 1: of Electricity episodes, you'll hear more about electroplating and how 402 00:26:52,080 --> 00:26:56,240 Speaker 1: that works. Now, since the individual components on microprocessors now 403 00:26:56,320 --> 00:27:00,440 Speaker 1: measureing the nanometers, it's critical that from this point forward, 404 00:27:00,880 --> 00:27:06,199 Speaker 1: all potential contaminants have to be eliminated from the fabrication area. 405 00:27:06,680 --> 00:27:09,360 Speaker 1: That means anyone working within the environment has to wear 406 00:27:09,400 --> 00:27:13,440 Speaker 1: a special suit often called a bunny suit, to eliminate 407 00:27:13,480 --> 00:27:18,200 Speaker 1: the possibility of dust, skin, or hair getting into the environment. Also, 408 00:27:18,480 --> 00:27:21,359 Speaker 1: they tend to have very powerful air conditioning systems which 409 00:27:21,359 --> 00:27:25,040 Speaker 1: are circulating and filtering the air on a very frequent basis. 410 00:27:25,400 --> 00:27:28,679 Speaker 1: Uh intel They're setup has air coming in from the 411 00:27:28,720 --> 00:27:31,399 Speaker 1: ceiling vents and the ceiling allow air to come down, 412 00:27:31,960 --> 00:27:35,359 Speaker 1: and then vents on the floor pull air away and 413 00:27:35,400 --> 00:27:38,840 Speaker 1: it does this constantly, so you're constantly circulating and filtering 414 00:27:38,880 --> 00:27:42,240 Speaker 1: the air to remove any potential pollutants that could ruin 415 00:27:42,800 --> 00:27:48,600 Speaker 1: a microprocessor. This is what gives us the term clean room, 416 00:27:48,640 --> 00:27:52,600 Speaker 1: and the clean rooms in semiconductor facilities tend to be 417 00:27:52,640 --> 00:27:56,960 Speaker 1: a Class one clean room, meaning there more free of 418 00:27:56,960 --> 00:28:01,800 Speaker 1: pollutants than even the most advanced hospitals. Are so extremely 419 00:28:01,920 --> 00:28:07,199 Speaker 1: clean environments, obviously, all the equipment that's being used has 420 00:28:07,240 --> 00:28:10,520 Speaker 1: to be absolutely spotless, because again, you introduce a tiny 421 00:28:10,560 --> 00:28:13,840 Speaker 1: impurity and you ruin a microchip, or worse yet, you 422 00:28:13,920 --> 00:28:16,919 Speaker 1: might ruin an entire wafer, which means all of the 423 00:28:16,920 --> 00:28:19,520 Speaker 1: microchips that would have been produced on that wafer are 424 00:28:19,560 --> 00:28:23,480 Speaker 1: a loss. You you'd have to throw them out. From 425 00:28:23,480 --> 00:28:27,600 Speaker 1: the beginning, Intel had to work on ways to design, miniaturize, 426 00:28:27,600 --> 00:28:30,919 Speaker 1: and imprint circuit layouts onto silicon, and this continues to 427 00:28:30,960 --> 00:28:34,639 Speaker 1: be an engineering challenge as companies like Intel attempt to 428 00:28:34,720 --> 00:28:37,960 Speaker 1: keep up with the observations Gordon Moore made decades ago. 429 00:28:38,760 --> 00:28:42,760 Speaker 1: The individual transistors are acting like switches, so they either 430 00:28:42,880 --> 00:28:46,560 Speaker 1: complete a circuit and allow electrons to move through, or 431 00:28:46,600 --> 00:28:49,520 Speaker 1: they break a circuit and they keep electrons from moving through. 432 00:28:49,600 --> 00:28:52,880 Speaker 1: And they're turned on and off by these devices called gates. 433 00:28:53,400 --> 00:28:57,000 Speaker 1: So a gates either open or it's closed, and that 434 00:28:57,080 --> 00:29:02,040 Speaker 1: tells you you know, essentially that the microprocessors. Ultimately their 435 00:29:02,120 --> 00:29:06,200 Speaker 1: job is traffic management. They're managing the movement of electrons 436 00:29:06,320 --> 00:29:11,479 Speaker 1: because they represent single pieces of information, either a zero 437 00:29:11,720 --> 00:29:14,040 Speaker 1: or a one and off or an on, a no, 438 00:29:14,440 --> 00:29:17,959 Speaker 1: or a yes. And collectively, when you get lots of 439 00:29:18,000 --> 00:29:20,800 Speaker 1: these dull pieces of information, you can describe much more 440 00:29:20,840 --> 00:29:25,680 Speaker 1: complex concepts than just on or off, and that's where 441 00:29:25,680 --> 00:29:31,000 Speaker 1: you get into the very basics of computer science. So 442 00:29:33,640 --> 00:29:38,760 Speaker 1: all of this, these different components have to be etched 443 00:29:38,920 --> 00:29:41,560 Speaker 1: onto the silicon, and that's because in the later stages 444 00:29:41,560 --> 00:29:44,720 Speaker 1: of manufacturing, the individual elements of the microprocessor have to 445 00:29:44,760 --> 00:29:46,800 Speaker 1: be deposited on the wafer. That can be up to 446 00:29:46,880 --> 00:29:52,880 Speaker 1: thirty layers of of material put onto a wafer before 447 00:29:52,880 --> 00:29:56,680 Speaker 1: it becomes a chip, and each incredibly tiny element needs 448 00:29:56,680 --> 00:29:58,880 Speaker 1: to be in its proper place. And these days, chip 449 00:29:58,920 --> 00:30:03,880 Speaker 1: manufacturers use variation of lithography to essentially print the components 450 00:30:04,440 --> 00:30:08,800 Speaker 1: onto the circuit etchings to build the microprocessor layer by layer. 451 00:30:08,880 --> 00:30:15,960 Speaker 1: It is incredibly precise and difficult to imagine. But we'll 452 00:30:15,960 --> 00:30:19,320 Speaker 1: talk more about that a little bit later on. First, 453 00:30:19,960 --> 00:30:29,400 Speaker 1: let's take a quick break to thank our sponsor. Back 454 00:30:29,400 --> 00:30:32,240 Speaker 1: to the four zero zero four. It had an impressive 455 00:30:32,440 --> 00:30:37,040 Speaker 1: number of transistors for the time. It had about two 456 00:30:37,080 --> 00:30:42,080 Speaker 1: thousand three transistors on this one chip. Uh, the micro 457 00:30:42,240 --> 00:30:46,360 Speaker 1: processors of today leave this processor behind in the dust. 458 00:30:46,400 --> 00:30:49,000 Speaker 1: That doesn't exist in those clean rooms I just talked about. So, 459 00:30:49,080 --> 00:30:53,680 Speaker 1: for example, the Broadwell E family of Intel processors, which 460 00:30:53,720 --> 00:30:57,800 Speaker 1: is actually from a couple of generations back well, they 461 00:30:58,000 --> 00:31:00,000 Speaker 1: launched in two thousand and sixteen and have a trans 462 00:31:00,000 --> 00:31:04,120 Speaker 1: sister account of about three point four billion. So you 463 00:31:04,160 --> 00:31:07,720 Speaker 1: went from two thousand, three hundred way back in nineteen 464 00:31:07,840 --> 00:31:11,640 Speaker 1: seventy one to three point four billion in two thousand, sixteen. 465 00:31:11,680 --> 00:31:15,280 Speaker 1: So yeah, Moore's law is no joke. While today's chips 466 00:31:15,280 --> 00:31:18,080 Speaker 1: have components that can measure just a smidge more than 467 00:31:18,120 --> 00:31:21,840 Speaker 1: a dozen nanometers in width, the four zero zero four 468 00:31:22,080 --> 00:31:28,240 Speaker 1: circuit line was ten microns wide, or ten thousand nanometers, 469 00:31:28,280 --> 00:31:31,800 Speaker 1: So the transistors of nineteen seventy one measured about ten 470 00:31:31,840 --> 00:31:35,720 Speaker 1: thousand nanometers wide. Today it's more like fourteen if you're 471 00:31:35,760 --> 00:31:41,720 Speaker 1: getting a top of the line processor, fourteen nanometers instead 472 00:31:41,720 --> 00:31:45,360 Speaker 1: of ten thousand. If you're wondering exactly how much that is, 473 00:31:45,400 --> 00:31:47,520 Speaker 1: because it's still kind of hard to imagine even like 474 00:31:47,640 --> 00:31:50,320 Speaker 1: ten thousand nanometers, how white is that, well, that's one 475 00:31:50,360 --> 00:31:54,440 Speaker 1: tenth the width of your typical human hair. Human hair 476 00:31:54,520 --> 00:31:58,840 Speaker 1: tends to be around a hundred thousand nanometers wide, not 477 00:31:59,000 --> 00:32:03,560 Speaker 1: long wide, or at least so I'm told I have 478 00:32:03,640 --> 00:32:06,320 Speaker 1: to take a lot about hair on faith these days. 479 00:32:07,560 --> 00:32:12,200 Speaker 1: I miss having hair. Well. The four zero zero four 480 00:32:12,280 --> 00:32:16,680 Speaker 1: launched and changed the world of electronics and computers. In 481 00:32:16,760 --> 00:32:20,280 Speaker 1: nineteen seventy two, Intel would expand, opening up an assembly 482 00:32:20,320 --> 00:32:26,040 Speaker 1: plant in Penang, Malaysia. This was Intel's first international manufacturing facility. 483 00:32:26,160 --> 00:32:30,520 Speaker 1: Intel also acquired a company called Microma, which was experimenting 484 00:32:30,520 --> 00:32:34,280 Speaker 1: with a new technology themselves. They were offering up digital 485 00:32:34,280 --> 00:32:38,160 Speaker 1: watches that had liquid crystal displays. This was pretty new 486 00:32:38,200 --> 00:32:42,160 Speaker 1: in the early seventies. So Intel got into the wearables 487 00:32:42,200 --> 00:32:45,080 Speaker 1: business just a couple of years after it was founded, 488 00:32:45,400 --> 00:32:48,080 Speaker 1: and people would think, oh, I thought Intel got into 489 00:32:48,120 --> 00:32:52,000 Speaker 1: the wearables business pretty recently with providing chips that could 490 00:32:52,000 --> 00:32:55,040 Speaker 1: be found in lots of different types of products out 491 00:32:55,040 --> 00:32:59,200 Speaker 1: there now. As it turns out, this early attempt to 492 00:32:59,200 --> 00:33:03,200 Speaker 1: get into wearables didn't really pan out for Intel. Uh. 493 00:33:03,240 --> 00:33:07,080 Speaker 1: Intel held onto Microma for about six years, but they 494 00:33:07,080 --> 00:33:11,720 Speaker 1: found that they had real trouble meeting consumer expectations and 495 00:33:11,760 --> 00:33:15,600 Speaker 1: figuring out exactly what consumers wanted. Because Intel was not 496 00:33:15,760 --> 00:33:21,880 Speaker 1: a consumer electronics company. Intel was building products for other businesses. 497 00:33:22,520 --> 00:33:26,440 Speaker 1: So Intel would create a processor that some other computer 498 00:33:26,520 --> 00:33:31,280 Speaker 1: manufacturer would use in its products. Intel wasn't building stuff 499 00:33:31,360 --> 00:33:36,680 Speaker 1: specifically for the end consumer, and as a result, they 500 00:33:36,760 --> 00:33:40,000 Speaker 1: weren't really good at running Microma as a business. They 501 00:33:40,040 --> 00:33:42,520 Speaker 1: held onto it for about six years and then they 502 00:33:42,560 --> 00:33:45,400 Speaker 1: sold it for a big loss. They lost about fifteen 503 00:33:45,440 --> 00:33:49,120 Speaker 1: million dollars on it. They sold it for less than 504 00:33:49,120 --> 00:33:51,840 Speaker 1: what they bought it, so it was a a tough 505 00:33:52,000 --> 00:33:55,880 Speaker 1: lesson for Intel in those early days. Jumping back to 506 00:33:56,640 --> 00:34:00,480 Speaker 1: two again, Intel created the first eight bit micro processor, 507 00:34:00,680 --> 00:34:04,560 Speaker 1: also known as the eight zero zero eight. Now, this 508 00:34:04,600 --> 00:34:07,360 Speaker 1: one obviously was more powerful than the four zero zero four, 509 00:34:07,400 --> 00:34:10,320 Speaker 1: but it still wouldn't really transform the world. It wouldn't 510 00:34:10,320 --> 00:34:13,600 Speaker 1: be until nineteen seventy four when the company introduced the 511 00:34:13,680 --> 00:34:16,719 Speaker 1: eight zero eight zero or the eight. This was the 512 00:34:16,760 --> 00:34:21,280 Speaker 1: first true general purpose microprocessor. So the four zero zero 513 00:34:21,320 --> 00:34:23,680 Speaker 1: four and the eight zero zero eight had laid the 514 00:34:23,719 --> 00:34:27,800 Speaker 1: groundwork for the eight, but it was this a D 515 00:34:27,960 --> 00:34:31,600 Speaker 1: eight that would become the basis for microprocessors in everything 516 00:34:31,719 --> 00:34:35,880 Speaker 1: from computers and calculators to traffic lights because it was 517 00:34:35,920 --> 00:34:39,279 Speaker 1: a true general purpose processor that could be used for 518 00:34:39,320 --> 00:34:42,360 Speaker 1: all sorts of different applications. So we often think of 519 00:34:42,440 --> 00:34:46,480 Speaker 1: Intel as the company that makes the processors in computers, 520 00:34:46,520 --> 00:34:49,319 Speaker 1: but the truth is they make processors that are in 521 00:34:49,560 --> 00:34:53,440 Speaker 1: all sorts of different gadgets and devices and products, not 522 00:34:53,600 --> 00:34:57,920 Speaker 1: just computers and phones. This made into one of the 523 00:34:57,960 --> 00:35:01,680 Speaker 1: most competitive companies in the space. They offered up a 524 00:35:01,760 --> 00:35:05,640 Speaker 1: computing solution in a small package for an aggressive price, 525 00:35:06,080 --> 00:35:09,560 Speaker 1: and as a result, Intel became the global leader in 526 00:35:09,600 --> 00:35:12,120 Speaker 1: microchips and held onto that title for a while. The 527 00:35:12,200 --> 00:35:15,240 Speaker 1: cost of the eight was just three d sixty dollars, 528 00:35:15,280 --> 00:35:18,360 Speaker 1: which you know they were. Intel was saying, this is 529 00:35:18,360 --> 00:35:21,240 Speaker 1: a computer on a chip. You get all the power 530 00:35:21,480 --> 00:35:24,279 Speaker 1: of a computer on a single chip that can be 531 00:35:24,320 --> 00:35:29,480 Speaker 1: incorporated into lots of different applications, and it's three sixty bucks, 532 00:35:29,480 --> 00:35:33,120 Speaker 1: which was much cheaper than full computer systems. So Intel 533 00:35:33,200 --> 00:35:36,600 Speaker 1: was doing really, really well in these days. It was 534 00:35:36,640 --> 00:35:40,000 Speaker 1: also dominant in the area of memory chips at the time, 535 00:35:40,360 --> 00:35:43,279 Speaker 1: so remember Intel had started by making chips that were 536 00:35:43,440 --> 00:35:47,080 Speaker 1: for memory, not for processing. By the end of nineteen 537 00:35:47,080 --> 00:35:50,719 Speaker 1: seventy four, Intel held eighty two point nine percent of 538 00:35:50,760 --> 00:35:54,719 Speaker 1: the d RAM chip market. Now this would change over 539 00:35:54,760 --> 00:35:57,440 Speaker 1: the next decade because other companies would get involved in 540 00:35:57,560 --> 00:36:01,440 Speaker 1: making memory chips and the competition got really fierce. And 541 00:36:01,480 --> 00:36:04,440 Speaker 1: once you get a ton of different companies all competing 542 00:36:04,440 --> 00:36:07,480 Speaker 1: with each other making the same stuff, you'll see that 543 00:36:07,560 --> 00:36:10,279 Speaker 1: product your your market share will start to drop over time, 544 00:36:10,360 --> 00:36:15,799 Speaker 1: unless you're making crazy deals by undercutting your competition. So 545 00:36:16,000 --> 00:36:20,880 Speaker 1: by a decade later, Intel's sharing the d RAM market 546 00:36:20,880 --> 00:36:22,839 Speaker 1: would have dropped all the way down to one point 547 00:36:22,920 --> 00:36:26,120 Speaker 1: three per cent. But while that memory chip business had 548 00:36:26,160 --> 00:36:30,760 Speaker 1: become incredibly competitive. Intel was still dominant in the microprocessor market, 549 00:36:31,440 --> 00:36:35,000 Speaker 1: so it wasn't as big a deal even though their 550 00:36:35,040 --> 00:36:41,160 Speaker 1: memory chip business was leaking away over time, or at 551 00:36:41,200 --> 00:36:44,719 Speaker 1: least was getting uh was being less dominant in the 552 00:36:44,760 --> 00:36:47,000 Speaker 1: market over time to the point where they had dropped 553 00:36:47,000 --> 00:36:50,920 Speaker 1: down to one point three within a decade. Because they 554 00:36:50,920 --> 00:36:55,160 Speaker 1: were the definitive name in microprocessors and more than balanced out, 555 00:36:55,760 --> 00:36:59,040 Speaker 1: they were able to recapture a lot of that success 556 00:36:59,080 --> 00:37:03,600 Speaker 1: with the microprocesss. In nineteen Robert Neiss, who was again 557 00:37:03,600 --> 00:37:06,120 Speaker 1: while the co founders, became the chairman of the board 558 00:37:06,160 --> 00:37:09,400 Speaker 1: at Intel. Noise was known as an executive who issued 559 00:37:09,520 --> 00:37:13,319 Speaker 1: the lavish trappings that many CEOs and chairpersons have indulged in. 560 00:37:14,280 --> 00:37:17,360 Speaker 1: He remained in leadership positions at Intel until his retirement, 561 00:37:17,840 --> 00:37:21,240 Speaker 1: which must not have suited him very well because after 562 00:37:21,239 --> 00:37:23,920 Speaker 1: he retired, he then went on to become the leader 563 00:37:23,960 --> 00:37:29,160 Speaker 1: of a semiconductor manufacturing consortium called Sema Tech. And Noise 564 00:37:29,200 --> 00:37:31,960 Speaker 1: would pass away in nineteen nine at the age of 565 00:37:32,000 --> 00:37:35,400 Speaker 1: sixty two. His leadership style would remain a very powerful 566 00:37:35,440 --> 00:37:39,120 Speaker 1: influence at Intel. It became sort of the model of 567 00:37:39,320 --> 00:37:44,040 Speaker 1: how leaders were expected to behave over at Intel, largely 568 00:37:44,080 --> 00:37:46,200 Speaker 1: that you were supposed to have kind of a let's 569 00:37:46,200 --> 00:37:49,799 Speaker 1: get to work sort of attitude and not have too 570 00:37:49,800 --> 00:37:54,640 Speaker 1: many lavish trappings of the executive's life. Uh, the idea 571 00:37:54,719 --> 00:37:59,279 Speaker 1: being that everyone should benefit, not just whomever's at the top. UM. 572 00:37:59,520 --> 00:38:03,040 Speaker 1: I don't know if that's how Intel's culture is now. 573 00:38:03,200 --> 00:38:06,920 Speaker 1: I haven't ever been at Intel's corporate headquarters, but I 574 00:38:06,960 --> 00:38:11,040 Speaker 1: know for a long time his values specifically were upheld 575 00:38:11,080 --> 00:38:15,640 Speaker 1: as the model for future Intel executives. Now, as you 576 00:38:15,719 --> 00:38:19,319 Speaker 1: might imagine, Intel's line of processors became more powerful with 577 00:38:19,400 --> 00:38:22,680 Speaker 1: every generation and stuck pretty close to Gordon Moore's predictions. 578 00:38:23,120 --> 00:38:26,600 Speaker 1: The processor's power tended to double every two years or so. 579 00:38:27,040 --> 00:38:32,000 Speaker 1: In nine, Intel introduced the sixteen bit eight eight six 580 00:38:33,200 --> 00:38:37,880 Speaker 1: eight six could work on to eight bit instruction sets simultaneously, 581 00:38:38,000 --> 00:38:41,280 Speaker 1: so it had a parallel processing component in it. However, 582 00:38:41,520 --> 00:38:45,239 Speaker 1: at the time, most software relied on eight bit processors 583 00:38:45,400 --> 00:38:49,120 Speaker 1: and couldn't take advantage of this sixteen bit capability, so 584 00:38:49,160 --> 00:38:52,440 Speaker 1: it's almost like it was future proofed that the software 585 00:38:52,480 --> 00:38:57,360 Speaker 1: that was out wasn't really able to enjoy the benefit 586 00:38:57,440 --> 00:39:01,560 Speaker 1: of the sixteen bit processor at the time. However, there's 587 00:39:01,600 --> 00:39:06,240 Speaker 1: a generally held belief that's been proven true multiple times 588 00:39:06,239 --> 00:39:10,360 Speaker 1: that if you build a really super strong processor, it 589 00:39:10,440 --> 00:39:12,759 Speaker 1: does not take long before someone figures out a way 590 00:39:12,800 --> 00:39:15,239 Speaker 1: to take advantage of that to build the software that 591 00:39:15,440 --> 00:39:18,279 Speaker 1: makes the best use or at least takes up a 592 00:39:18,320 --> 00:39:23,080 Speaker 1: great deal of that new super fast processors abilities. Uh. 593 00:39:23,160 --> 00:39:24,920 Speaker 1: In fact, you can get to a point where software 594 00:39:24,920 --> 00:39:31,320 Speaker 1: bloat outpaces processor performance. So it feels like generation over generation, 595 00:39:31,360 --> 00:39:33,879 Speaker 1: your processors are getting slower, But that's not really the case. 596 00:39:33,920 --> 00:39:37,520 Speaker 1: It's just that software is getting more bloated. Sometimes it's 597 00:39:37,560 --> 00:39:42,040 Speaker 1: a combination of the two. Actually. Well, by this time, 598 00:39:42,160 --> 00:39:45,680 Speaker 1: Intel was starting to face competition with Motorola, which had 599 00:39:45,719 --> 00:39:48,600 Speaker 1: introduced its own line of microprocessors in a line known 600 00:39:48,640 --> 00:39:51,680 Speaker 1: as the sixty eight thousand. This was the type of 601 00:39:51,719 --> 00:39:55,160 Speaker 1: microprocessors that would end up being used in Apple products 602 00:39:55,160 --> 00:39:59,080 Speaker 1: for a very long time. Motorola was the big uh 603 00:39:59,600 --> 00:40:03,320 Speaker 1: provide leader of micro processors for Apple. For many years, 604 00:40:04,840 --> 00:40:08,160 Speaker 1: there was a race to see which microprocessor architecture would 605 00:40:08,280 --> 00:40:11,000 Speaker 1: end up becoming the industry standard, which one is going 606 00:40:11,040 --> 00:40:14,680 Speaker 1: to be used in personal computers the most, and a big, 607 00:40:14,760 --> 00:40:19,759 Speaker 1: juicy contract helped assure that space for Intel. The contract 608 00:40:19,800 --> 00:40:24,239 Speaker 1: came from International Business Machines Corporation, which is better known 609 00:40:24,280 --> 00:40:28,600 Speaker 1: as IBM. IBM chose the eight bit processor from Intel, 610 00:40:28,640 --> 00:40:32,080 Speaker 1: called the eight eight, to be the processor of choice 611 00:40:32,080 --> 00:40:36,320 Speaker 1: in the official IBM personal computer. Line eight was based 612 00:40:36,320 --> 00:40:40,719 Speaker 1: off the design of the eighty six, the sixteen bit processor, 613 00:40:41,560 --> 00:40:45,160 Speaker 1: but the eight was an eight bit processor, so essentially 614 00:40:45,160 --> 00:40:48,960 Speaker 1: the eight had half of the address bus disabled to 615 00:40:49,080 --> 00:40:52,200 Speaker 1: make it an eight bit processor. Eventually, the pairing of 616 00:40:52,280 --> 00:40:55,640 Speaker 1: Intel chips with the Windows operating system a few years 617 00:40:55,800 --> 00:40:59,640 Speaker 1: down the line would lead to the nickname Windtell for 618 00:40:59,760 --> 00:41:03,440 Speaker 1: I b M Compatible machines, and it just showed that 619 00:41:03,520 --> 00:41:08,600 Speaker 1: Intel was considered a standard part of the IBM compatible universe. 620 00:41:09,160 --> 00:41:12,280 Speaker 1: You had an Intel processor and you had the Windows 621 00:41:12,320 --> 00:41:15,720 Speaker 1: operating system. If you didn't have those things, people didn't 622 00:41:15,719 --> 00:41:18,120 Speaker 1: really think of it as a PC. They thought of 623 00:41:18,160 --> 00:41:20,920 Speaker 1: it as something else, uh, which is a little weird, 624 00:41:21,239 --> 00:41:24,320 Speaker 1: but that just goes to show that, you know, Intel 625 00:41:24,360 --> 00:41:28,480 Speaker 1: had made some very savvy moves to become indispensable in 626 00:41:28,520 --> 00:41:31,480 Speaker 1: the personal computing industry, which ended up taking off like 627 00:41:31,520 --> 00:41:35,520 Speaker 1: gangbusters in the eighties. Now, around this time, Intel also 628 00:41:35,600 --> 00:41:39,759 Speaker 1: introduced a new innovation in computing called coprocessors. Now, these 629 00:41:39,760 --> 00:41:43,080 Speaker 1: were chips designed to take on certain tasks that typically 630 00:41:43,600 --> 00:41:47,160 Speaker 1: would go to the computer's CPU. But by freeing up 631 00:41:47,200 --> 00:41:50,959 Speaker 1: the CPU from doing those little tasks, they could run 632 00:41:51,120 --> 00:41:56,200 Speaker 1: software more efficiently. They could concentrate on the harder computational problems, 633 00:41:56,440 --> 00:41:59,640 Speaker 1: and these coprocessors would free up the CPUs by taking 634 00:41:59,640 --> 00:42:03,560 Speaker 1: on those more mundane tasks because they were very specialized 635 00:42:03,640 --> 00:42:06,960 Speaker 1: microprocessor chips and they were very efficient at handling very 636 00:42:07,000 --> 00:42:11,480 Speaker 1: specific types of computer problems. This kept the chips efficient 637 00:42:11,560 --> 00:42:13,839 Speaker 1: and kept their price down, and it also opened up 638 00:42:13,840 --> 00:42:17,480 Speaker 1: a new market for Intel to dominate. The company was 639 00:42:17,520 --> 00:42:20,760 Speaker 1: really growing rapidly at this time. Back when it was founded, 640 00:42:20,800 --> 00:42:26,280 Speaker 1: Intel had a grand total of twelve employees, but by 641 00:42:26,280 --> 00:42:30,319 Speaker 1: that number had grown to more than fifteen thousand employees. 642 00:42:30,880 --> 00:42:33,960 Speaker 1: Intel's founders tried hard to avoid falling into the same 643 00:42:34,000 --> 00:42:37,480 Speaker 1: traps they had encountered over at fair Child Semiconductor. They 644 00:42:37,480 --> 00:42:41,360 Speaker 1: were really striving to have open communication between executives and 645 00:42:41,440 --> 00:42:43,760 Speaker 1: everybody else. They wanted to make sure that their plans 646 00:42:43,760 --> 00:42:46,520 Speaker 1: were transparent, that everyone knew which way that the company 647 00:42:46,560 --> 00:42:50,560 Speaker 1: was going, and that people's concerns were listened to. Because 648 00:42:50,880 --> 00:42:54,040 Speaker 1: the founders of Intel had come from an environment where 649 00:42:54,080 --> 00:42:56,319 Speaker 1: they felt like they weren't being listened to, and they 650 00:42:56,400 --> 00:42:58,759 Speaker 1: didn't want Intel to turn into the same sort of thing. 651 00:42:59,560 --> 00:43:02,120 Speaker 1: But I've see that becomes a challenge when your workforce 652 00:43:02,239 --> 00:43:06,080 Speaker 1: numbers in the thousands of employees. So it was not 653 00:43:06,200 --> 00:43:10,640 Speaker 1: an easy thing for Intel to try and do. Now, 654 00:43:10,680 --> 00:43:13,160 Speaker 1: if you remember my podcast about the story of Macintosh, 655 00:43:13,239 --> 00:43:16,360 Speaker 1: you know I don't like to focus on every single 656 00:43:16,480 --> 00:43:20,359 Speaker 1: product that a company offers, because it just bogs down 657 00:43:20,440 --> 00:43:24,000 Speaker 1: the show. With Intel, it's particularly important. I don't do 658 00:43:24,040 --> 00:43:27,600 Speaker 1: that because the company has has had dozens of different 659 00:43:27,640 --> 00:43:31,400 Speaker 1: microprocessors in its history and hundreds of other types of products. 660 00:43:31,760 --> 00:43:34,120 Speaker 1: So we're gonna focus on a few important highlights rather 661 00:43:34,160 --> 00:43:36,440 Speaker 1: than a rundown of all the chips. Now, I don't 662 00:43:36,480 --> 00:43:38,719 Speaker 1: want to be here for a week going over an 663 00:43:38,719 --> 00:43:42,920 Speaker 1: impressive but exhaustive list of specs. I do want to 664 00:43:42,920 --> 00:43:47,200 Speaker 1: talk about some of the important early processors, though, because 665 00:43:47,360 --> 00:43:51,520 Speaker 1: I think it's it shows Intel strategy and also some 666 00:43:51,600 --> 00:43:53,880 Speaker 1: of the mistakes the company has made over its years. 667 00:43:55,719 --> 00:43:58,040 Speaker 1: The first one I would like to talk about is 668 00:43:58,080 --> 00:44:03,080 Speaker 1: the eight one six, which followed a sixteen bit architecture. 669 00:44:03,160 --> 00:44:07,720 Speaker 1: Is based off that eight six line from before starting 670 00:44:07,760 --> 00:44:11,800 Speaker 1: with the until began to incorporate components in the micro 671 00:44:11,960 --> 00:44:16,680 Speaker 1: processor that would normally be independent on a computer's motherboard. So, 672 00:44:16,719 --> 00:44:19,359 Speaker 1: in other words, Intel was looking at different parts of 673 00:44:19,360 --> 00:44:21,400 Speaker 1: the mother board and saying, well, we can lump this 674 00:44:21,480 --> 00:44:26,000 Speaker 1: in with the CPU. We can increase the response of 675 00:44:26,040 --> 00:44:29,520 Speaker 1: the CPU, we can make it more efficient, we can 676 00:44:29,600 --> 00:44:33,160 Speaker 1: decrease latency, and as a result, the entire computing experience 677 00:44:33,200 --> 00:44:37,720 Speaker 1: goes faster. This does not necessarily mean the CPU itself 678 00:44:37,840 --> 00:44:42,320 Speaker 1: is clocking faster. It's just again making that architecture more efficient. 679 00:44:43,160 --> 00:44:49,200 Speaker 1: So that was Intel strategy. The integration made sense. It 680 00:44:49,280 --> 00:44:52,800 Speaker 1: made the CPUs more efficient, made the computers much more fast. 681 00:44:53,120 --> 00:44:55,839 Speaker 1: And later that same year, Intel would release the eight 682 00:44:56,000 --> 00:44:58,600 Speaker 1: two eighties six, and at this point we would just 683 00:44:58,640 --> 00:45:01,440 Speaker 1: call it the two eight six. We would also end 684 00:45:01,520 --> 00:45:04,719 Speaker 1: up calling the next two computers the three six and 685 00:45:04,760 --> 00:45:07,720 Speaker 1: the four eight six. They were all from the same 686 00:45:07,880 --> 00:45:12,400 Speaker 1: kind of family of microprocessors. There were obviously changes and 687 00:45:12,480 --> 00:45:16,719 Speaker 1: redesigns with each generation, but they were all based off 688 00:45:16,719 --> 00:45:24,879 Speaker 1: that same architecture. In Intel Creative It's first reduced instruction 689 00:45:25,120 --> 00:45:31,160 Speaker 1: set computer or RISK microprocessor. These are more specialized processors 690 00:45:31,160 --> 00:45:35,280 Speaker 1: that concentrate on a relatively narrow band of computer instructions, 691 00:45:35,280 --> 00:45:38,319 Speaker 1: which means it can go really fast as long as 692 00:45:38,360 --> 00:45:41,280 Speaker 1: the instructions sent to it are within that narrow band. 693 00:45:41,320 --> 00:45:46,640 Speaker 1: It's kind of like a bureaucracy. Bureaucracies can handle forms 694 00:45:46,680 --> 00:45:50,160 Speaker 1: that are filled out properly very well. That's what a 695 00:45:50,200 --> 00:45:53,440 Speaker 1: bureaucracy does. But if you have a case that is 696 00:45:53,480 --> 00:45:59,719 Speaker 1: outside of the normal line, bureaucracies are terrible because they're 697 00:45:59,719 --> 00:46:03,960 Speaker 1: not signed to process stuff that's outside the norm. If 698 00:46:04,000 --> 00:46:07,040 Speaker 1: they can process something that's inside the norm, it should 699 00:46:07,040 --> 00:46:09,480 Speaker 1: go very fast. Same sort of thing is true with 700 00:46:09,600 --> 00:46:14,359 Speaker 1: risk microprocessors, except we're talking about computational problems, not licenses 701 00:46:14,400 --> 00:46:17,520 Speaker 1: and other issues that we would have when we encounter bureaucracies. 702 00:46:19,040 --> 00:46:22,480 Speaker 1: The first person to theore rise a risk microprocessor was 703 00:46:22,560 --> 00:46:26,239 Speaker 1: John Coke of IBM Research in nineteen seventy four, and 704 00:46:26,320 --> 00:46:29,880 Speaker 1: David Patterson, who taught at the University of California at Berkeley, 705 00:46:30,120 --> 00:46:33,040 Speaker 1: has the credit of coming up with the term. Intel's 706 00:46:33,160 --> 00:46:37,759 Speaker 1: I nine sixty risk microprocessors and its descendants would find 707 00:46:37,880 --> 00:46:44,359 Speaker 1: use in many applications, including military systems. In the three 708 00:46:44,760 --> 00:46:48,240 Speaker 1: six debuted, and it was the first CPU Intel designed 709 00:46:48,239 --> 00:46:53,600 Speaker 1: to be fully backwards compatible with previous CPUs. This was 710 00:46:53,640 --> 00:46:57,080 Speaker 1: a design decision that was carried forward into future microprocessors, 711 00:46:57,080 --> 00:46:59,640 Speaker 1: and it was also Intel's first thirty two bit x 712 00:46:59,680 --> 00:47:02,239 Speaker 1: a D six processor. It could support up to four 713 00:47:02,320 --> 00:47:06,520 Speaker 1: gigabytes of system memory, which at the time was a 714 00:47:06,680 --> 00:47:10,960 Speaker 1: truly enormous amount. Nobody really expected to use for gigabytes 715 00:47:11,000 --> 00:47:16,000 Speaker 1: of memory back in the mid eighties, but the architecture 716 00:47:16,080 --> 00:47:19,640 Speaker 1: texture was capable of supporting it. Now, that's not to 717 00:47:19,680 --> 00:47:22,720 Speaker 1: say that everything Intel touch turned into gold at this time. 718 00:47:23,080 --> 00:47:25,719 Speaker 1: The company also designed a processor called the i A 719 00:47:25,880 --> 00:47:29,600 Speaker 1: p X thirty two, and as you would guess by 720 00:47:29,640 --> 00:47:32,359 Speaker 1: the name and number, this did not follow the same 721 00:47:32,480 --> 00:47:36,319 Speaker 1: architecture as the X eight six chips did. This was 722 00:47:36,440 --> 00:47:40,520 Speaker 1: meant to expand its product line and differentiate microprocessors by 723 00:47:40,560 --> 00:47:44,520 Speaker 1: following a brand new architecture, but the design process did 724 00:47:44,560 --> 00:47:48,120 Speaker 1: not go smoothly. There were lots of flaws that were 725 00:47:48,120 --> 00:47:51,320 Speaker 1: introduced into the final processor design, and it was a 726 00:47:51,400 --> 00:47:54,480 Speaker 1: much more complicated design than the eight six line, and 727 00:47:54,520 --> 00:47:58,799 Speaker 1: it wasn't particularly efficient. So ultimately Intel would end up 728 00:47:58,800 --> 00:48:02,600 Speaker 1: shelving the product and saying that this was not a success. 729 00:48:03,520 --> 00:48:06,120 Speaker 1: Later in another attempt to break away from the x 730 00:48:06,160 --> 00:48:08,759 Speaker 1: A D six processor line, Intel would introduce the I 731 00:48:09,040 --> 00:48:14,240 Speaker 1: eight sixty Risk. While the early Risk microprocessors weren't intended 732 00:48:14,239 --> 00:48:17,040 Speaker 1: to go into personal computers, this was not true of 733 00:48:17,040 --> 00:48:20,840 Speaker 1: the I eight sixty. Unfortunately, like the I A p 734 00:48:21,160 --> 00:48:24,759 Speaker 1: X four thirty two, the microprocessor was flawed. It would 735 00:48:24,800 --> 00:48:27,960 Speaker 1: stall out when the processor encountered computational problems that were 736 00:48:27,960 --> 00:48:31,560 Speaker 1: outside of scope. So again, it's like that bureaucracy. When 737 00:48:31,880 --> 00:48:36,160 Speaker 1: a form doesn't have the right capability of capturing the 738 00:48:36,200 --> 00:48:38,480 Speaker 1: information you need, and you go to the bureaucracy, you're 739 00:48:38,480 --> 00:48:40,719 Speaker 1: gonna get the run around, and it's gonna be painful 740 00:48:40,760 --> 00:48:44,880 Speaker 1: and slow and laborious. Same thing was true with this processor. 741 00:48:45,000 --> 00:48:48,880 Speaker 1: It could not handle all computational problems efficiently, and that 742 00:48:49,280 --> 00:48:53,000 Speaker 1: eventually led to it not being a particularly powerful or 743 00:48:53,080 --> 00:48:58,520 Speaker 1: popular UH product in Intel's line. But the four six 744 00:48:58,600 --> 00:49:00,960 Speaker 1: was a different story that the first x A D 745 00:49:01,080 --> 00:49:05,399 Speaker 1: six CPU to incorporate an element called the functional processing unit, 746 00:49:05,560 --> 00:49:09,280 Speaker 1: or FPU, which until that time had been a separate 747 00:49:09,320 --> 00:49:14,480 Speaker 1: component from the CPU. This decreased latency between the FPU 748 00:49:14,680 --> 00:49:18,759 Speaker 1: and CPU, which translated again to faster computing speeds. The 749 00:49:18,840 --> 00:49:21,720 Speaker 1: first of these were running at clock speeds of fifty 750 00:49:21,800 --> 00:49:26,600 Speaker 1: mega hurts or fifty million clock cycles per second. Starting 751 00:49:26,600 --> 00:49:31,240 Speaker 1: a n Intel's ad campaign, Push made the company's slogan 752 00:49:31,400 --> 00:49:36,279 Speaker 1: Intel inside a common phrase in computer circles, into on 753 00:49:36,400 --> 00:49:40,840 Speaker 1: a really clever way of convincing computer companies to include 754 00:49:40,880 --> 00:49:45,160 Speaker 1: Intel chips and to use this phrase Intel inside, both 755 00:49:45,200 --> 00:49:48,840 Speaker 1: in the marketing strategy and on the computers themselves. The 756 00:49:48,880 --> 00:49:52,960 Speaker 1: way Intel convinced companies to do this as they said, Hey, 757 00:49:53,000 --> 00:49:55,640 Speaker 1: if you include our chips and you include the phrase 758 00:49:55,760 --> 00:50:00,440 Speaker 1: Intel inside in your marketing, we will pay for half 759 00:50:00,440 --> 00:50:04,960 Speaker 1: of your marketing costs. So they would shoulder half of 760 00:50:05,040 --> 00:50:09,400 Speaker 1: the marketing budget of these companies for print and television ads. 761 00:50:10,000 --> 00:50:13,239 Speaker 1: That's an enormous amount of money and Intel spent millions 762 00:50:13,239 --> 00:50:16,279 Speaker 1: of dollars doing this. But as a result, they were 763 00:50:16,320 --> 00:50:22,080 Speaker 1: able to essentially guarantee that their name would become synonymous 764 00:50:22,080 --> 00:50:25,080 Speaker 1: with computers. People thought, well, if I'm buying a computer, 765 00:50:25,239 --> 00:50:28,160 Speaker 1: a PC at any rate, not a Mac, then it's 766 00:50:28,200 --> 00:50:30,600 Speaker 1: gonna have that Intel and side sticker on there. It's 767 00:50:30,600 --> 00:50:33,280 Speaker 1: gonna have Intel inside on all the advertising, and Intel 768 00:50:33,320 --> 00:50:36,400 Speaker 1: became a household name, so it was a very valuable 769 00:50:36,560 --> 00:50:42,440 Speaker 1: move on Intel's part. In n Intel introduced the Pentium 770 00:50:42,600 --> 00:50:46,279 Speaker 1: line of processors. I remember when that happened because it 771 00:50:46,320 --> 00:50:49,080 Speaker 1: confused me because I was used to two eighty six 772 00:50:49,120 --> 00:50:51,480 Speaker 1: three eight six for eighty six, and then it went 773 00:50:51,560 --> 00:50:55,200 Speaker 1: to Pentium. These were still built on the X eight 774 00:50:55,320 --> 00:50:58,760 Speaker 1: six processor architecture, but they did depart from that eighty 775 00:50:58,800 --> 00:51:01,799 Speaker 1: six naming system. They decided to go with more trademark 776 00:51:01,920 --> 00:51:04,759 Speaker 1: names and less about just numbers. They part of this 777 00:51:05,000 --> 00:51:10,080 Speaker 1: was probably to appeal to a larger audience of computer consumers, 778 00:51:10,640 --> 00:51:15,360 Speaker 1: people who probably didn't really care for referring to processors 779 00:51:15,400 --> 00:51:20,160 Speaker 1: by what seemed to be meaningless numbers to them. Now, 780 00:51:20,160 --> 00:51:22,360 Speaker 1: as you can imagine, these chips were more efficient and 781 00:51:22,400 --> 00:51:25,080 Speaker 1: faster than their predecessors, and this helped usher in an 782 00:51:25,120 --> 00:51:29,680 Speaker 1: era of PC gaming. Among other things, the faster processors 783 00:51:29,719 --> 00:51:33,120 Speaker 1: allowed game developers to create software that was graphically intense 784 00:51:33,160 --> 00:51:36,279 Speaker 1: and fast paced. And while games had always been part 785 00:51:36,400 --> 00:51:40,480 Speaker 1: of personal computer history ever since personal computers had first 786 00:51:40,640 --> 00:51:44,720 Speaker 1: come on the scene, they now could rival consoles, which 787 00:51:44,840 --> 00:51:49,879 Speaker 1: are essentially specific computers designed to run very particular types 788 00:51:49,920 --> 00:51:54,399 Speaker 1: of software. So a console video game console like the 789 00:51:54,600 --> 00:51:57,080 Speaker 1: Xbox or back in the day of this the Super 790 00:51:57,160 --> 00:52:03,560 Speaker 1: Nintendo Nintendo Entertainment System. It's just a specialized computer. But uh, 791 00:52:03,680 --> 00:52:07,760 Speaker 1: Intel was able to show that with these Pentium processors, 792 00:52:08,960 --> 00:52:13,000 Speaker 1: personal computers could run games that would rival even the 793 00:52:13,040 --> 00:52:18,080 Speaker 1: best consoles. But there was a big problem. Those first 794 00:52:18,120 --> 00:52:21,160 Speaker 1: Pentium chips had a design flaw in them that would 795 00:52:21,239 --> 00:52:26,919 Speaker 1: cause the microprocessor to make it incorrect calculation for certain processes. Now, 796 00:52:26,960 --> 00:52:31,280 Speaker 1: this incorrect calculation didn't happen frequently, but it did happen, 797 00:52:32,000 --> 00:52:34,600 Speaker 1: and while Intel had hoped that perhaps the flaw was 798 00:52:34,640 --> 00:52:38,280 Speaker 1: so minor as to go unnoticed, that did not happen. 799 00:52:39,320 --> 00:52:43,239 Speaker 1: The reason that was noticed was due to a mathematician 800 00:52:43,520 --> 00:52:48,600 Speaker 1: named Thomas Nicely. It was a professor Nicely was actually 801 00:52:48,640 --> 00:52:52,880 Speaker 1: working on a math problem involving twin primes. So a 802 00:52:52,960 --> 00:52:56,000 Speaker 1: twin prime is a set of two prime numbers that 803 00:52:56,120 --> 00:53:00,480 Speaker 1: differ by just two. So three and five are both 804 00:53:00,480 --> 00:53:03,480 Speaker 1: prime numbers and their twin primes because five miles three 805 00:53:03,480 --> 00:53:07,520 Speaker 1: equals two, five and seven are twin primes. Seven milus 806 00:53:07,560 --> 00:53:11,400 Speaker 1: five is two. But as numbers get larger, primes become 807 00:53:11,680 --> 00:53:16,880 Speaker 1: less frequent, and twin primes become even more rare. The 808 00:53:16,880 --> 00:53:21,080 Speaker 1: philosopher Euclid created a proof that suggested there's an infinite 809 00:53:21,160 --> 00:53:24,640 Speaker 1: number of prime numbers, and he believed that the same 810 00:53:24,800 --> 00:53:27,000 Speaker 1: was going to be true of twin primes, but he 811 00:53:27,000 --> 00:53:31,080 Speaker 1: had no way of proving it well. Euclid was thinking 812 00:53:31,200 --> 00:53:34,520 Speaker 1: that way back in three hundred b c. It wouldn't 813 00:53:34,520 --> 00:53:36,920 Speaker 1: be until the twentieth century before someone came up with 814 00:53:36,920 --> 00:53:40,640 Speaker 1: a means of describing this mathematically, and that person was 815 00:53:40,880 --> 00:53:45,839 Speaker 1: Vigo Brune, a Norwegian mathematician. He said that the sum 816 00:53:46,120 --> 00:53:50,120 Speaker 1: of the reciprocals of twin primes would converge to a 817 00:53:50,200 --> 00:53:53,680 Speaker 1: constant sum that later on we would call Bruns constant. 818 00:53:54,360 --> 00:53:56,880 Speaker 1: And in case you're rusty on your mathematical terms, a 819 00:53:56,960 --> 00:53:59,400 Speaker 1: reciprocal of a number is what you get when you 820 00:53:59,480 --> 00:54:02,880 Speaker 1: take one and divide it by that number in question. 821 00:54:03,360 --> 00:54:07,279 Speaker 1: So let's say you start with five. The reciprocal of 822 00:54:07,360 --> 00:54:14,040 Speaker 1: five is one five, because again you're looking at the 823 00:54:14,120 --> 00:54:16,279 Speaker 1: one divided by the number you had started with, or 824 00:54:16,560 --> 00:54:19,239 Speaker 1: point two if you prefer. So. Brun figured out that 825 00:54:19,280 --> 00:54:22,560 Speaker 1: if you took the reciprocals of twin primes and you 826 00:54:22,680 --> 00:54:26,480 Speaker 1: added the two reciprocals together, they converged on a number 827 00:54:26,520 --> 00:54:29,160 Speaker 1: that would end up again being called Bruins constant, and 828 00:54:29,200 --> 00:54:31,840 Speaker 1: by nineteen seventy six that constant was calculated to be 829 00:54:31,880 --> 00:54:37,680 Speaker 1: approximately one point nine oh two one six o five 830 00:54:38,000 --> 00:54:42,680 Speaker 1: four for twin primes up to one hundred billion. But hey, 831 00:54:42,719 --> 00:54:44,799 Speaker 1: that's not good enough for math. We need to be 832 00:54:44,880 --> 00:54:49,160 Speaker 1: more specific than that. So enter Thomas Nicely, who bought 833 00:54:49,160 --> 00:54:52,960 Speaker 1: a computer in n with a Pentium processor and he 834 00:54:53,000 --> 00:54:55,960 Speaker 1: wanted to run calculations related to Bruin's constant. But as 835 00:54:56,000 --> 00:54:58,319 Speaker 1: he did so, he was coming up with errors that 836 00:54:58,360 --> 00:55:01,200 Speaker 1: should not have been there, and he was eventually able 837 00:55:01,200 --> 00:55:04,360 Speaker 1: to track it down to a design flaw and Intel's 838 00:55:04,440 --> 00:55:06,640 Speaker 1: chips that only showed up if you were doing some 839 00:55:06,680 --> 00:55:11,040 Speaker 1: really big calculations. That's when the flaw would make itself apparent, 840 00:55:11,120 --> 00:55:15,000 Speaker 1: but most people would never see it. At first, Intel 841 00:55:15,040 --> 00:55:18,880 Speaker 1: tried to shirk this issue, but intense pressure caused the 842 00:55:18,920 --> 00:55:22,480 Speaker 1: company to change its tune. So eventually Intel issued a 843 00:55:22,520 --> 00:55:26,480 Speaker 1: recall offering a free replacement of the affected print Pentium 844 00:55:26,560 --> 00:55:30,279 Speaker 1: chips with a modified chip that corrected this error, and 845 00:55:30,360 --> 00:55:34,080 Speaker 1: that recall would end up costing Intel four hundred seventy 846 00:55:34,200 --> 00:55:40,760 Speaker 1: five million dollars. Ouch. Well, we've got a bit more 847 00:55:40,800 --> 00:55:43,840 Speaker 1: history of Intel to talk about But before I wrap 848 00:55:43,880 --> 00:55:46,839 Speaker 1: all this up, let's take another quick break to thank 849 00:55:46,880 --> 00:55:58,759 Speaker 1: our sponsor. In Intel would debut the Pentium two. They 850 00:55:58,800 --> 00:56:01,400 Speaker 1: actually had a slight missed up between the Pentium and 851 00:56:01,400 --> 00:56:04,440 Speaker 1: the Pentium two called the Pentium Pro that didn't go 852 00:56:04,480 --> 00:56:06,400 Speaker 1: over so well, but the Pentium two was a slightly 853 00:56:06,440 --> 00:56:10,320 Speaker 1: different story, and Intel also introduced two new product lines 854 00:56:10,320 --> 00:56:14,120 Speaker 1: of processors, the Cellern and Zion brands. Now, these were 855 00:56:14,200 --> 00:56:18,879 Speaker 1: based off of Intel's existing architecture, but they had very 856 00:56:18,920 --> 00:56:23,320 Speaker 1: specific implementations where some of the bells and whistles weren't 857 00:56:23,360 --> 00:56:25,440 Speaker 1: present on them. This allowed them to be used for 858 00:56:25,520 --> 00:56:31,000 Speaker 1: very specific implementations, like you could have mid range UH 859 00:56:31,160 --> 00:56:36,399 Speaker 1: servers for example, or lower priced computers because you had 860 00:56:37,480 --> 00:56:39,799 Speaker 1: the same sort of processors but with some of the 861 00:56:39,840 --> 00:56:46,120 Speaker 1: features turned off essentially in these more mid range machines UH. 862 00:56:46,160 --> 00:56:49,600 Speaker 1: That was Intel's attempt at serving all different levels of 863 00:56:49,640 --> 00:56:55,640 Speaker 1: the market, not just the premium customers. So later on, 864 00:56:56,040 --> 00:57:00,359 Speaker 1: from that point, Intel introduced the Pentium three, which would 865 00:57:00,360 --> 00:57:03,600 Speaker 1: become the first microprocessor to break the one giga hurts 866 00:57:03,719 --> 00:57:06,839 Speaker 1: clock speed. Intel had been in a race with its 867 00:57:06,840 --> 00:57:10,799 Speaker 1: competitor a m D to create the first commercial processor 868 00:57:11,280 --> 00:57:14,560 Speaker 1: that could hit a giga hurts clock speed, and at 869 00:57:14,560 --> 00:57:17,560 Speaker 1: one point Intel even had a model of the Pentium 870 00:57:17,600 --> 00:57:20,160 Speaker 1: three that had a processing speed of or a clock 871 00:57:20,200 --> 00:57:23,680 Speaker 1: speed I should say of one point one three giga hurts, 872 00:57:23,720 --> 00:57:28,640 Speaker 1: but that one was unstable. Uh A. A review found 873 00:57:28,880 --> 00:57:32,080 Speaker 1: the performance was unstable and so that one got recalled. 874 00:57:33,000 --> 00:57:36,040 Speaker 1: But still gigga hurts clock speed meant that it could 875 00:57:36,080 --> 00:57:42,000 Speaker 1: complete a billion instructions per second under ideal circumstances. By 876 00:57:42,040 --> 00:57:44,920 Speaker 1: the late nineteen nineties, Intel was starting to get into 877 00:57:44,960 --> 00:57:48,680 Speaker 1: the business of buying other businesses. It was acquiring other 878 00:57:48,760 --> 00:57:52,880 Speaker 1: businesses to get into new markets that included wireless communications products, 879 00:57:53,400 --> 00:57:57,840 Speaker 1: networking components, and controlled chips for specific types of applications, 880 00:57:57,840 --> 00:58:00,320 Speaker 1: such as the kind you might find in v equal 881 00:58:00,360 --> 00:58:03,800 Speaker 1: control systems. And if you listen to my shows about 882 00:58:03,840 --> 00:58:06,680 Speaker 1: the Macintosh, you know that. In two thousand five, Steve 883 00:58:06,800 --> 00:58:10,200 Speaker 1: Jobs shocked the tech world when he announced that Mac 884 00:58:10,240 --> 00:58:13,600 Speaker 1: computers would move away from their traditional CPUs and use 885 00:58:13,760 --> 00:58:19,320 Speaker 1: Intel CPUs instead. Intel maintained a competitive advantage over others 886 00:58:19,360 --> 00:58:23,280 Speaker 1: in the space, but this also came at a pretty 887 00:58:23,320 --> 00:58:28,160 Speaker 1: stiff price. In two thousand nine, the European Union found 888 00:58:28,160 --> 00:58:33,840 Speaker 1: guilty Intel guilty of alleged monopolistic actions. In other words, 889 00:58:33,880 --> 00:58:36,800 Speaker 1: they said that Intel was effectively acting like a monopoly 890 00:58:36,960 --> 00:58:40,680 Speaker 1: within the European Union, and they hit Intel with a 891 00:58:40,760 --> 00:58:45,920 Speaker 1: fine that was one point four five billion dollars. That 892 00:58:46,040 --> 00:58:49,040 Speaker 1: same year, Intel had to pay its competitor, A m D, 893 00:58:49,520 --> 00:58:53,560 Speaker 1: a hefty sum of one point to five billion dollars 894 00:58:54,040 --> 00:58:56,280 Speaker 1: as part of a settlement in which am D had 895 00:58:56,320 --> 00:59:00,640 Speaker 1: accused Intel of using leverage on computer company is two 896 00:59:01,440 --> 00:59:06,760 Speaker 1: put Intel products in instead of other processors into those computers. 897 00:59:06,800 --> 00:59:09,479 Speaker 1: So essentially am D was saying Intel was being anti 898 00:59:09,520 --> 00:59:14,600 Speaker 1: competitive and really putting the screws to computer manufacturer saying 899 00:59:14,600 --> 00:59:17,800 Speaker 1: you've got to use Intel products and not anyone else, 900 00:59:18,520 --> 00:59:22,280 Speaker 1: and uh, that is not legal, which is why the 901 00:59:22,400 --> 00:59:26,280 Speaker 1: lawsuit happened, and eventually the two settled. Now I could 902 00:59:26,320 --> 00:59:30,400 Speaker 1: talk about how every generation of Intel's chips were faster, 903 00:59:31,560 --> 00:59:36,360 Speaker 1: included more on board memory, had faster bus speeds, greater 904 00:59:36,440 --> 00:59:39,680 Speaker 1: clock speeds, lower latency, but we can just sum it 905 00:59:39,760 --> 00:59:42,760 Speaker 1: up to say the company kept following a strategy that 906 00:59:42,880 --> 00:59:47,280 Speaker 1: it branded the tick talk strategy. So let me tell 907 00:59:47,280 --> 00:59:50,880 Speaker 1: you what TikTok it means and and why that has 908 00:59:50,920 --> 00:59:55,920 Speaker 1: recently changed. So TikTok is a two part strategy to 909 00:59:56,040 --> 01:00:03,280 Speaker 1: developing the microprocessors of today or really yesterday at this point, 910 01:00:03,280 --> 01:00:07,840 Speaker 1: because again we're just now beyond TikTok. The tick part 911 01:00:08,720 --> 01:00:14,160 Speaker 1: involves looking at the architecture you designed on your previous 912 01:00:14,200 --> 01:00:19,480 Speaker 1: generation of micro chips. So you look at that architecture 913 01:00:19,520 --> 01:00:22,360 Speaker 1: and you look at the size of it, and you decide, 914 01:00:22,480 --> 01:00:25,520 Speaker 1: how can I shrink down these components to the next 915 01:00:25,720 --> 01:00:30,520 Speaker 1: smaller size. That is your tick. So you might look 916 01:00:30,560 --> 01:00:34,320 Speaker 1: at a generation of processors that have forty five nanometer components, 917 01:00:34,960 --> 01:00:37,400 Speaker 1: and your challenge is to figure out how to shrink 918 01:00:37,480 --> 01:00:41,520 Speaker 1: down that same design so that the components measure thirty 919 01:00:41,520 --> 01:00:45,600 Speaker 1: two nanometers not forty five. And you follow essentially the 920 01:00:45,640 --> 01:00:48,480 Speaker 1: same blueprint as you did before, only you can cram 921 01:00:48,720 --> 01:00:52,680 Speaker 1: more stuff into that blueprint because you reduced the size 922 01:00:52,680 --> 01:00:55,560 Speaker 1: of the individual components. You freed up some space by 923 01:00:55,600 --> 01:01:02,080 Speaker 1: making everything smaller. However, at these sizes, how those components 924 01:01:02,120 --> 01:01:06,480 Speaker 1: are arranged matters more and more. So you could just 925 01:01:06,600 --> 01:01:10,880 Speaker 1: keep shrinking the components down as your research and development 926 01:01:10,880 --> 01:01:17,120 Speaker 1: allows you to make ever smaller pieces, but that only 927 01:01:17,160 --> 01:01:19,600 Speaker 1: gets you so far. What you really need to do 928 01:01:19,720 --> 01:01:23,480 Speaker 1: is figure out how to arrange those new smaller pieces 929 01:01:23,520 --> 01:01:27,200 Speaker 1: in the ideal configuration to get the most efficient use 930 01:01:27,280 --> 01:01:30,240 Speaker 1: out of them. And this is the TALK step. So 931 01:01:30,280 --> 01:01:33,080 Speaker 1: in TICK you figure out how to shrink stuff down further, 932 01:01:33,520 --> 01:01:36,640 Speaker 1: and talk you figure out how to rearrange these new 933 01:01:36,760 --> 01:01:40,240 Speaker 1: smaller components so that they work the best way you 934 01:01:40,320 --> 01:01:44,280 Speaker 1: possibly can make them. So in our example, we would 935 01:01:44,320 --> 01:01:48,200 Speaker 1: be looking at those thirty two nanometer components and figuring 936 01:01:48,200 --> 01:01:53,760 Speaker 1: out the right architecture to maximize their efficiency. And the TikTok. 937 01:01:54,000 --> 01:01:58,400 Speaker 1: Generations of a single family of processors are going to 938 01:01:58,440 --> 01:02:02,480 Speaker 1: have the same size opponents, They're just gonna have different configurations. 939 01:02:02,920 --> 01:02:05,400 Speaker 1: The TICK is going to be based on the previous generation. 940 01:02:05,840 --> 01:02:11,760 Speaker 1: The TALK is the new architecture to maximize the efficiency 941 01:02:11,880 --> 01:02:15,960 Speaker 1: of the new sized components. And then your next TICK 942 01:02:16,640 --> 01:02:19,400 Speaker 1: is going to be taking those thirty two nanometer components 943 01:02:19,440 --> 01:02:21,760 Speaker 1: and figuring out how to shrink them down even further, 944 01:02:22,360 --> 01:02:26,640 Speaker 1: but within the same general architecture as your previous generation. 945 01:02:27,080 --> 01:02:32,640 Speaker 1: Tick talk, tick talk. That was pretty much how Intel 946 01:02:32,760 --> 01:02:38,240 Speaker 1: ran things uh for quite a few generations of processors. 947 01:02:39,240 --> 01:02:43,360 Speaker 1: Now Intel follows a new strategy. It calls it the 948 01:02:43,480 --> 01:02:49,480 Speaker 1: process architecture Optimize pattern others will call this tick talk 949 01:02:49,560 --> 01:02:54,280 Speaker 1: talk because again, the first one is shrinking components down 950 01:02:54,320 --> 01:02:59,160 Speaker 1: to a new smaller size, the second means finding a 951 01:02:59,200 --> 01:03:03,520 Speaker 1: better arch at lecture for those sized components, and the 952 01:03:03,560 --> 01:03:07,520 Speaker 1: third one is refining that design even further, so you're 953 01:03:07,560 --> 01:03:12,400 Speaker 1: staying on the same size of components for three generations 954 01:03:12,600 --> 01:03:15,720 Speaker 1: in a row. UH. This means that you don't have 955 01:03:15,760 --> 01:03:17,440 Speaker 1: to spend so much time trying to figure out how 956 01:03:17,480 --> 01:03:20,000 Speaker 1: are you going to shrink things down even further as 957 01:03:20,040 --> 01:03:25,240 Speaker 1: you gradually get closer and closer to a fundamental physical limitation. UH. 958 01:03:25,280 --> 01:03:29,640 Speaker 1: That being where quantum physics comes in and doesn't play 959 01:03:29,760 --> 01:03:32,800 Speaker 1: nicely with your designs anymore, and you get things like 960 01:03:32,840 --> 01:03:39,120 Speaker 1: electron tunneling where electrons seem to leak through transistors, and 961 01:03:39,160 --> 01:03:43,080 Speaker 1: since transistors are meant to control the flow of electrons, 962 01:03:43,120 --> 01:03:46,000 Speaker 1: this is what we in the computer biz often call 963 01:03:46,640 --> 01:03:51,200 Speaker 1: a bad thing. It introduces the possibility of errors and miscalculations, 964 01:03:51,560 --> 01:03:55,960 Speaker 1: and you don't want that in your processor, so it 965 01:03:56,080 --> 01:03:59,680 Speaker 1: ends up extending the amount of time Intel spends on 966 01:04:00,560 --> 01:04:05,880 Speaker 1: a specific size of dye for their their chips, but 967 01:04:05,960 --> 01:04:10,560 Speaker 1: it also maximizes the efficiency of that while engineers continue 968 01:04:10,560 --> 01:04:15,280 Speaker 1: to work on the next breakthrough. One other thing I 969 01:04:15,280 --> 01:04:19,240 Speaker 1: need to touch on before I conclude is the concept 970 01:04:19,320 --> 01:04:23,320 Speaker 1: of multi core processors, because Intel story doing that as well. 971 01:04:24,280 --> 01:04:28,680 Speaker 1: Intel's work with parallel processing way back in the nineties 972 01:04:29,120 --> 01:04:33,920 Speaker 1: provided the basis for multi core processors. Multi core processors 973 01:04:33,920 --> 01:04:39,240 Speaker 1: can handle several computational problems simultaneously, which brings the clock 974 01:04:39,320 --> 01:04:41,760 Speaker 1: speeds to bear on the problems to solve them in 975 01:04:41,800 --> 01:04:45,520 Speaker 1: parallel rather than in sequence. So one core might be 976 01:04:45,560 --> 01:04:48,520 Speaker 1: working on one problem, another core could be working on 977 01:04:48,720 --> 01:04:51,520 Speaker 1: a separate problem, So dual core processors. You could do 978 01:04:51,600 --> 01:04:54,760 Speaker 1: two of these, quad core four and so on and 979 01:04:54,800 --> 01:04:57,720 Speaker 1: so forth. And there's also threading as well, you can 980 01:04:57,800 --> 01:05:04,080 Speaker 1: thread different copy stational problems. But ultimately the concept we 981 01:05:04,120 --> 01:05:06,760 Speaker 1: need to really focus on here is that idea of 982 01:05:06,800 --> 01:05:11,880 Speaker 1: parallel processing, because for certain types of computational problems, parallel 983 01:05:11,920 --> 01:05:17,120 Speaker 1: processing is much much faster than doing sequential processing where 984 01:05:17,160 --> 01:05:21,040 Speaker 1: you're just going down a list of instructions. Even if 985 01:05:21,080 --> 01:05:25,520 Speaker 1: you make a really really super fast CPU, if it's 986 01:05:25,520 --> 01:05:28,720 Speaker 1: tackling a very long list of instructions that could otherwise 987 01:05:28,720 --> 01:05:32,400 Speaker 1: be divided up, a multi core processor might be more 988 01:05:32,440 --> 01:05:38,440 Speaker 1: effective than a very fast single core CPU. I like 989 01:05:38,520 --> 01:05:41,160 Speaker 1: to use an analogy whenever I talk about this, and 990 01:05:41,280 --> 01:05:44,160 Speaker 1: long time listeners of tech stuff are probably familiar with 991 01:05:44,200 --> 01:05:46,680 Speaker 1: us because I've used it before, but I find it's 992 01:05:46,920 --> 01:05:50,240 Speaker 1: very helpful if you're trying to understand the difference between 993 01:05:50,480 --> 01:05:56,479 Speaker 1: a super fast CPU and a fast but not as 994 01:05:56,720 --> 01:06:01,680 Speaker 1: crazy fast multi core processor. So we're going to imagine 995 01:06:01,840 --> 01:06:08,680 Speaker 1: a math class and you've got essentially sixteen kids in 996 01:06:08,720 --> 01:06:12,280 Speaker 1: this math class, right, one of those kids is a 997 01:06:12,320 --> 01:06:16,600 Speaker 1: math genius. She's a prodigy. She's so smart she can 998 01:06:16,680 --> 01:06:20,080 Speaker 1: complete any problem in a fraction of the time of 999 01:06:20,120 --> 01:06:23,680 Speaker 1: any of her other classmates. Her classmates, by the way, 1000 01:06:23,800 --> 01:06:27,680 Speaker 1: are smart. They're they're not They're not slower anything, they're 1001 01:06:27,680 --> 01:06:32,040 Speaker 1: just not geniuses like she is. Now, on any given 1002 01:06:32,240 --> 01:06:35,920 Speaker 1: singular problem, the genius is always going to solve it 1003 01:06:36,000 --> 01:06:40,640 Speaker 1: faster than her her other classmates. They're just never going 1004 01:06:40,680 --> 01:06:44,440 Speaker 1: to be as fast as she is for any one problem. 1005 01:06:44,560 --> 01:06:47,880 Speaker 1: But imagine the teacher hands out the test, and the 1006 01:06:47,920 --> 01:06:51,680 Speaker 1: test has fifteen problems on it, So there are fifteen 1007 01:06:51,720 --> 01:06:55,640 Speaker 1: problems you have to solve, and they're sixteen kids in 1008 01:06:55,640 --> 01:07:00,160 Speaker 1: the class, including the genius. The teacher gives the genius 1009 01:07:00,200 --> 01:07:04,000 Speaker 1: a test that has all sixteen, all fifteen problems on it. 1010 01:07:04,440 --> 01:07:07,440 Speaker 1: So she has to solve all fifteen of those problems 1011 01:07:07,480 --> 01:07:10,160 Speaker 1: as quickly as she can. But he tells the other 1012 01:07:10,280 --> 01:07:14,640 Speaker 1: fifteen students, you each will tackle one of these problems, 1013 01:07:14,680 --> 01:07:18,200 Speaker 1: and he assigns them. Student one has problem one, Student 1014 01:07:18,280 --> 01:07:21,000 Speaker 1: two has problem to, and so on all the way 1015 01:07:21,040 --> 01:07:24,960 Speaker 1: down the fifteen problems, and they have their job is 1016 01:07:25,000 --> 01:07:30,080 Speaker 1: to complete their problem, their one problem before the genius 1017 01:07:30,120 --> 01:07:35,240 Speaker 1: can complete all fifteen problems. Well, this is sort of 1018 01:07:35,240 --> 01:07:39,080 Speaker 1: what multi core processors are capable of doing. They divide 1019 01:07:39,120 --> 01:07:43,600 Speaker 1: up problems into different parts, and collectively they can solve 1020 01:07:43,680 --> 01:07:47,880 Speaker 1: that big problem faster than a really super fast processor 1021 01:07:48,000 --> 01:07:51,920 Speaker 1: could nine times out of ten, ninety nine times out 1022 01:07:51,920 --> 01:07:55,680 Speaker 1: of a hundred, nine times out of a thousand. Those 1023 01:07:55,760 --> 01:07:59,240 Speaker 1: fifteen students are going to finish their individual problems before 1024 01:07:59,280 --> 01:08:02,760 Speaker 1: the math genie can work through all fifteen on her test. 1025 01:08:03,960 --> 01:08:05,800 Speaker 1: And you might say, well, that's not fair to like, 1026 01:08:05,920 --> 01:08:08,120 Speaker 1: that's not the point. It's an analogy to explain how 1027 01:08:08,200 --> 01:08:12,800 Speaker 1: multi core processors work from a kind of high level approach. 1028 01:08:13,240 --> 01:08:15,760 Speaker 1: It gets a lot more technical than that, obviously, but 1029 01:08:15,840 --> 01:08:20,240 Speaker 1: that's just to explain that for certain types of computational problems, 1030 01:08:20,560 --> 01:08:24,280 Speaker 1: parallel processing is much more effective. Now, there are types 1031 01:08:24,320 --> 01:08:28,600 Speaker 1: of computational problems that cannot be broken into parallel processing, 1032 01:08:28,920 --> 01:08:33,280 Speaker 1: and for those a super fast CPU is still more 1033 01:08:33,439 --> 01:08:35,320 Speaker 1: often than not going to be more effective than a 1034 01:08:35,360 --> 01:08:38,920 Speaker 1: multi core processor. So it really just depends upon the application. 1035 01:08:39,720 --> 01:08:43,040 Speaker 1: But more and more we're seeing parallel processing problems being 1036 01:08:43,080 --> 01:08:47,799 Speaker 1: the type that computers tackle. Now, how long will Intel 1037 01:08:47,840 --> 01:08:52,120 Speaker 1: and other microprocessor manufacturers be able to keep Moore's law alive? 1038 01:08:52,520 --> 01:08:55,479 Speaker 1: Because people are constantly predicting the end to Moore's law 1039 01:08:55,600 --> 01:08:59,360 Speaker 1: has happened numerous times throughout history. From the eighties on. 1040 01:09:00,000 --> 01:09:01,960 Speaker 1: People will say, oh, Moore's law is coming to an 1041 01:09:02,040 --> 01:09:04,960 Speaker 1: end because it's not physically possible for us to keep 1042 01:09:05,040 --> 01:09:07,840 Speaker 1: up with it. But so far, engineers have been able 1043 01:09:07,880 --> 01:09:12,200 Speaker 1: to stave that off, partly through innovative architectures that don't 1044 01:09:12,200 --> 01:09:15,799 Speaker 1: necessarily increase the number of components by a factor of two, 1045 01:09:16,040 --> 01:09:20,040 Speaker 1: but rather increase the output of the processor by a 1046 01:09:20,040 --> 01:09:23,960 Speaker 1: factor of two every two years. So it requires some 1047 01:09:24,040 --> 01:09:28,400 Speaker 1: reinterpretation of what Moore's law means. It may mean that 1048 01:09:28,479 --> 01:09:31,480 Speaker 1: you fudge a bit on the amount of time necessary 1049 01:09:31,520 --> 01:09:34,360 Speaker 1: to get to that factor of two. And it might 1050 01:09:34,400 --> 01:09:40,240 Speaker 1: mean reinterpreting it as processing power versus number of discrete elements. 1051 01:09:41,120 --> 01:09:44,160 Speaker 1: But I think the important thing for us as consumers 1052 01:09:44,439 --> 01:09:48,799 Speaker 1: is the performance, not whether the chip actually has twice 1053 01:09:48,800 --> 01:09:51,400 Speaker 1: as many transistors as the one from two years ago. 1054 01:09:52,439 --> 01:09:57,520 Speaker 1: So that is still a battle that's going on today. Uh. 1055 01:09:57,640 --> 01:10:01,479 Speaker 1: It does mean that maybe we'll hit that fundamental limit 1056 01:10:01,640 --> 01:10:03,960 Speaker 1: at some point and we will not be able to 1057 01:10:04,000 --> 01:10:09,040 Speaker 1: continue More's law using traditional microprocessor designs, and then we 1058 01:10:09,120 --> 01:10:14,519 Speaker 1: may need another true revolution in computer processing to create 1059 01:10:14,640 --> 01:10:19,480 Speaker 1: a new means of keeping up with that power output. 1060 01:10:20,240 --> 01:10:23,960 Speaker 1: But maybe stepping outside of what a processor is supposed 1061 01:10:24,000 --> 01:10:28,840 Speaker 1: to look like, that's a possibility today. Intel is doing 1062 01:10:28,880 --> 01:10:31,920 Speaker 1: pretty well as of the recording of this podcast. When 1063 01:10:31,960 --> 01:10:34,840 Speaker 1: I checked it was its shares were trading at around 1064 01:10:34,840 --> 01:10:38,080 Speaker 1: thirty five dollars a share. That puts the market cap 1065 01:10:38,160 --> 01:10:40,960 Speaker 1: value of Intel at about a hundred and sixty five 1066 01:10:41,320 --> 01:10:46,479 Speaker 1: billion dollars. The company may nearly sixty billion in revenue 1067 01:10:46,560 --> 01:10:49,680 Speaker 1: in twenty six, which is not too bad for a 1068 01:10:49,720 --> 01:10:53,639 Speaker 1: company that was found by a couple of traders. Well done, 1069 01:10:54,000 --> 01:10:59,840 Speaker 1: More and Noise. That wraps up our two part episode 1070 01:11:00,000 --> 01:11:03,120 Speaker 1: on the story of Intel. Obviously, there's a lot more 1071 01:11:03,120 --> 01:11:06,800 Speaker 1: I could talk about from the various sensors and processors 1072 01:11:06,800 --> 01:11:10,000 Speaker 1: Intel has been developing for all sorts of applications to 1073 01:11:10,120 --> 01:11:14,000 Speaker 1: their partnerships with various companies throughout its history, and maybe 1074 01:11:14,000 --> 01:11:16,320 Speaker 1: in future episodes I'll touch on some of those, but 1075 01:11:16,400 --> 01:11:18,240 Speaker 1: I thought that it was really important to just kind 1076 01:11:18,240 --> 01:11:21,720 Speaker 1: of hit the high points to understand where Intel came 1077 01:11:21,800 --> 01:11:24,720 Speaker 1: from and how it developed over its years. If you 1078 01:11:24,720 --> 01:11:27,960 Speaker 1: guys have any suggestions for future episodes of Text Stuff, 1079 01:11:28,200 --> 01:11:31,240 Speaker 1: whether it's a topic I should cover, someone I should 1080 01:11:31,280 --> 01:11:34,639 Speaker 1: have on to interview, a guest co host who might 1081 01:11:34,640 --> 01:11:37,840 Speaker 1: be able to tackle a specific subject with me, let 1082 01:11:37,840 --> 01:11:41,160 Speaker 1: me know. Send me an email. The address for the 1083 01:11:41,160 --> 01:11:45,560 Speaker 1: show is tech Stuff at how stuff works dot com, 1084 01:11:45,680 --> 01:11:48,720 Speaker 1: or you can drop me a line on Twitter or Facebook. 1085 01:11:48,840 --> 01:11:51,640 Speaker 1: The handle for both of those is text Stuff h 1086 01:11:51,920 --> 01:11:55,400 Speaker 1: s W. Remember you can tune in on Wednesdays and 1087 01:11:55,439 --> 01:12:00,000 Speaker 1: Friday's to see me stream this podcast live. You can 1088 01:12:00,000 --> 01:12:03,400 Speaker 1: watch me record it live, which includes all the ridiculous 1089 01:12:03,439 --> 01:12:06,200 Speaker 1: mistakes I make and when I have to restart, and 1090 01:12:06,280 --> 01:12:08,920 Speaker 1: sometimes I'm chatting with the chat room and just answering 1091 01:12:09,000 --> 01:12:12,240 Speaker 1: questions or shooting the breeze. Go to twitch dot tv 1092 01:12:12,439 --> 01:12:15,439 Speaker 1: slash tech stuff. To see the schedule there and join 1093 01:12:15,520 --> 01:12:19,280 Speaker 1: me sometime, won't you. And in the meantime, I hope 1094 01:12:19,280 --> 01:12:21,519 Speaker 1: you guys have a great week and I'll talk to 1095 01:12:21,520 --> 01:12:29,799 Speaker 1: you again really soon. For more on this and thousands 1096 01:12:29,800 --> 01:12:41,880 Speaker 1: of other topics, is it staff works dot com.