WEBVTT - TechStuff Tidbits: CPU Basics 0:00:04.360 --> 0:00:12.440 Welcome to tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.440 --> 0:00:15.480 and welcome to tech Stuff. I'm your host, Jonathan Strickland. 0:00:15.560 --> 0:00:19.120 I'm an executive producer with iHeartRadio. And how the tech 0:00:19.200 --> 0:00:23.560 are you? It's time for a tech Stuff Tidbits episode 0:00:23.920 --> 0:00:26.599 and I thought today we could talk a little bit 0:00:26.680 --> 0:00:30.920 about CPUs and some of the stuff around them. So 0:00:30.960 --> 0:00:34.000 if you are someone who is you know, familiar with 0:00:34.040 --> 0:00:36.800 the term, but you don't know a whole lot about CPUs, 0:00:37.400 --> 0:00:38.960 this is kind of for you. For those of you 0:00:39.000 --> 0:00:42.720 all who have been over clocking your CPUs for years, 0:00:42.720 --> 0:00:45.839 this is going to be beyond basic. So that's just 0:00:45.880 --> 0:00:47.360 a warning for you. But if you want to come 0:00:47.360 --> 0:00:50.760 along for the ride, please do. I like your company. Okay, 0:00:51.320 --> 0:00:54.800 So CPU, what the heck is that? It is a 0:00:54.880 --> 0:00:59.040 central processing unit. You can think of this as the 0:00:59.080 --> 0:01:04.200 brains of your computer. This is the micro chip that 0:01:04.280 --> 0:01:08.800 will accept data. It'll actually send out and request data 0:01:08.880 --> 0:01:12.759 from memory for example, to be able to run operations 0:01:12.920 --> 0:01:17.440 on that data and create output. So you get input 0:01:17.520 --> 0:01:20.640 coming in, you get operations going on, like you know, 0:01:20.720 --> 0:01:24.440 mathematical operations, and then you get results coming out. Those 0:01:24.440 --> 0:01:26.840 results do stuff in the computer. It could be all 0:01:26.840 --> 0:01:30.080 sorts of things like it's it's literally all the things 0:01:30.120 --> 0:01:32.200 that need to happen. And if you think of the 0:01:32.240 --> 0:01:35.200 CPU kind of like a brain, it's not a one 0:01:35.240 --> 0:01:39.440 to one, don't take it like that, but it does 0:01:39.480 --> 0:01:43.480 a similar function. How everything our bodies are doing. Everything 0:01:43.520 --> 0:01:47.000 we are doing is ultimately coming as a result of 0:01:47.280 --> 0:01:51.960 brain activity. That you know, if you if you are 0:01:52.000 --> 0:01:55.800 thinking about it, then that's obviously active. You can understand that. 0:01:55.880 --> 0:01:58.720 But if it's stuff like you know, things that you're 0:01:58.720 --> 0:02:02.000 doing unconsciously, like linking, Like you're not thinking about blinking, 0:02:02.000 --> 0:02:05.760 are you? Please? Please don't think about blinking, because I 0:02:05.800 --> 0:02:08.680 don't want to be responsible for that. You're doing it. Now, 0:02:08.680 --> 0:02:11.639 you're going to think every time you blink for a while, 0:02:11.800 --> 0:02:14.160 and then you'll stop, you'll forget about it for a bit, 0:02:14.440 --> 0:02:15.919 then it'll come back to you later in the day 0:02:15.919 --> 0:02:18.800 and you'll start thinking about blinking again, and I'm that's 0:02:18.800 --> 0:02:24.040 all my fault anyway. CPUs are able to run processes 0:02:24.320 --> 0:02:27.200 on data. That's their main purpose. Now. They come in 0:02:27.280 --> 0:02:31.120 a couple of different varieties and from a couple of 0:02:31.120 --> 0:02:36.040 different manufacturers. The two major manufacturers of CPUs really the 0:02:36.080 --> 0:02:38.600 only two worth talking about when you get down to it, 0:02:38.639 --> 0:02:43.480 because the entire industry has centered around them. Are Intel 0:02:44.080 --> 0:02:51.920 and AMD, and these two companies dominate the CPU industry, 0:02:52.280 --> 0:02:57.160 and in fact the CPUs they make are incompatible with 0:02:57.160 --> 0:03:01.519 one another. Right, you can't swap an Intel chip out 0:03:01.560 --> 0:03:05.200 for an AMD chip on your computer because an Intel 0:03:05.360 --> 0:03:08.840 chip won't fit in an AMD slot and vice versa. 0:03:08.960 --> 0:03:13.600 So this also gives me a chance to mention motherboards, 0:03:13.760 --> 0:03:18.720 which I'll just do briefly. Motherboards are giant circuit boards. Essentially, 0:03:18.840 --> 0:03:21.760 this is like your peripheral nervous system in a way. 0:03:21.800 --> 0:03:24.640 This is what connects all the components to each other 0:03:24.760 --> 0:03:28.239 so that data can actually pass from one thing to 0:03:28.280 --> 0:03:32.120 another and your computer it can work. So you can 0:03:32.200 --> 0:03:35.760 think of this a sort of the vertebrae attaches everything together, 0:03:36.520 --> 0:03:40.840 and motherboards work with specific chips. So if you are 0:03:40.960 --> 0:03:44.400 ever going to build your own computer, then one thing 0:03:44.440 --> 0:03:48.040 you absolutely must make certain you do is to confirm 0:03:48.160 --> 0:03:52.440 that the processor you're interested in will work with the 0:03:52.640 --> 0:03:56.720 mother board you're interested in. If you're getting an AMD motherboard, 0:03:56.760 --> 0:03:59.520 like an AMD compatible motherboard, and you're getting an Intel chip, 0:04:00.040 --> 0:04:02.360 you're going to rapidly find out that you've you're gonna 0:04:02.400 --> 0:04:05.360 have to return something because either you're gonna need to 0:04:05.400 --> 0:04:08.280 get an AMD processor or you're going to get another motherboard. 0:04:08.680 --> 0:04:11.520 But motherboards are where you plug everything else in, right, 0:04:11.640 --> 0:04:13.760 Like if you've got a graphics processing unit, which we'll 0:04:13.760 --> 0:04:16.360 talk about in a second, or you know you've got 0:04:16.440 --> 0:04:18.719 various components that you want to add to it. Maybe 0:04:18.720 --> 0:04:21.400 you've got other expansion slots that you want to fill 0:04:21.480 --> 0:04:24.200 up with various cards of some sort, maybe a capture 0:04:24.240 --> 0:04:27.880 card something like that, that goes on to the motherboard, 0:04:28.080 --> 0:04:30.840 which again acts as kind of like the connectivity for 0:04:30.920 --> 0:04:34.719 all the different elements in your PC. Getting back to CPUs, 0:04:34.720 --> 0:04:36.080 because there are a couple of other things I want 0:04:36.080 --> 0:04:37.719 to talk about. I did mention that there are a 0:04:37.720 --> 0:04:42.360 couple of different types of CPU. Actually there's lots if 0:04:42.360 --> 0:04:45.400 you want to boil down to it. But really one 0:04:45.440 --> 0:04:48.280 thing we can talk about our cores, like a single 0:04:48.320 --> 0:04:53.279 core CPU versus a multi core CPU. For many years, 0:04:53.320 --> 0:04:57.120 single core was pretty much all you had, and you 0:04:57.120 --> 0:05:02.960 could increase the speed of how your CPU could process 0:05:03.000 --> 0:05:07.240 information by getting CPUs that had more and more components 0:05:07.240 --> 0:05:09.720 crammed onto them. But that was pretty much the only 0:05:09.760 --> 0:05:12.479 way you could speed things up. Then we get to 0:05:12.800 --> 0:05:18.120 multi core CPU design, and as the name suggests, these 0:05:18.160 --> 0:05:23.760 CPUs actually consists of kind of like miniature versions of 0:05:23.960 --> 0:05:28.640 CPUs all clumped together in a way. So think of 0:05:28.640 --> 0:05:32.039 it instead of it being one big brain, something like 0:05:32.120 --> 0:05:36.159 maybe for smaller brains or AID or whatever. However many 0:05:36.200 --> 0:05:42.320 cores there are. Multi core CPUs are effective when they're 0:05:42.360 --> 0:05:46.560 dealing with information that can be broken down into parallel 0:05:46.640 --> 0:05:51.919 lines of processing because you can assign different cores by you. 0:05:52.000 --> 0:05:54.160 I mean, this is all happening by the computer. You're 0:05:54.200 --> 0:05:57.480 not taking an active role, but the computer can assign 0:05:57.920 --> 0:06:02.000 different lines of processing to specif it cores, and that way, 0:06:02.040 --> 0:06:05.159 each core is focusing on part of the overall operation, 0:06:05.960 --> 0:06:09.279 and collectively they can solve it faster. So if you've 0:06:09.279 --> 0:06:11.200 listened to tex staff for a very long time, you 0:06:11.320 --> 0:06:13.720 know the analogy I'm about to give because it's one 0:06:13.760 --> 0:06:16.800 of my favorite go toos. Imagine that it's a class 0:06:16.839 --> 0:06:20.239 of math students and everyone in the class is great 0:06:20.240 --> 0:06:23.839 at math, but there's also one absolute genius at math 0:06:23.880 --> 0:06:30.680 who just consistently finishes tests fast and is perfect, and 0:06:30.760 --> 0:06:34.280 she's just the best. Everybody else is great, and they 0:06:34.320 --> 0:06:36.360 also get very good grades, but it takes them more 0:06:36.400 --> 0:06:39.560 time to complete, say a math test. And so as 0:06:39.800 --> 0:06:43.760 a fun experiment, the teacher assigns a pop quiz to 0:06:43.800 --> 0:06:47.880 the class, and the genius student she has all the 0:06:48.480 --> 0:06:51.080 problems on her sheet. She has to answer all of them. 0:06:51.120 --> 0:06:53.680 Let's say that there's six of them, and there are 0:06:53.720 --> 0:06:56.600 six other students. They're really smart ones, but they're not 0:06:56.720 --> 0:07:00.359 genius level. They're each given one of the problem. So 0:07:00.440 --> 0:07:03.000 student one gets problem one, student two gets problem two, 0:07:03.080 --> 0:07:06.440 and so on. They just have to solve their own problem. 0:07:06.480 --> 0:07:09.080 That's it, and then they can turn in their test. 0:07:09.600 --> 0:07:12.080 So the question is who finishes the test first. Well, 0:07:12.120 --> 0:07:14.600 in this approach, it's probably going to be the six students, right, 0:07:14.640 --> 0:07:18.400 because they're each focusing on just one problem. Whereas the genius, 0:07:18.400 --> 0:07:21.600 even though she's faster than any individual student, she has 0:07:21.640 --> 0:07:25.120 to complete all six and she can't do it simultaneously. 0:07:25.160 --> 0:07:28.600 She has to go sequentially. Well, that would be a 0:07:28.640 --> 0:07:33.080 case where the six students would outperform the single one, 0:07:33.080 --> 0:07:35.920 where you could say a multi core processor could handle 0:07:36.040 --> 0:07:40.320 that sort of computational problem more effectively than a single core. 0:07:40.880 --> 0:07:44.880 But that's not how all computer problems come, right, Like 0:07:45.000 --> 0:07:48.440 what if the test worked in such a way where 0:07:48.800 --> 0:07:51.920 the only way you could tackle problem number two is 0:07:51.920 --> 0:07:54.679 if you had the solution to problem number one first, 0:07:55.320 --> 0:07:57.560 then you could tackle problem number two. Same thing with 0:07:57.640 --> 0:07:59.240 problem number three. You have to wait until you have 0:07:59.280 --> 0:08:01.920 the solution to problem number two. Well, then the six 0:08:02.000 --> 0:08:05.800 students are going to lose because we've already established the genius. 0:08:05.880 --> 0:08:09.760 She can finish her test much faster, and meanwhile, student 0:08:09.840 --> 0:08:11.880 number two has to wait until student number one is 0:08:11.920 --> 0:08:15.760 finished before they can tackle problem number two. Right, So, 0:08:16.400 --> 0:08:21.000 depending on the computational problem, multicore may or may not 0:08:21.120 --> 0:08:24.040 be better than a single core. For a lot of 0:08:24.320 --> 0:08:28.560 gaming applications, multi coore can work pretty well, and also 0:08:29.320 --> 0:08:33.080 in a lot of graphics processing units or GPUs, which, 0:08:33.120 --> 0:08:37.319 as the name suggests, are all about handling the processing 0:08:37.640 --> 0:08:44.120 to generate graphics. Those are often parallel processing machines. Right, 0:08:44.160 --> 0:08:48.120 that's just like a chip designed to do parallel processing 0:08:48.760 --> 0:08:53.600 and coincidentally, or not even so coincidentally, but because of that, 0:08:54.320 --> 0:09:00.559 they also were incredibly important for mining of certain crypto currencies, 0:09:00.840 --> 0:09:03.880 initially stuff like bitcoin, but then bitcoin's value got to 0:09:03.920 --> 0:09:07.080 a point where it no longer made financial sense to 0:09:07.120 --> 0:09:12.360 go after graphics processing cards in order to mine bitcoin, 0:09:12.600 --> 0:09:15.840 because you just you would fall behind. People were going 0:09:15.880 --> 0:09:21.000 for purpose built machines in order to mind bitcoin more effectively. 0:09:21.000 --> 0:09:23.920 But then ethereum before it moved on to being a 0:09:23.920 --> 0:09:27.400 proof of steak versus proof of work. That was the 0:09:27.440 --> 0:09:32.080 cryptocurrency that drove people to scoop up as many graphics 0:09:32.080 --> 0:09:34.520 processing unit cards as they possibly could in order to 0:09:34.559 --> 0:09:39.880 create an entire network of computers using these to mine cryptocurrency, 0:09:40.160 --> 0:09:46.760 because like graphics processing, cryptocurrency mining really would benefit from 0:09:46.800 --> 0:09:51.600 parallel processing, which is what multicore processors can do. Okay, 0:09:52.000 --> 0:09:55.280 so that's your basic division, single core and multicore. There's 0:09:55.320 --> 0:09:57.040 some other stuff we need to talk about. One of 0:09:57.080 --> 0:10:01.240 the things is clock speed. So this can seem a 0:10:01.280 --> 0:10:05.080 little confusing at first because you're thinking, well, our experience 0:10:05.160 --> 0:10:08.000 with clocks is that these are devices meant to keep time, 0:10:08.040 --> 0:10:11.520 and therefore the speed of clocks should be uniform, it 0:10:11.520 --> 0:10:14.520 should be universal, right, because you just want to keep time. 0:10:15.200 --> 0:10:18.520 That's not what we're really talking about here. Clock speed 0:10:18.520 --> 0:10:22.800 with CPUs or even GPUs, it's the rate at which 0:10:23.679 --> 0:10:28.400 the processor is able to perform calculations, and we typically 0:10:28.480 --> 0:10:31.880 see this measured these days, especially in gigga hurts. Like 0:10:31.920 --> 0:10:35.599 when I remember from my days of working with computers 0:10:35.600 --> 0:10:38.680 when I was a kid, kill hurts and mega hurts 0:10:38.720 --> 0:10:40.960 were like that was the thing. But gigga hurts is 0:10:41.000 --> 0:10:46.920 definitely where we're at now, and hurts refers to a 0:10:48.000 --> 0:10:52.120 repetitive task in a given amount of time, like how 0:10:52.120 --> 0:10:55.520 many times are you doing something within a given amount 0:10:55.520 --> 0:10:59.000 of time like a second, So one hurts would be 0:10:59.160 --> 0:11:01.679 one repetition per second. When you're talking about gigga hurts, 0:11:01.679 --> 0:11:05.480 you're talking about a billion per second. So when you're 0:11:05.520 --> 0:11:09.200 talking about a clock speed of a processor and it's 0:11:09.360 --> 0:11:11.640 three point five gigga hurts, that means it's three point 0:11:11.640 --> 0:11:17.520 five billion operations per second. I'm going to explain that 0:11:17.559 --> 0:11:21.240 a little bit further because that's kind of a very 0:11:21.360 --> 0:11:25.440 high level version of what clock speed is. But before 0:11:25.440 --> 0:11:38.360 we get to that, let's take a quick break. Okay, 0:11:38.720 --> 0:11:42.800 So I mentioned before that clock speed reference is essentially 0:11:42.880 --> 0:11:48.160 the number of calculations or operations that a CPU can 0:11:48.280 --> 0:11:51.040 make in a given amount of time. In this case 0:11:51.080 --> 0:11:56.560 a second, operations can be actually multi step, so really 0:11:56.559 --> 0:11:59.520 you should think of this as steps that the processor 0:11:59.600 --> 0:12:03.160 can take per second. With some operations requiring only a 0:12:03.160 --> 0:12:06.520 single step, in some perhaps requiring multiple steps. So it 0:12:06.559 --> 0:12:10.120 gets a little more fuzzy when we start talking about operations. 0:12:10.160 --> 0:12:13.640 But generally speaking, the higher the clock speed, the faster 0:12:13.800 --> 0:12:18.800 that processor is at you know, processing information. It can 0:12:18.840 --> 0:12:21.800 process more information in the same amount of time, which 0:12:22.200 --> 0:12:26.120 we end up saying means the processor is faster even 0:12:26.120 --> 0:12:29.000 though you know most of the time it's not moving anywhere, 0:12:29.600 --> 0:12:32.040 at least not if your computer has been set up properly. 0:12:32.600 --> 0:12:37.040 So this also brings us to the concept of overclocking. 0:12:37.840 --> 0:12:45.240 Overclocking is when you end up pushing past the rated 0:12:45.480 --> 0:12:51.080 specifications of a piece of hardware in order to get 0:12:51.120 --> 0:12:54.360 more out of it. So, in other words, when you 0:12:54.400 --> 0:12:57.600 get a processor, it essentially is, you know, it's telling 0:12:57.640 --> 0:13:02.040 you what clock speed it's been set at. That speed 0:13:02.480 --> 0:13:08.520 has been determined by the manufacturer. Usually you can actually 0:13:08.559 --> 0:13:11.480 push a processor to go beyond that speed. You have 0:13:11.520 --> 0:13:13.360 to go into some settings to be able to do this. 0:13:13.360 --> 0:13:16.280 It gets very technical. I'll probably do an episode just 0:13:16.360 --> 0:13:19.320 about that at some point, but for now we'll set 0:13:19.360 --> 0:13:21.560 it aside. The point being is you can go in 0:13:21.600 --> 0:13:26.600 there and essentially remove the regulator that would otherwise keep 0:13:26.679 --> 0:13:28.800 you at that top speed and not let you go 0:13:28.880 --> 0:13:31.679 past it. You can take that down and allow your 0:13:31.800 --> 0:13:35.520 processor to work even harder. This is gonna do a 0:13:35.559 --> 0:13:37.560 couple of things. It's going to generate a lot more heat, 0:13:37.960 --> 0:13:39.520 so you have to have a way of dealing with 0:13:39.600 --> 0:13:43.760 that heat otherwise you run the risk of damaging the processor. 0:13:43.920 --> 0:13:46.840 And then sure it will run super fast for a while, 0:13:47.160 --> 0:13:50.800 but then ultimately it's going to break, it's going to 0:13:50.840 --> 0:13:53.760 wear out, and then you'll have to replace it. So 0:13:53.800 --> 0:13:55.920 you really need to have a way of carrying heat 0:13:55.920 --> 0:13:59.640 away to allow this to happen. And then like there's 0:13:59.679 --> 0:14:01.760 certain limits you just are not going to get past 0:14:02.960 --> 0:14:06.680 without damaging the processor. So overclocking is a thing that 0:14:06.880 --> 0:14:10.320 is possible, and a lot of people do it, and 0:14:10.480 --> 0:14:12.520 a lot of gamers do it in particular because it 0:14:12.559 --> 0:14:15.120 allows them to get even more performance out of their machines. 0:14:15.640 --> 0:14:17.480 But you need to know what you're doing so that 0:14:17.559 --> 0:14:20.680 you do it in a way that's not gonna wreck 0:14:20.760 --> 0:14:23.280 your computer rail of the gate and have to you'll 0:14:23.320 --> 0:14:25.600 have to go in and like maybe replace a CPU, 0:14:25.680 --> 0:14:28.840 or if you're really unlucky, you might have to replace 0:14:28.880 --> 0:14:32.560 a whole motherboard along with the CPU. So that's something 0:14:32.560 --> 0:14:36.040 to keep in mind. But yeah, overclocking is a way 0:14:36.080 --> 0:14:38.560 to get your computer to do more than what the 0:14:38.640 --> 0:14:42.120 manufacturer intended. Now, in some cases, in fact, this has 0:14:42.120 --> 0:14:46.560 happened in the past. Manufacturers have made processors like the 0:14:46.640 --> 0:14:51.640 same processor chip and specifically gone in and set different 0:14:51.760 --> 0:14:56.840 versions of that chip with different limitations of its clock speed. So, 0:14:57.000 --> 0:15:01.520 in other words, there's an artificial limit, like an artificial ceiling, 0:15:01.720 --> 0:15:05.600 to how fast that processor can work. You could go 0:15:05.640 --> 0:15:08.720 out and buy one version of the CPU that's maybe 0:15:09.080 --> 0:15:13.560 limited to one point five giga hurts, and then you 0:15:13.600 --> 0:15:15.560 go and you get to another one it's one point 0:15:15.600 --> 0:15:17.680 seven five, and then a third one and it's two 0:15:17.800 --> 0:15:21.160 point zero. And it's possible that all three of those 0:15:21.200 --> 0:15:26.080 processors are from an architecture standpoint identical. The only thing 0:15:26.120 --> 0:15:28.520 that's different is that the manufacturer has put a different 0:15:28.520 --> 0:15:30.640 limit in there and set a different price tag on 0:15:30.720 --> 0:15:35.320 each of those chips, and it makes it really cheap 0:15:35.400 --> 0:15:39.800 to produce them, and you're just artificially limiting what the 0:15:40.080 --> 0:15:43.560 hardware can do. So there are ways of getting past that. 0:15:43.760 --> 0:15:46.880 And I know that there are a lot of computer 0:15:47.000 --> 0:15:49.360 enthusiasts out there and when they find out that kind 0:15:49.360 --> 0:15:52.120 of thing. They get really upset, which is understandable because 0:15:52.800 --> 0:15:56.680 you're thinking to yourself, if this device is capable of 0:15:56.760 --> 0:16:00.240 performing at a higher level that's still within the rate 0:16:00.320 --> 0:16:03.400 limit of the manufacturer for the you know, whatever the 0:16:03.440 --> 0:16:08.600 top of the line is, then it's so unfair that's 0:16:08.640 --> 0:16:12.520 being sold to me with brakes put on so it 0:16:12.640 --> 0:16:14.880 cannot reach that level. I'm going to remove the brakes. 0:16:15.760 --> 0:16:18.200 That has happened in the past. It's not something that 0:16:18.240 --> 0:16:21.360 necessarily happens frequently, but it does happen. But then there 0:16:21.360 --> 0:16:23.520 are also cases where someone just goes out and they 0:16:23.560 --> 0:16:26.240 buy a really fast CPU and they say, I want 0:16:26.240 --> 0:16:29.920 to make it more faster. Then they'll go in and 0:16:29.920 --> 0:16:34.320 overclock it. Overclockers often will experiment with other pretty wild 0:16:34.400 --> 0:16:38.280 stuff like liquid nitrogen cooling and that kind of thing 0:16:38.320 --> 0:16:40.840 in order to get the absolute most out of the 0:16:40.880 --> 0:16:43.760 performance of their machines. But then you're starting to talk 0:16:43.800 --> 0:16:47.880 about setups that are truly crazy, and you're not likely 0:16:48.440 --> 0:16:52.960 to encounter those things. One other thing I want to 0:16:52.960 --> 0:16:57.920 talk about with CPUs or processors in general, are flops. 0:16:58.240 --> 0:17:00.760 Now not talking about products that and do well in 0:17:00.800 --> 0:17:04.800 the market. I'm talking about floating point operations per second 0:17:04.960 --> 0:17:09.800 or flops. You'll often hear flops being thrown around as 0:17:09.800 --> 0:17:15.440 a measure of computational performance, and you know we have 0:17:15.480 --> 0:17:18.160 the clock speed is one version of this, well, flops 0:17:18.320 --> 0:17:23.760 is something similar. It's the number of floating point operations 0:17:24.040 --> 0:17:28.800 that the processor is able to handle per second. And 0:17:28.880 --> 0:17:33.240 this is a place where GPUs or graphics processing units 0:17:33.280 --> 0:17:38.160 often can outperform CPUs, So we look at flops. That's 0:17:38.200 --> 0:17:42.960 another way really of just measuring how powerful or fast 0:17:43.640 --> 0:17:47.080 a processor is. When we look at things like supercomputers, 0:17:47.480 --> 0:17:51.480 we are often referencing their processing power in terms of 0:17:51.520 --> 0:17:54.400 the number of flops like peda flops. You're talking about 0:17:54.640 --> 0:17:59.440 astronomical numbers here when you're talking about floating point operations 0:17:59.440 --> 0:18:03.360 per second. But it's just really another way of judging 0:18:03.480 --> 0:18:06.640 the processing speed of a processor. And as I said, 0:18:06.680 --> 0:18:11.520 it's more frequently something that we associate with GPUs than CPUs, 0:18:11.520 --> 0:18:14.119 but it is an important thing to keep in mind. 0:18:14.760 --> 0:18:18.960 So those are some real basic elements of CPUs. Again, 0:18:19.440 --> 0:18:21.800 you can go into so much more detailed that I 0:18:21.840 --> 0:18:26.080 thought that this would be a way to understand from 0:18:26.119 --> 0:18:28.520 a very high level what CPUs are doing and why 0:18:28.560 --> 0:18:31.600 they're important we can talk about other things in the future, 0:18:31.640 --> 0:18:36.000 like hyper threading, for example, to talk about how that 0:18:36.160 --> 0:18:39.359 plays into say a multi core processor. But I figured 0:18:39.400 --> 0:18:41.800 that would go a little beyond just a tech Stuff 0:18:41.840 --> 0:18:44.720 Tidbits episode. I'll probably do more of these in the future, 0:18:44.720 --> 0:18:47.920 where I'll look at things like memory and other elements 0:18:48.200 --> 0:18:53.960 in computers that are basic but often hard to understand 0:18:54.040 --> 0:18:56.720 from for a newcomer, someone who has maybe heard the 0:18:56.800 --> 0:18:59.560 term but they don't really understand what it means or 0:18:59.560 --> 0:19:04.240 how it fits into the overall picture of technology. And 0:19:04.280 --> 0:19:07.040 we'll do it beyond computers as well. We'll talk about 0:19:07.040 --> 0:19:10.960 other types of technology and have these little tidbit episodes occasionally. 0:19:11.720 --> 0:19:14.199 I hope you are all well. If you would like 0:19:14.240 --> 0:19:15.679 to reach out to me, the best way to do 0:19:15.760 --> 0:19:18.159 that is on Twitter. The handle for the show is 0:19:18.280 --> 0:19:23.520 tech Stuff HSW and I'll talk to you again really soon. 0:19:30.119 --> 0:19:34.800 Tech Stuff is an iHeartRadio production. For more podcasts from iHeartRadio, 0:19:35.160 --> 0:19:38.840 visit the iHeartRadio app, Apple Podcasts, or wherever you listen 0:19:38.880 --> 0:19:39.919 to your favorite shows.