เกี่ยวกับตอนนี้
Ever wonder what’s really doing the thinking inside your computer or phone? Let's talk about the CPU.
Become a Member
https://members.sideofburritos.com
Don't just host your podcast, own it
🟡 Yellowball is a podcast hosting service I built to give podcasters more control and less nonsense. Whether you’re starting fresh, moving your show over, or just curious, check it out at https://yellowball.fm.
References
https://newsroom.intel.com/tech101/the-transistor-explained
https://www.computerhistory.org/revolution/digital-logic/12/285
https://www.computerhistory.org/revolution/digital-logic/12/284
https://www.intel.com/content/www/us/en/history/virtual-vault/articles/the-intel-4004.html
https://ieeemilestones.ethw.org/Milestone-Proposal%3AIntel_4004_Microprocessor
https://en.wikipedia.org/wiki/Central_processing_unit
https://en.wikipedia.org/wiki/ARM_architecture_family
แสดงหมายเหตุ 🔗
สำเนาบทสนทนา 🔗
00:00
I had my first experience with computer hardware when I was 8 years old.
00:05
My dad had brought home an old computer from work that wasn't being used anymore.
00:10
We took it apart, dove inside the case, made sure to avoid touching any capacitors, and
00:16
just looked around.
00:18
We took out the floppy drive, the motherboard,
00:19
sticks of RAM, and the hard drive. We even took the case of the hard drive apart
00:24
so we could look at the shiny platters inside. To this day, that's probably one of my favorite
00:31
pieces of computer hardware. There's something mesmerizing about those perfectly reflective
00:36
platters. And the second you touch them, your fingerprints are all over them. Fun fact, there
00:42
are some powerful rare earth magnets in there that you can take out and repurpose. Just be careful
00:48
you don't pinch your fingers between them. So that was the first time I ever took apart a computer,
00:53
but building one, that happened in college. It was my second year, and I wanted a server to run
01:00
virtual machines on from my dorm room. My college had a slash 16 network, which means every network
01:07
jack you plugged into had a public IP. You could plug in a computer, and boom, it was accessible
01:13
on the internet. So I specced it out and placed my Newegg order. This was before Newegg was acquired
01:19
by a Chinese company. Then for the next few weeks, as parts came in, I would hike up the hill to the
01:26
mail room, pick up each package, and make the one mile walk back to my dorm room. A mile in both
01:32
directions with each individual component really makes you understand just how many different parts
01:38
you need for a computer. Once all the parts came in, I got it assembled and up and running.
01:45
So in this season, I want to cover each main component of a computer. We all use them every
01:51
day, but do we really understand them? The component that does everything you ask it,
01:56
on this episode of In The Shell. The wrong thing to do is just go out and buy a computer
02:03
and then learn about it. You'll learn, but you'll learn a lot of things that maybe you didn't want
02:08
to learn. A computer that you buy today will likely be obsolete six months from now, and there's not a
02:15
dang thing that you can do about it. My name is Josh, and I'm able to keep this podcast independent
02:22
and advertisement-free because of support from listeners like you. If you are finding value in
02:28
what I'm doing here, consider becoming a paid supporter at members.sideofburritos.com. And as
02:35
a thank you, members get early access to new videos, ad-free versions of everything, bonus content,
02:41
and access to a live monthly Q&A. Thanks for considering. Now let's get back to the show.
02:48
At its core, pun intended, the CPU is a processor. It processes instructions. When your computer runs
02:56
a program, or even just boots up, the CPU fetches instructions from memory, decodes them, and executes
03:03
them one after another at lightning speed. This cycle of fetch, decode, execute repeats billions of
03:11
times per second, synchronized by the CPU's internal clock. It also has a small set of specialized tools
03:18
to help get its job done. One tool is the Arithmetic Logic Unit, ALU, which handles basic math and logical
03:26
comparisons. Another is the Control Unit, which manages the flow of data, telling memory and input
03:33
and output devices where to send or receive information. The CPU also uses tiny storage areas
03:40
called registers to keep data it's actively working on. Modern CPUs often include multiple processing
03:48
units called cores on one chip. A quad-core CPU has four cores that can work on tasks in parallel,
03:57
allowing the CPU to handle multiple tasks at the same time more efficiently. Despite all these
04:03
advancements, the CPU's fundamental job remains the same, to execute instructions and process data.
04:10
But how does this tiny chip actually switch and calculate so incredibly fast? The transistor.
04:17
A transistor is essentially a microscopic electrical switch. It can turn
04:22
on or off to either allow electrical current to flow or to block it. This on-off behavior
04:28
is exactly what we need to represent the ones and zeros of binary code, the language of computers.
04:35
Each transistor can be in an on state, allowing current representing a one,
04:40
or an off state, blocking current representing a zero. By arranging billions of transistors
04:46
into complex circuits, engineers create logic gates and high-level components that perform all
04:53
the calculations and decision making inside the CPU. For example, combinations of transistors form
05:01
logic gates like AND, OR, and NOT, which in turn combine to implement everything from addition
05:07
and subtraction to multimedia processing. It's incredible to realize that even the most advanced
05:14
computation,
05:14
ultimately boil down to transistors switching on and off. Historically, transistors replaced
05:21
their larger predecessors, vacuum tubes, in the 1950s and 60s. Early computers like the 1940s
05:29
ENIAC used around 18,000 vacuum tubes as switches, and those tubes ran hot and burnt out frequently.
05:37
Transistors were first demonstrated in 1947 at Bell Labs. They were tiny, used far less power,
05:45
ran cooler, and were more reliable. Over the next decade, engineers dramatically improved
05:52
transistor designs. By the mid-1960s, a new manufacturing approach, the planar process,
05:59
made it possible to build many transistors on a single slice of silicon. This led to the first
06:05
integrated circuits and eventually...
06:07
the first microprocessors in the early 1970s. For perspective, Intel's first CPU, the 4004
06:15
in 1971, had about 2,300 transistors, compared to modern CPUs that can house tens of billions
06:23
of transistors on one chip. This rapid growth in transistor counts followed a trend known as
06:31
Moore's Law, the observation that roughly every two years, the number of transistors on a chip
06:37
doubles, which I talked about more in Season 2, Episode 2 on Gordon Moore. When you hear that a
06:44
CPU runs at 3GHz, what that actually means is that its internal clock ticks 3 billion times per second.
06:52
Each tick is an opportunity for the CPU to move one tiny step forward in the fetch-decode-execute cycle.
06:59
One other important aspect of CPUs is their architecture.
07:03
CPU architecture means the basic set of machine instructions a CPU understands to carry out fundamental operations, things like arithmetic, logic, and moving data around.
07:16
Over the years, different families of CPUs have developed their own languages.
07:21
The most common architectures you'll hear about today are x86, x86-64, which is x86's 64-bit extension, and ARM.
07:32
Modern Intel and AMD desktop CPUs use the x86-64 architecture, which is basically a 64-bit evolution of the older 32-bit x86 standard.
07:44
The jump to x86-64 around the early 2000s was a big deal because it expanded the CPU's capability.
07:52
Older 32-bit x86 chips could only directly use about 4GB of RAM, whereas 64-bit CPUs can address
08:02
vastly more memory, theoretically up to 16 exabytes. Another reason CPU architecture is a big deal
08:10
is the difference in performance and efficiency. x86 and x86-64 chips, the kind in most desktops
08:17
and laptops. They have complex instruction sets, often called CISC, spelled C-I-S-C, which include
08:26
many built-in operations. These CPUs tend to be very powerful and capable of handling heavy workloads
08:33
like video editing, gaming, and software development, but they also use more electricity
08:39
and run hotter as a trade-off. On the other hand, ARM CPUs use a reduced
08:44
instruction set called RISC, spelled R-I-S-C, which is focused on efficiency.
08:45
.
08:45
.
08:45
.
08:51
ARM chips are designed to do more with less. They draw far less power and run cooler,
08:56
which is why they dominate in smartphones, tablets, and other battery-powered devices
09:01
where energy efficiency is crucial. An ARM-based phone can run all day on a small battery.
09:09
Historically, the trade-off was that ARM's CPUs weren't as powerful as their x86 counterparts.
09:15
They handled everyday tasks just fine, but weren't geared for the really demanding stuff.
09:21
However, that line is blurring today. A great example is Apple's M1 chip and its successors
09:28
used in newer Macs. It's an ARM-based CPU that proved ARM chips can be both extremely powerful
09:35
and power-efficient.
09:37
In some cases, Apple's ARM CPUs are on par with or even outperformed traditional PC x86 chips while still running cool.
09:53
And look at how the CPU actually fits into a PC when you are building one.
09:58
Physically, a CPU is a small squarish chip that needs to be installed onto the motherboard.
10:03
The motherboard is something we'll be covering in detail in a future episode.
10:08
The motherboard has a special CPU socket, a slot that is designed to hold the CPU, and connect it to the rest of the system.
10:16
Every motherboard model supports only a particular type of socket, and a CPU will only work if it matches that socket on the motherboard.
10:26
This is why, if you pick an AMD Ryzen CPU,
10:29
You need a motherboard with the correct AMD socket for that chip.
10:33
Once the CPU is seated in the motherboard, the next crucial step is keeping it cool.
10:39
As powerful as these chips are, they generate a lot of heat.
10:43
Without proper cooling, a CPU would overheat and shut down, or throttle itself dramatically to avoid damage.
10:51
So in your PC build, after the CPU is in the socket, you'll attach a CPU cooler on top of it.
10:57
This might be an air cooler, a metal heatsink with a fan, or a liquid cooler, but either way it serves the same purpose, to carry heat away from the CPU.
11:09
Thermal paste is applied between the CPU and the cooler.
11:13
It's a tiny dab of gray stuff that ensures there are no air gaps and maximizes heat transfer between the CPU's surface and the cooler's surface.
11:21
base. When done correctly, this cooling setup keeps the CPU at a happy operating temperature
11:28
even under load. As I was going through my script for this episode, there was one question I kept
11:35
asking myself that I really wanted to know, and that's why are these tiny transistors made of
11:41
silicon? Silicon is one of the most abundant elements on earth. Common sand is largely silicon
11:48
dioxide, which means it's cheap and widely available. But more importantly, silicon is a
11:54
semiconductor, meaning it can act as an electrical conductor or an insulator under different conditions.
12:01
In pure form, silicon doesn't conduct electricity well. However, by doping it, introducing tiny
12:09
amounts of other elements, like phosphorus or boron, we can tweak its electrical properties.
12:14
Doped silicon can carry current in a controlled way, creating regions that either have excess electrons, N-type, or excess holes, P-type.
12:24
Put an N-type or P-type silicon together, and you get a transistor that can switch on or off at your command.
12:31
Additionally, silicon is friendly to manufacturing.
12:35
It can be refined into extremely pure, single crystals, and then sliced into thin wafers you might have seen in photos of chip factories.
12:45
Those wafers are the platform on which billions of transistors are fabricated, using processes like photolithography.
12:53
The abundance of silicon means we can produce these wafers in huge quantities relatively cheaply.
12:59
And the material stability and reliability mean the resulting chips can run for years without degrading.
13:06
It really is amazing when you stop to think about it.
13:10
A CPU is this small piece of material that sits quietly inside your computer, yet it does so much.
13:17
That old machine I took apart with my dad felt so powerful,
13:21
yet today's CPUs run thousands of times faster and pack far more complexity.
13:30
In the Shell is written, researched, and recorded by me,
13:34
the ARM-based podcaster,
13:37
if you are listening in an app that lets you rate shows,
13:40
please take a minute to rate this one.
13:43
I would truly appreciate it.
13:46
I always heard that if you're hiking in the woods, bring a piece of fiber optic cable with you.
13:53
If you get lost, you can bury it,
13:55
and 15 minutes later, five backhoes will be there to dig it up.
14:01
That's it. Take care, and I'll see you next time.