[MUSIC] So we just heard from Gaetano how to encode integers and floats. And now in this new section we're going to start looking at machine codes. Okay, so by machine code recall, we saw in the, in the first section, are this zeros, and ones that tell the computer what to do when executing the program. Okay, and that's going to, we also going to start looking at an introduction to assembly language. And what you have to understand, in terms of basic machine organization, in order to learn about assembly programming. And learn how a computer work works down a little bit lower level. Okay? So here's what we're going to see in this section. We're going to see what instruction set architecture is, which is the heart of software interface. It's the contract between hard and software. We going to look at briefly at the history of Intel processors and architectures. And then we're going to see how C in assembly code machine code relates, and it's going to start looking at x86 basics. So, we saw this picture in the very first section, and if you recall, this is the lifetime. The, the , the life time of a program starts with coding time when you write the code. Compile time, when you translate from the high level language all the way down to machine code, 0s and 1s that compose what the computer understands. Okay? And then run time is what happens when you actually executing the program. Okay? Now also recall that, what makes programs fast is well, your algorithm, how you write your code. Okay? what the compiler does in other words, how the compiler translates from C code to the To the hardware, software interface, to DISA, the instructions that are chosen with the compiler. And then the assembler converts that to 0s and 1s, to, to machine code. And then, you know, depending on how the hardware executes it, it's going to make your program run faster, okay. So just, just to review, repeat what I just said. You know, the, the time required to execute a program depends well, on the program itself. What algorithm you've used, how the compiler translates that code written by the program into assembly instructions. It also depends on the instructions of that architecture, which is a set of basic instructions offered by the processor to the compiler, or to whoever's writing assembly code. And then, ultimately, of course, the program performance is also a function of the actual hardware implementation. How's the hardware itself how the hardware itself is organized to execute those instructions. Okay, so now let's spend a little bit more time on understanding exactly what the ISA is. OKay? So the ISA defines three things, three important things. The first one is a system state. What defines the state of your system, okay, of your imperial system. That includes the contents of registers, which is as Gaetano told you you know, a basic unit of storaging inside the CPU itself. Okay? It includes the program counter, which tells what instruction's being executed at that time in the, in the processor. And also, it includes all of the contents of memory. Everything that's stored in registries, memory and including the program counter is part of the system state. Okay? So the ISA also defines the instructions that you execute, that the CPU can execute in the most basic level. Think of that as very, very simple instructions, things that do like, an add or a subtraction or say, a load from memory. Okay that gets there from memory in to a register as another example, of a, of an instruction, okay? And finally, the ISA also defines what exactly happens when these instructions are executed. Because one, one of these, when one of these basic instructions are executed, it changes the state of your computer system. Okay? So, and how this state Changes his function of, also function of what the ISA specifies. Okay? So here are the major decisions when designing an ISA. First of all, obviously, what instructions are available like: add, subtract, multiplication, floating point operations and so on. Okay? And what do they do? Exactly how they change the system state, and also how they are encoded. Because if you recall, when say you have an add instruction in assembly, it becomes some sequences of zeros and ones, that's, that's what the computer actually understands. And encoding is actually this mapping between the instruction in the ISA and the actual bit sequence that determines that instru-, that that encodes that instruction. So the ISA also tells how many registers your computer has. You know, how many, you know, explicit units of storage you have inside your processor. Okay? So how many you have available, and also, how wide they are. Right, for example, are they 32 bits, are they 64 bits, or are they smaller, like in case of a very simple processors and so on. Okay? And finally, the ISA also defines something very important, which is, how do you specify a location in memory? Okay? So how do you specify an address in your assembly code, or how in, in your instructions? Okay? How do you tell instructions what addresses they should be manipulating. Okay? It's also called you know, what are the addressing modes supported. Okay? So let me write it here, addressing modes. The, all the ways that it can specify an instructions in an address in your program. So now x86 is one type of ISA. Okay? so it's very, very popular. Okay? So that's and in fact it's so popular that x86, which is, you, you probably know, Intel Intel Corporation makes x86 processors, or processors that implement x86 ISA. And these processors completelly dominate the server, desktop and laptop markets for now. And they actually had a very revolutionary design. You know, in fact if you have code written 8086 all the way when it was designed in 1978, in fact very close to when I was born. so code written for the 8086, still runs in the most modern Intel x86 processors. Since we had very, very, evolutionary design, and more features were added as time, as time went on. An x86 ISA is something that we, is a type of[UNKNOWN] that we call CISC which stands for Complex Instruction Set Computer. Okay? What that says is that the instructions, the basic, the basic instructions specified by the x86 ISA are very rich. Okay? They can do things like you know, move a, move a string in memory. They do things that are much more basic than just a simple add or subtraction or move a little bit of data here and there. These things are much richer than that. Okay? So, and we call it CISC, that is a Complex Instruction Set Computer. As opposed to RISC, which stands for Reduced Instruction Set Computers. Okay? A RISC machine has much fewer instructions, and instructions are much much simpler. Okay? So, in more advanced computer architecture courses, you will see that this has a lot of implications of how the hardware itself is designed. Okay? But it's a little bit out of the scope of this class to give you the details of why whether an ISA being RISC or CISC how, why and how it affects computer design. So, le, let's look at some numbers here just to, to give an idea of how amazing this evolution was. Okay? So if you go back to 1978, when the 8086 was introduced. That processor had 29,000 transistors, and it ran on between five and ten MHz. That's the cycle time of the processor, between five and ten Million cycles per second. Okay? At that time that was the first 16 bit processor. That means that the word size was 16 bits or two bytes, and it was the basis for the I, the IBM PC and DOS. It really popularized personal computing. That's a very very important processor. And the total address space, a maximum memory you could you could address was one MB. Can you imagine that? It can't even hold a photograph. [LAUGH] But and then another major step was that the, the 386 that was introduced in 19, seven years later in 1985. And it had ten times as many transistors, it ran about you know three times faster in it's clock's frequency. But it was significant for a number of things. Okay? It was the first 32-bit microprocessor. Okay? It was referred to as the IA-32. you had a flat addressing mode. In that, it didn't have to divide the me-, the, the memory into regions called segments and addressed them individually. You had a single address that could point anywhere in memory. So, that made the process of managing memory and accessing memory much, much simpler. Okay? So, and this machine is also capable of, capable of running an operating system called Unix. Which is very similar to what we call Linux today. Okay? And so it's very similar, Linux is a type of Unix Operating system. Okay? And and today, a 32 bit Linux GCC targets, targets the 386 by default. So the 386 are very, very important architecture as well. Now let's jump 20 years later and look at what happened. Now with the Pentium 4, when, when it was introduced in 2005, had 230 million instructions. We're talking about a 10,000, 10,000 fold increase in the number of transistors. Okay? And, you know? A sim-, in about, about a a huge increase in, in frequency as well. Okay? And then the Pentium 4F was the first 64 bit processor as,as also introduced by first 64 bit x86 processor, introduced by Intel. It's also referred to x86-64. The AMD was actually involved in, in defining what the x86-64 looks like, which we're going to see in a second. A little bit more history of Intel processors, the 486 was introduced in 1989, was also a very important processor. Then there was a series of Pentium processors that hit the that had a lot of evolutions. In-, including like around this time, they were starting to worry about power consumption as well. So and then, it was a Core 2 Duo, you know, which was a dual core processor. A processor that had multiple multiple compute cores in the same chip. And then, more recently, we had the i7, that for example, here we have a dipex. It's a picture of the actual silicon of the i7, and it has four cores. It is essentially, you know, these are exactly, each, each one of these boxes here, okay. Are each, each one of these boxes is a processor, and you have four of them on a chip, and they all share a bunch of cache memory. That we, we are going to see exactly what that means in the later sections of this class. Okay? So, and along this history, there are a lot of other things that were introduced along the way. One was support for multimedia operations. Essentially, multimedia operation tends to have a lot of fine grain pair at instruction level, we called it vector type of parellism. So, in order to take advantage of that Intel introducedt an exception the ISA called SIMD. Which stands for Single Instruction Multiple Data, and that that's that instruction actually operates in multiple data items simultaneously. So taking advantage parallels that the application has. Okay. So and also, along this way like in this ISA evolution, not only the ISA architecture evolution. There were instructions that were introduced to do more efficient condition operations. Okay? We'll see later, in this class, that there are these instructions called branches that changes the flow of execution to the processor. And these instructions can be very, very expensive. So in part of the innovation that happens in this evolution of microprocessor was to make some of that much cheaper. And then, obviously, as I mentioned before, there are lots of cores introduced along the way. Okay? So if you're interested in learning more about the specifics of of Intel's architecture. You can go to the automated relational knowledge base great name. that's available in Intel website. And also you can look at the list of Intel micro processors in Wikipedia, which is great to get more information as well. But now, let's not forget about AMD, a company called Advanced Micro Devices that made clones of x86. By clones it means other processors that had completly different hardware, but also implemented this same instructor set of architecture. That's one example of the beauty of abstractions. See it had the same abstraction call the instruction, the instruction for that architecture, the[INAUDIBLE] interface was the same. The implementation underneath doesn't really affect the software as long as the ISA stays this, the same. And so AMD came up with completely different hardware that implemented x86 ISA, and it had some advantage at first. You know, they were a lot cheaper and were a tiny bit slower but it made sense in terms of, of, of price performance. OKay. Typically just followed behind Intel but then AMD just got really, really aggressive in terms of taking advantage of new secret design techniques and they designed very, very, very fast processors. And you know one great example is the AMD Opteron which is really, really tough competitor. To the Pentium 4. And AMD, was also instrumented, was also instrumental in developing the x86-64. Which is a 64 bit extension of uh,of the x86 ISA, that later on actually became a good part of what x86-64 looks today, even for Intel processors. It has really, really interesting you know, quote, unquote collaboration between two companies that are essentially competitors. Okay? So now to to end this mod-, first module of the section, just to tell, we are going to cover now this section in, in x86 program, which will be the, the following section. we are going to look at what's called IA-32, which is the traditional 32 bit, 32 bit ISA. And um,[INAUDIBLE] here, 32-bit[LAUGH] , 32-bit ISA. And then we're going to also talk about the 64-bit version of the x86 ISA. Okay? And keep in mind that all lab assignment's actually going to use x86-64. But in teaching you assembly programming, it, it, it's going to be larger transparent to you. We you're going to know when it matters, okay? Thank you and see you soon.