Hambone Blues Jam

Home Decoration Tips

Mod-03 Lec-03 Instruction Set Architecture

Hello and welcome to today’s lecture on
instruction set architecture and in this lecture I shall discuss various topics such as what
do you mean by an instruction. Then, what do you mean by instruction set architecture,
classification of instruction architecture and as sport of the instruction at architecture
will find there are you have to perform various operations. So, there is classification of
operations then classification of operands instruction format addressing modes, then
evolution of the instruction set. Then, there are two bases classes of the instruction set
architecture RISC verses CISC discuss and finally, I shall discuss about the MIPS instruction
set architecture. So, let me start with by an instruction what
you really mean by an instruction you know whenever you want to talk to somebody you
have to know the language of that person. So, whenever you are writing a program you
are essentially communicating with the processer and you have to talk the language the processor
understands the language. The processor understands is essentially represented by instruction
set architecture, so an instruction can be considered as a word. Instruction is equivalent to a word, in conventional
you know language, and then instruction set architecture can be considered as the vocabulary
for the processors. So, you can say that an instruction can be consider as a word in process
language and question naturally arises what information an instruction should covey to
the CPU as we know in our traditional language a word coveys a meaning. So, similarly, an
instruction should convey something to the processer what it should convey let us try
to understand first of all an instruction can be represented by an instruction format.
Different components of the instruction can be represented by different fields, first
thing that an instruction should convey to the processer is the op code the operation
to be performed by the processer. You know the task of a processer is to execute instruction
one after the other obviously the instruction should tell what operation to be performed.
Later on we shall see the various type of operation that can be perform by a processer
like adipindict operations logical operations and like that.
Then, apart from the op code, the other operates or the other fields of the instruction are
number one is address of operant one. You know whenever you have to perform some operations
say let say assume you have to perform c is equal to a plus b question is where from you
will get a that is the address of operant of 1, then comes the question of where from
you will get b. So, address of operant two-third is you have to store this results somewhere
c value of c, so that is known as destination address. So, these are the four things and
instruction should convey to n to the processer, finally this instruction execution is over
what processer should do now processer has to phase the next instruction where from it
will get. So, that information should also be present
here, so address of the next instruction, so this are the different fields of an instruction
if everything is conveyed explicitly. Now, that means what I am trying to tell you if
all the information likes source address of operant one so address of operant to destination
address of the next instruction. If everything is explicitly specified as spot of the instruction,
and then instruction will be too long. For example, say suppose this op code requires
1 byte and address may require 2, 4 four bytes, let us assume it requires 2 byte for simplicity
2 byte, 2 byte and 2 byte. So, you can see even in this very simple format
it requires 9 bytes per instruction now, so 9 if 9 bytes is required for a single instruction
and as you know a program consist of a large number of instructions written one after the
other. So, a program will require very large memory size because each instruction is requiring
5 bytes if all the fields are specified explicitly. Then, they have to store in the memory and
as you know from the memory have to transfer to the processer whenever the execution take
place. So, not only the program will occupy very large memory because is each instruction
is long to fetch them from the memory to the CPU.
It will take long time because if the one size is 3, you know 16 bit, you can fetch
only 2 bytes at a time. So, you will require, I mean say for 9 bytes you will require at
least you know 5 cycles, 5 machine cycles to fetch it. So, each instruction fetching
will require long time in other words execution time will be long. So, storage requirement
is long process is time processing time is long if all the fields are specified explicitly
in that case what is the alternative is to specify implicitly rather than explicitly
how can it be done. For example, you can use a special register
known as program counter pc program counter and it is implied that address of the next
instruction is always present in the program counter. When the processer is initialized,
I mean whenever you put reset button then program counter will be loaded by the address
of the next instruction to be executed. Then, as one instruction is fetch and it is executed,
it will automatically load the address of the next instruction in the program counter.
So, program counter is considering to be this source of the next instruction, so you do
not require this field. This field is no longer required in the instruction
if you see is providing the address of this next instruction, so it is implied that program
counter will be provide the address of the next instruction. So, this field is not required,
now let us consider another possibility say you are using three memory locations to store
the sources source of the operant and distention address, what you can say that address of
operant one will be the destination address you make the assumption that address of the
operant one will be the destination address where the result will be stored.
So, in such case there is no need for this field destination address field is not required,
now the question of a 2 this two fields address of operant 1 and address of operant 2 can
you get read of them? You can get read of them in this way you assume that one of the
operands will be always will be always able available from the resistant known as accumulator
acc accumulator so accumulator will hold. One of the two operant always that means you
have to load the load one of two operant in the accumulator before you execute an instruction.
So, whenever you make that assumption maybe this particular address of operant 1 is not
required. However, you will require another field address
of operant two will be required as part of the instruction can you get read of this you
can you cannot really get read of this fully, but what you can do instead of specifying
the full address. You can specify the address in a in a various way, for example what you
can do you can use a register as the address of the operant two in such a case may be the
op code can be little bigger, say some a part of the op code can be used to specify the
register which will be the address of the operant 2 in such a case.
Again, you will be able to overcome this, I mean this field not be present in an instruction,
so you find that ultimately you have been able to have an instruction with only one
field which specify the op code. Remaining things can be implicit, implicitly specified
available from some special purpose register processers knows where from they will get
it. Also, what if you register, then the size will be small the reason for that is you know
size of the memories pretty long may be 64 kilo bytes minimum few megabytes for now a
days few gigabytes. So, you will require 16 beta address or 32 beta address as it happens
in present the computers. So, instead of 32 bit, if the number of register
available is maybe say 16 only 4 bites is required, so four bites can be provided as
part of the op code field. So, in such a case you will require a very small instructions
size this is how the instruction size can be reduced. So, based on this discussion,
we shall see you know different techniques have involved at the c modes and another things.
So, instruction is too long if everything is specified explicitly, so it requires more
space in memory. It requires longer execution time question
is how can you reduced reduce, how can you reduce the size of the instruction specifying
information. Implicitly, as I have already told by using program counter by using accumulator
by using general purpose register and stark pointer I shall discuss about it little later
to implement some special data structures, so this is the basic idea about an instruction. Now, coming to instruction set architecture
what it really means an instruction set architecture is a structure of a computer that a machine
language program are must understand to write a correct program for that machine. So, instruction
set architecture is equivalent to vocabulary of the processer which the programmer must
know and with the help of each he can write a program in a simple language or machine
language. So, what the instruction architecture defines, it defines the operation that the
processer can execute various data transfer mechanism and how to excess data either from
memory or from registers. Then, various control mechanism like branch
jump and so on, so it is essentially a contract between programmer and compiler and the hardware.
So, you have got hardware and software and instruction set architecture is the essential
a contract between the hardware and software. Knowledge of instruction set architecture
is important not only from the programmer’s perspective. That means if the programmer
does not know the instruction set architecture, he cannot write program in machine language.
Not only is that important, it is also important from another perspective from processer design
and impletion perspective as well. That means not only a programmer should know an instruction
set architecture, but the designer of the processer should also know it because design
is essentially the specification. It sharps at this specification to the designer because
this is what the processer has to do have implemented those instructions. So, this instruction
can be executed by that processer, so the instruction set architecture essentially subs
as specification to the designer of the processer. Now, let us focus on the programmer visible
part of a processer number one is registers that means a process with help of this registers.
I mean there is at the registers are available where data are located you can store data
you can access data with the help of the instructions. Then, the instruction set architecture also
provide addressing modes various addressing modes with the help of each data can be access
either from register or form memory. Then, instruction format you will see that
and I have already mention about the instruction format, he may call it at instruction format
the various fields of an instructions a were with the help of which can specify a various
things. Then, it should be also know exceptional conditions what do you mean by exceptional
conditions exception can occur in two ways, exception can be generated from the outside
world. For example, interrupt it can be generated by an I O or it can be generated by a by user,
so reset interrupted inputs are the external interrupts or exception are coming from outsides
of the processer. In addition to that, some exception are generated
from within the processer, so whenever in instruction is executed. There are may be
something wrong, how that code that is being executed may not be op code may be invalid
code because of wrong alignment of the memory. That can happen or while performing, you know
execution of an instruction there are some situation like divide by 0. So, those situations
are known as exception generated within the processer whenever an instruction is executed
by the processer. So, whenever that happens what happens if
something goes wrong that also has as to be you know specified by the I mean should be
known to the programmer. So, whenever this exception happens for example, if interrupted
occurs it knows that there is a interrupt of a sub routine there is the specified, specific
address to which it will jump. Then, all this to together represent this instruction said
what operation can be performed, so this are the different parts programmer visible parts
provided by the instructions set architecture. Now, there are various instruction set architecture
design choices for example, types of operation supported, you know the processor has to do
some data processing what kind of data the processer can process. So, this can be like
arithmetical logical data transfer control transfer system calls floating point operation
addition multiplication division and decimal addition, subtraction, decimal operation,
string operation, bit manipulation operation. So, these are the various type of the operation
which can be supported, so it is not necessary a processor should everything depending on
the application its design. One can decide that this is the subset of
the operation a process instruction set architecture will provide and accordingly the process has
to implementer. So, it depends on the application for which a processor design, then comes the
types of operant supported. So, types of operant means whether it is byte operant the operation
can be perform on bytes or it can be performed on 16 bytes or it can be performed on 32 bytes.
So, it can be byte character digit half word double word and also floating point number
you can see the operands sizes can vary. Therefore, format can also vary, some can
be the fix point operand’s and some can be floating point numbers then types of operand’s
storage allowed again where the operand’s will be stored obviously either in register
or on memory. These are the two alternatives available, so there can be various type of
storage facility like stack accumulator then registers can be up two types special propose
register, general purpose register, then memory by memory. We really mean the main memory
perform the process or axis various operant and also instruction, then as I have already
told there is the possibility of specifying either explicitly or implicitly.
So, implicit verses explicit operand’s in instruction and numbers of each, I have already
explain this particular feature how instead of specifying everything explicitly you can
specified some of the things in implicit manner then orthogonally operand’s. That means
whether each operation will support all different types of addressing modes like that so that
that defines the orthogonality of operands. Then, operands location and addressing mode,
so these are the various design choices of a instruction set architecture. Now, let us have some kind of classification
of instruction set architecture, so it is actually determine by the means used for storing
data in CPU. So, I have already told you can store register or memory for storing operants.
Depending on that, this instruction set architecture can be classified broadly into three types
like stack architecture accumulator based architecture and register based architecture.
So, there are three possible alternatives in case of stack architecture operands are
implicitly on top of the stack, so everything you are doing with the help of a stock stack
and you can access from the top of the stack. So, as you know stack is the data structure
last in first out type of data structure, so you can access always from top of this
stack and then it you can be accumulator at this architecture. So, as I have already one
operant is in the accumulator which is the special purposes register and others are elsewhere.
I mean it can be either in some register or it can be some memory locations and essentially
this is one register machine. That means whenever it is accumulator based machine, it is implies
that the processer has got only one register. That is why it is called one register base
machine and this is particularly available in older machine, you know in earlier years
the implementation of registers was very costly hardware was costly.
So, only one register was allowed to be provided as part of the processer and that is how the
accumulator base processor, where popular in the earlier of computers were subsequently
as we shall discuss that restriction was overcome. Then, the general purpose registers comes,
where operands are in the register or specific memory locations. So, it is a general purpose,
you have a set of register may be 16 or 13 register, they are general purpose in nature. So, you can access operands from the registers
or it can be also from specific memory location. So, you can see how the instructions set architecture
has involved over the years as I said in the early years it was all accumulator based architecture
like ESDSAC, IBM 701, this were single accumulated based or IBM 700 series. Back in 1953 these
were all accumulator based architecture that means you have only one accumulator and operands
is always available taken from the accumulator and result also stored in the accumulator.
However, this second operands can be taken from memory, and then came special purpose
register architecture. I have already mentioned about this special
purpose register like a program counter stack pointer and there are some more special purposes
register provided a from where you can access operands. Then, came the era of general purpose
register architecture, so where it can be register memory type.
That means one operands is taken from the register second operant is from the memory,
so these are known as register memory architecture. For example, process like IBM 360, DEC PDP
11, Intel 80386, these are all you know belong to this this class and they are essentially
you know one operant is taken from the register operant is taken from the memory. They belong
to the category of SISC architecture complex instruction set architecture. So, then came
subsequently the RISC architecture where it is register architecture that means both the
offers and taken from register and result is also store in the register, register architecture.
That means both the operant are taken from the registers and result is also stored in
the register. That means whenever you performing various operation at arithmetic and logical
operations the always the operands are taken from the register result is also stored in
the register. So, this is known as register, register architecture and this are also known
as load store architecture because you have to perform. You know loading of the register
from the memory with the help of explicitly instruction and stored the results from the
register to the memory with the help of storage instruction.
So, those this are also known as loads store architecture and the RISC processer belongs
to this category and they are also known has load store architecture like CDC 6, 6, 0,
0 MIPS processer DEC alpha. These are all register, register architecture processer,
so you see it is started with simple and then gradually it has become complex. As I have already told, we can do the classification
in three ways three ways stack based where the operands are taken from the top of this
stack top to elements in stacks. Destination is also top of this stack, everything is done
with the help of this stack in accumulator base the processer as I have already told,
and one of the operands is from the accumulator. Other operands is from the memory and result
is store in the accumulator and in case of register a general register, register memory
or register, register processer. It can be register or memory and destination also can
be register or memory that means source of operant and destination both can be register
or memory. For example, let me illustrate these three
architectures with the help of this very simple operation. Consider you have to perform this
simple processing c is equal to a plus b in case of stack machine this will be the 4 instructions
push a, push b add and pop c. So, this is how the operands will be taken and the processing
will be performing then in case of accumulator based machines, first you have to load the
accumulator from the memory. Then, you will be adding the one index from taken from the
accumulator and second operant is from the memory and you are performing the operation,
then storing the result in accumulator, then result can be stored in memory locations c.
So, this is accumulator memory accumulated type of processer architecture and then register
memory in case of register memory as we can see you have got the register r 1 which is
the general purpose register where you are loading one of the operands. Then, you are
performing the operations adding contain the register r 1 with contain the memory location
b and then you are storing the result. Here, the result is available in register
r 1 and result store in memory location c in the third instructions and in case of register,
register architecture you are always doing with the help of register. So, accumulator
is loaded I mean register r 1 is loaded with the first operands, and then you are loading
the second operands in register r 2. So, this is the essentially that loads store architecture
and then you are performing the addition operation result is stored in another register r 3 by
adding contain of r 1 with contain of r 2. Then, you can store the result with the help
with the help of store instruction in memory locations c. So, you can see this is the register,
register architecture or loads store architecture, so we have compare four different architecture
that is possible. Now, let us come to the classification of
operations the operations can be classified into four different categories, as you can
see data transfer data transfer means you are move that data transfer is taking place
from register to memory or memory to register. So, we can say the data transfer so you have
got some registers and memory, so this is your storage place. So, it will be either
transferring from register to memory, so it can be called store or you will be transferring
from memory to register load or from one register to another register. So, this is your normally
known as move, so this are the three possible transfer the data transfer that can take place
between the between register and memory or within the registers. Then, the data manipulation
operations data manipulation, so data manipulation operations can be broadly divided into arithmetic
operations or logical operations arithmetic operations are like add subtract multiply
divide. So, again you have got a different possibilities
sign or unsigned integer or floating point. So, you can have addition on unsigned addition
or signed addition or multiply sign multiply unsigned multiply or integral multiply or
floating point a arithmetic operations. So, you have got several alternatives, similarly
you have got logical like conjunction disjunction shift left shift right and or like that. So,
these are the various data manipulation operations where you are performing manipulation of the
data the value of data is changing. Then, there are some instructions which are known
as status manipulation what do you mean by status manipulation. In case of status manipulation instructions,
neither you are transferring data note transfer of data take place no manipulation of data
take place. That means neither the data is moved from register to memory or memory to
register nor you are changing the value, what happens this status of the processing changes.
There are some instructions like that, for example, set carry there is carry flag bit
as you known. So, set carry flag bit here the status of the processer changes or say
enable interrupt or disable interrupt. So, this are some of this status manipulation
instruction with the help of which status of the computation can be change, but neither
data transfer nor data manipulation take place. Then comes the control transfer instructions
which can be conditional branch like branch on equal branch on not equal set on less than
or unconditional jump. So, there can be various types of control transfer instruction, now
operation which can confirm, so this are the different classification of operation that
is provided by the instruction set architecture. Now comes the instruction format whenever
we are considering instruction format it is very important that the length of the instructions
should be multiple bytes why it is necessary to be a multiple bytes because you will be
storing instruction in memory. The accessible unit from the memory is byte that is why we
call it byte addressable byte addressable memory. So, your instruction has to be either
8 bit 1 byte or it can be 16 bites for 16, normally it is a 8 bit for eight bit processer,
16 bit for 16 bit processer. That means 1 byte or 16 bit 2 bits or it can be 32 bit
4 bits, so there are it has to be multiple bytes. So, you can access it from this point
and edit or from this point or form this point and you can get a complete instruction.
That is why I was telling that alignment is important if it is red from the middle for
example, whenever the instruction is 16 bit and if you read from here you will not really
get the instructions in a properly. So, whenever you try to execute it, it will give error
or exception, now instruction and coding can be variable or fixed there can be fixed format.
In other words, what I am trying to tell an instruction can be of fix size say 32 bit
fixed all instruction are of 32 bit. So, that is one possibility another possibility is
that it can be multiple of some bytes say one can be of 16 bit, one instruction format
another can be 32 bit and third one can be 48 bit.
That means in these three cases in this particular case that instruction can be of one bytes
size 2 bytes sizes or 3 bytes sizes. So, we call it variable format, so instruction we
will call it variable format, so you can either have variable format or fixed format particularly.
Later on, we shall see we can see this 6 processers have variable format, on the other hand the
RISC processor have fixed format in general that is true. So, variable format leads to
vary variable encoding tries use few bits as represent a program as possible, but at
the cost of complicity of the decoding. So, that means here the complexity of the decoding
is present whenever you use variable format. Later, we shall discus on MIPS instruction;
we will see that DEC alpha ARM MIPS instruction. They have fixed instruction sixes because
they belong to the RISC category this processers and this is the general instruction format
for this processer up code. Then, three addresses one, address two, address three, they can
be usually, they represent the register, but we should discuss about various format with
the help of the example particularly MIPS instruction. Now, address modes describe how an instruction
can find the location of the operands, so you can have varieties of addressing modes
to facilitate accessing of operands from registers as well as from memory. You will see that
either from register be that means you have to do calculation of the address from where
you will get the operant and that is called the effective address. So, effective address
is calculated whenever you are operands are in memory, what you can do that effective
address calculation may involve readings some value from register and some part of the instruction
that can be used two compute the effective address.
Later on I shall discuss about the how effective address is calculated in various situations
our in different addressing modes. So, effective address is essentially the actual memory address
this specifying specified by an addressing modes the mechanism by phase the effective
address is calculated varies from one addressing mode to the another addressing mode. Here, I shall discuss about the various addressing
modes, for example inherent addressing 0 addresses. So, here for example you have nothing but
up code, so rest of the things are implicit obviously one of the operands is in the accumulator.
So, it can involve a an accumulator based instruction since such a case say it can be
say implement of a value implement the content of the accumulator or compliment the content
of the accumulator. In such a case it involves only one operator, so if it is in the accumulator
that is implemented or inverted and result will be also in the accumulator.
So, there is no need to have any other operands address, so that is why this are called this
is known as inherent a 0 address modes only the up codes need to be explicitly specified
rest of information are implicit. Then, an immediate addressing the operand itself profile
is part that mean you have got op code and first operand may be in the accumulator, second
operand is provided as part of the instruction an explicitly. That means you will add the
contain of the accumulator, I mean if it is add operation with the with the operands which
is provided as part of the instruction. So, this is the second addressing mode third
addressing mode is known as absolute or direct addressing mode in such a case as you can
see the address of the operands particular address of the second operand first operand
is a. Assume in the accumulator address of the operand second is available from the memory
and that address is explicit the specified and as part of the instruction, so that is
why it is called absolute or direct addressing. Then, you can have to incorporate little bit
of flexibility, you can have indirect addressing in indirect addressing instead of specifying
the address from the memory, sorry instead of specifying address. Explicitly, what you
can do that address can have the address of the operands instead of the operand as you
can see here in this case this is the address, but this the address does not contain the
operand. The address contain the address of the operand, so you got a kind of indirection,
so this address is pointing to another memory equation which is the address of the operand
so that a 2 is the address of the operand. So, again you have to read from memory location
a to locate the operand, so this is known as indirect addressing essentially to provide
little bit of flexibility in addressing. Then, you can have register addressing as I have
already told some parts of the op code your small portion of the op code, I mean the that
instruction the op code is little smaller. Here, you can provide the register name register
number and which will give provide the operands, so here is your register, set of register
or register bank whatever you call it. This is pointing to the operand, I mean register
where the operand is stores, then you can have register in direct.
So, just like your indirect addressing here also that register is not having the operands
but, is having the address of the operands. So, you can sometimes use the concept of register
appear because we are single register may not able to hold the entire address. For example,
you are register is 16 bit, but address is of 32, so you require a pair of register in
profile the full address. So, such a case we use the concept of register pair, so as
you can see here a register pair is holding the address and that address can be used to
phase the operand. So, that means effective address here is coming from the registers,
so this is how in different situation the effective address generated as you can see. Then comes the paged addressing, sometimes
we use the concept of paging and particularly to reduce the size of the instruction we use
different pages or you can use a page register. So, in such a case an offset is provided and
then direct page register is used this two together provides the full address where operands
is available. So, this is known as paged addressing that is done with the page of special purpose
register which is available in the as part of the processer. Then, index addressing is
used to facilitated implementation of data structures like a q and as you can see we
use a special purpose register x. So, this register gives you the upset value
and another register base address is provided as part of the instruction this together is
used to generate the address and you get the operand here, this is very useful for accessing
ray implements. So, the index register will provide that base register, this base address
is the starting address of clearly and then index register will join to different elements
of the ray. So, base of this addresses fixed and then you can change the index register
value to point to different a elements. If its 0 its point to the 0, 8 elements, if it
is one it points to be first element like that, so in this way you can access and a
ray a very convenient by index addressing I said that this is useful implement some
data structure like r a q and so on. Then comes the based addressing based addressing,
where a special purpose register is none as based register is used and the constant is
added to that to point to the operands. So, this particular earlier what have you seen
that register value was changing, now this value will change, I mean this base register
is added with the constant to generate operands at this.
This is particularly used in situation like re location, now sometimes you have to do
relocation in a multi user environment. Then, you require best addressing to change the
content of the based register to relook at the instruction to different parts of the
memory. So, then you can combine based indexed that in this you can have indexed register
and a based register both of them can be used simultaneous to have based in addressing.
Then, you can have relative addressing relative addressing is you are essentially providing
an address with respect to some register particularly program counter with respect to program counters.
So, op code and the displacement is provided as part of the instruction which is added
with the program counter to generate the effective address and which is the address of the operand.
So, this is known as relative or when you call the program counter register is used
then you call it PC relative. Then, there are various addressing mode related
to stack, so you have got a stack pointer which is pointing to the bottom of the top
of this stack and you know whenever you post some element and it is decrement and that
is stored. So, you can perform push a, so you can sit this pointing to it is static
point requirement to point to the next locations and value of a is stored here. Then, another
push you can see it is again implemented and top of tag his having the much holding the
value b. Similarly, you can perform pop operation, so with the help of the push and pop operation
you can store or load operands from in a memory. So, whenever you do pop operation you do can
read it from this location then it will point to next location. So, in this way you can
see stage you can do the help of this two instruction push and pop. So, these are the various addressing modes,
but our discussion will not be completed without the mentioning about RISC and CISC controversy.
I have already told that there are two possible architecture, one is known as RISC instruction
set computer is architecture another is complexes instruction set computer or RISC architecture
and what is the genesis of a CISC architecture. So, implementing commonly using instruction
in hard ware can lead to significant performance benefits that is means what you are trying
to do you are trying to minimize the semantic gap with high level languages. That means
whatever the high level language instruction can perform you are trying to do with the
help of machine language and instruction by making the instruction more and more complex.
As a result single high level language instruction will lead to very few machine language instructions
may be one or few. So, this semantic gap between the high level language and machine language
reduce that is the basic idea behind complex instruction set architecture. So, what is
happen as the as the various program in language involved and various operation which are provided
statements can performed in high level languages those are implemented a complex instruction.
So, that is the genesis of the this architecture for example, use of a 14 point processer can
lead do performance improvement and they will be complex 14 point operation which can be
specified with the help of complex instruction. On the other hand, in case of RISC architecture
the rarely used instructions can be eliminated to save chip space on chip case cache and
large number of registers can be provided. Here, there is a story behind it particularly
IBM did some study on various type of instruction which are used by the compiler. You know you
are writing a program in high level language then compilers is generating the machine language
code. So, which instructions are used by the compilers different compilers, it was found
that a lot of stimulation works were carried out by IBM before. It was found that very
complex instructions are really used the other compilers use only simple instruction for
generating the object code. So, what was decided the instruction which
is really used by the complier try to implement them. So, they remove that complex instruction
leading to what is to known as RISC processer. Here, complex instruction where moved and
only a simple instruction which are commonly used by the complier are retained and that
is the genesis of RISC architecture. These are the features of rich architecture
rich instructions set, some simple some can be very complex addressing modes. Then, many
instruction taken multiple cycle as I told they can be the instruction format can be
complex and I mean variable and there can be a large variation of CPI, some instruction
may take 1 cycle, some instructions may take 2 cycles. They may take a 10 cycles depending
on the complexity of the instructions, then instruction of variable sizes I have told
then small number of registers because it uses primarily a resist.
It is a register memory architecture, so one operands in the register other operands is
from the memory. That is why you require small number of registers and micro code control;
you know that the control unit is implemented with by micro programming micro programming
control unit. Instead of hard ware control unity you may have studied there are two ways
of implementing control unit one is hardware control unit and another is micro program
control unit. Normally, this CISC processer use micro program control unit and it is very
difficult to implement pipelining in CISC processer. So, obviously you one instruction could do
work of several instructions because this quite complex and there will be many variation
of RISC instructions because it involves memory and register. On the other hand, the RISC processer have
small number of instruction small number of addressing mode large number of register this
is one very important features of RISC processer. They requires very few large number of register
and instructions executive one or two cycle clock cycles you will find later on when I
shall discuss the MIPS processer architecture, you will see all the instructions will require
a single format. Also, it will involve only 1 cycle or 2 cycles for these executions,
then uniform length instruction and fixed instruction format as I already told and is
essentially register, register architecture where you require separate load and storage
instructions. That is why this called load store architecture
and separate instruction and data cache this is a very important feature. Later on, I shall
discuss on little more detail, you will require two separate cases cache memories one for
instruction one for data instead of a single a cache memory. Then, the control unit has
to be hardwired instead of micro program control unit it has to be hardware control unit. So,
these are the features and finally, I shall discuss in detail later on when we shall discuss
about pipelining. You will see the RISC processer can be RISC
instructions can be very easily pipe line compare in contrast two CISC processer. So,
implementing pipeline architecture for RISC processer is very difficult which can be very
easily done in case of RISC processer. So, with this let us come to the end of instruction
set architecture in by next class I shall continue our discussion on RISC, CISC, and
also consider a representative processor architecture that is the MIPS processor architecture in
my next class. Thank you.

Leave comment

Your email address will not be published. Required fields are marked with *.