Detailed explanation of super complete knowledge points of instruction system

Command system

One 、 Machine instructions

Instruction system is the language system of computer hardware , Also called machine language , Refers to the set of all instructions that the machine has , It is the main interface between software and hardware , It reflects the basic functions of the computer . From the perspective of system structure , It is the main attribute of the computer seen by the system programmer . Therefore, the instruction system represents the basic functions of the computer and determines the required capabilities of the machine , It also determines the format of instructions and the structure of the machine . Designing instruction system is to choose some basic operations in computer system ( Including operating system and high-level language ) Should it be implemented by hardware or software , Selecting some complex operations is implemented by a special instruction , It is still implemented by a series of basic instructions , Then determine the instruction format of the instruction system 、 type 、 Operations and access methods to operands .

An instruction is a statement in machine language , It's a meaningful set of binary code , The basic format of the instruction is as follows: ： Opcode fields + Address code field , The operation code indicates the operation nature and function of the instruction , The address code gives the operand or the address of the operand .

1. General format of instructions

The basic format of the instruction is as follows: ： Opcode fields + Address code field , The operation code indicates the operation nature and function of the instruction , The address code gives the operand or the address of the operand .

Opcode fields	Address code field
Indicates the operation nature and function of the instruction	Gives the operand or the address of the operand

（1） opcode .

• The operation code is used to reflect the operation to be completed by the instruction .
• The number of operation code digits reflects the basic allowable operation quantity of a certain type of machine , That is, the number of instructions .
• The length of the opcode can be fixed or variable .

Extended opcode technology

• Extended opcode is an instruction optimization technique , That is, let the length of the opcode decrease with the number of addresses Less but more （ That is to expand ）. According to different address instruction formats , Such as three addresses 、 2. Address 、 Single address instruction, etc , The number of digits of the opcode can be selected in different ways , So that on the premise of meeting the needs , The instruction length is effectively shortened .
• Code making 4 Bits can indicate 16 Orders , Because only 15 strip , So there is still a state left 1111, You can mark two address instructions .
• Pay attention to the three addresses here 、 2. Address 、 The single address is different from the address code division below .

（2） Address code

Address code , Used to indicate the source operand address of the instruction 、 Result address 、 The address of the next instruction （ If we agree on the rules, we don't need to record ）. The address here can be the address of main memory 、 Register address 、IO Device address .

• The length of an address code field , It reflects the address range it can cover . So we have to find a way to lengthen the length of the address code field .
• If the opcode length is fixed , Fixed command word length , Then if the number of separate address fields is more , Then the shorter the length of a field . We need to reduce the number of fields .
• there N The address instruction is related to the previous , But it's different .

Instruction type	Instruction content	performance	How to optimize
Four address instructions	There are four address fields , They are No 1、2 Operands , Result storage location , Next command position	Access the memory four times , If we have a program counter PC Words , The next command position is not permanent	Cancel the field of the next position
Three address instructions	Basically the same as the four fields	Or access the memory four times , But by PC Control the next command	The result of each operation does not have to be placed in main memory , It can also be placed in the position of the previous operand , perhaps CPU In the register
2. Address instruction	Deleted the result storage location	If it exists in the position of the previous operand ,4 Time . Put the register on 3 Time	Suppose that the result of our previous operation has been placed in the register , We can discard the first operand
Single address instruction	We added registers ACC, Last calculation result , Put it in , For the first operation , Is to take data and put it
Zero address instruction	There is no opcode in the instruction word	Used for some instructions without operands or implied operands

2. Instruction word length

• Machine word length ： The number of bits of binary data that a computer can process in an integer operation .
• Instruction word length ： The total number of bits of binary code in machine instructions , The instruction word length depends on the length of the slave opcode 、 The length of the operand address and the number of operand addresses . The word length of different instructions is different .
• Storage word length ： A storage unit stores a string of binary code , The number of bits in this string of binary code is called the storage word length , The storage word length can be 8 position 、16 position 、32 I'm waiting for you .

Two 、 Operand type and operation type

1. Operand type

Address 、 Numbers 、 character 、 Logical data, etc .

2. How data is stored

（1） Boundary alignment

If we have a data , When it is stored, it crosses two machine words , Then you need to access and store data many times , The allocation of addresses is also unclear .

Data aligned storage actually means that the data address is stored according to an integer multiple of the size of its data type , The purpose is to make the data read continuously with the least number of times .

If 32 When the bit system reads data, one word is 32 Bits or 4 Bytes , When data is stored , Try to allocate addresses according to an integer multiple of the size of the data type .

• -  For example, in the data ：
  - - int by 4 Bytes ;
    - long long yes 8 Bytes ;
    - char by 1 Bytes .
  -  Then in memory address allocation ：
  - - int It can only be placed at the first address 0,4,8 Location ;
    - long long The first address is 0,8,16 Location ;
    - char It can be placed anywhere .
  

When the computer is reading , about int Type data only needs one cycle ,long long Two cycles . But if stored in natural order ,int Type may require two cycles , At this time int Type data spans the storage space of two words , Equivalent to a warehouse with several rooms （ word ）, Every room has 4 Numbered sequentially （ Address ） Box of （ byte ）, Every time you enter a room is equivalent to a cycle .

When the stored data is less than one word long , Blank bytes can be added to fill .

（2） Byte order

Byte order , As the name implies, the order of bytes , Two more sentences are the storage order of data larger than one byte type in memory ( A byte of data, of course, does not need to talk about the order of the problem ).

In fact, most people seldom deal directly with byte order in actual development . Only in cross platform and network programs, byte order is a problem that should be considered . In all articles introducing byte order, it will be mentioned that byte order is divided into two categories ：Big-Endian and Little-Endian. Quoting standard Big-Endian and Little-Endian Is defined as follows ：

• a) Little-Endian That is, the low byte is discharged at the low address end of the memory , The high byte is placed at the high address end of the memory .
• b) Big-Endian That is, the high byte is discharged at the low address side of the memory , The low byte is placed at the high address of the memory .
• c) Network byte order

For low byte low address and high byte low address , In the vernacular, it is , We have this number from left to right , Or read from right to left .

num_real = 123456
Little-Endian = 1 2 3 4 5 6
Big-Endian = 6 5 4 3 2 1  Instead, the high-order numbers are the low-order numbers in our real data 

3. Operation type

Operation type	eg
Data transfer	MOV, LOAD, STORE
Arithmetic Logic	Add, subtract, multiply and divide
displacement	Arithmetic shift , Logical shift , Cyclic shift
IO

Transfer

type
To transfer unconditionally
Conditional transfer	SKP, BRO
Call and return	CALL, RETURN
Traps and trap instructions	Traps are unexpected interruptions , Trap instructions are generally implicit instructions ,CPU Automatic execution

3、 ... and 、 Addressing mode

The purpose of addressing , One is to find the data you need now , One is to find the location of the next instruction .

1. Command addressing

Basically sequential addressing and skip addressing . Sequential addressing depends on PC Self increasing , Jump addressing is the basic jump

2. Data addressing

There are many ways to address data , You need to set a field in the instruction word to specify the type . And the address code field is not necessarily the real address , It is generally called formal address , Remember to do A.

For real addresses , Write it down as EA.

（1） Address immediately

The direct instruction itself carries data , in other words , Formal address A The content in is no longer an address , Instead, it is directly the operand .

The operand is in the instruction , Immediately following the opcode , Stored as part of the instructions in the code segment of memory , The operand is immediate , This addressing method is called immediate addressing . Data is usually stored in the form of complement . It is often used to assign initial values to registers （ effect ）

• ① The immediate number can be sent to the register 、 One storage unit (8 position )、 Two consecutive storage units (16 position ) In the middle ;

• ② An immediate can only be used as a source operand , Cannot be used as destination operand ;

• ③ With A～F The leading numbers , Numbers must be added in front 0.

advantage ： The instruction has provided operands , No need to access the memory again . Provide operands The fastest .
shortcoming
① The operand is part of the instruction , Do not modify , It is applicable to operations such as assigning initial values to a register or storage unit .
② In the instruction A Limit of digits Of the immediate number expressed by such instructions Range .

（2） Direct addressing

A Address for storing data directly in .

（3） Implicit addressing

Do not explicitly give the operand address , Instead, it indicates in a register .

（4） Indirect addressing

There will be a transition in the middle ,A The address is stored in , This address points to the address of the target in the field memory .

（5） Register addressing （ indirect ）

The number of the register is given , Not the address of the address unit , Registers can store data , You can also save the address （ indirect ）.

（6） Base and indexed addressing

From the definition of both , It doesn't make much difference . It's just Is the reference address or offset address stored in the register . Why not regard it as one ？ Let's explain in detail .

Base addressing

• Definition ： The instruction gives a register number and a formal address , The contents of the register are the reference address , The formal address is used as an offset .
  The reference address plus the offset is the valid address of the operand .
• The contents of the base register in base addressing are usually determined by the operating system or the hypervisor , The median value during program execution is immutable . Its offset is variable . Mainly system oriented .
  A typical application of base addressing is program relocation . The target program is called into memory by the operating system , The user doesn't know where the memory is . The address used in user programming is actually a logical address , It will be converted to the actual physical address at run time in the future . In the base address addressing mode , When the program relocates, the operating system assigns a reference address to the user （ In the reference register ）, When the program is executed, it can be mapped to a physical address .
• And reference addressing can expand the addressing range （ The number of bits of the base address register is greater than the number of bits of the formal address ）. for instance ： Main memory 16M, Base register 24 position （ Not a register address 24 position ！）. The address segment in the instruction uses 16 position ,2 Bits are used to address registers （ Suppose by 4 A register , You need two addresses ）, The remaining 14 Bit gives displacement , Accessible 16KB Continuous storage space . Every time you modify the value of the base register , Then base address plus index can find continuity every time 16KB Space . Significantly improve addressing performance

Addressing

1. Definition ： Instruction gives a register number and formal address , The contents of the register are used as offsets , The formal address is used as the reference address . Base address plus offset to get effective address .
2. Indexed addressing is user oriented , The contents of the index register can be changed by the user , Form address unchanged （ Written directly in the instruction ）. Commonly used in arrays . You can set the first address of the format address bit group , Each time by changing the value of the index register to achieve the operation of the array .
3. That is, it can be modified by indexing , Make the same piece of code can access different content

（7） Relative addressing

PC+A

（8） Stack addressing

Stack addressing : Operands are stored on the stack , Implicitly use stack pointers (SP) As an operand address .
The stack is a register ( Or special register group ） Press... For a specific piece of " Last in, first out (LIFO)"
Principle managed storage , Read in this store / The address of the write unit is given in a specific register , This register is called a stack pointer (SP)