70 Views

Download Presentation
##### What Is State?

**An Image/Link below is provided (as is) to download presentation**

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -

**CS1103 電機資訊工程實習**What Is State? Prof. Chung-Ta King Department of Computer Science National Tsing Hua University (Contents from http://ai-depot.com/FiniteStateMachines/FSM-Background.html, www.dcs.shef.ac.uk/~ahw/comp_arch/L3.ppt )**這是什麼爛問題?**• State有不同的類型! • 新、舊 • 運作良好、堪用、待修、報廢 • 停止、行進、牽行 • ...**A Better Question**機車的「運作狀態」有幾種? 熄火靜止、熄火行進、 發動靜止、發動行進**不同「運作狀態」之間如何轉換?**• State transition 熄火 行進 發動 推動 熄火 靜止 發動 行進 熄火 停止 啟動 發動 發動 靜止 煞車停止 熄火 Finite state machine**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**Finite State Machines (FSM)**• or Finite State Automation (FSA) • Models of behaviors of a system or object, with a limited number of defined conditions or modes, where modes change with circumstance • 4 main elements: • States: define behavior and may produce actions • State transitions: move from one state to another • Rules (or conditions): must be met to allow a state transition • Input events: externally or internally generated; may trigger rules and lead to state transitions • An initial state and a current state Good for representation and modeling sequences of behaviors**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**Consider the Program**main() { i=5; j=6; k=0; k=i+j; if (k>0) i=0; else j=0; } • Can this program be modeled with a FSM? What is its “state”?**Initial State**main() { i=5; j=6; k=0; k=i+j; if (k>0) i=0; else j=0; } Memory j 6 CPU i 5 k 0**State 1**main() { i=5; j=6; k=0; k=i+j; if (k>0) i=0; else j=0; } Memory j 6 CPU i 5 k 11**State 2**main() { i=5; j=6; k=0; k=i+j; if (k>0) i=0; else j=0; } Memory j 6 CPU i 0 k 11**State 3**main() { i=5; j=6; k=0; k=i+j; if (k>0) i=0; else j=0; } Memory j 0 CPU i 5 k 11**FSM Representing the Program**2 k=i+j k<=0 1 4 V k>0 3**Compare with Flow Chart**k=i+j yes no k>0 j=0 i=0**Some Questions**• Can FSM be used to represent any program? Any algorithm? • What is the “state” of a program/process? Or a computer? • Why is this important? • 系統更新的回復點 • 電腦休眠之後回復原來畫面 • Context switch • Interrupt and resume • Is there a machine that can model any computation? Turing machine**State of a Procedure?**• Fibonacci number (iterative version) unsigned int fib(unsigned int n) { unsigned int i=1, j=0, k, t; for (k=1; k<=n; k++) { t=i+j; i=j; j=t; } return j; } What need to be stored when the procedure is suspended and later resumed (as if nothinghad happened)?**Why Is It Important?**• Fibonacci number (recursive version) unsigned int fib(unsigned int n) { unsigned int i=n-1, j=n-2; if (n<2) return n; else return fib(i) + fib(j); } What happens when n, i, and j eachrefer to the same memory location across procedure calls?**How to Solve for Recursion?**• If we could save the “state” of the calling procedure in a safe/separate place that the called procedure will not overwrite, then the recursion can be executed correctly • Big idea: “procedure state” in run-time stack • Allocate the “state” of each called procedure in the run-time stack procedure frame • All references to the “state” go to the stack • Each invocation will push “state” down the stack • How to write assembly code to do that? Memory reference relative to stack pointer Chapter 8 of CS2422**Summary: A View of a “State”**• Storage contents of the code/system that must be saved away when the code/system is suspended and resumed later, as if nothing had happened • State of a procedure on procedure calls • State of a thread/process on context switch make time-sharing possible • State of a process on migration to another PC • State of a computer on error recovery • State of a computer on hibernation • ...**Implication of the View**• We can do anything as long as we do not disturb the “state” z = x * 3.1412; if (z > 10) j = j + k; • Since FP multiplication takes a long time, can we execute the addition at the same time? out-of-order, speculative**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**FSM Also Good for Hardware**Processor Register Memory AX … BX x a data Controller b 01 0000001 1000010 MOV AX, a 11 0000001 1000110 ADD AX, b 01 1000011 0000001 MOV x, AX ALU … clock IR PC PC: program counter IR: instruction register address**Step 1: Fetch (MOV AX, a)**Register Memory AX … BX x a data Controller b 01 0000001 1000010 MOV AX, a 11 0000001 1000110 ADD AX, b 01 1000011 0000001 MOV x, AX ALU … clock IR PC 01 0000001 1000010 0000111 address**Step 2: Decode (MOV AX,a)**Register Memory AX … BX x a data Controller b 01 0000001 1000010 MOV AX, a 11 0000001 1000110 ADD AX, b 01 1000011 0000001 MOV x, AX ALU … clock IR PC 01 0000001 1000010 0000111 address**Step 3: Execute (MOV AX,a)**Register Memory AX … 00000000 00000001 BX x 00000000 000000001 a data Controller b 01 0000001 1000010 MOV AX, a 11 0000001 1000110 ADD AX, b 01 1000011 0000001 MOV x, AX ALU … clock IR PC 01 0000001 1000010 0000111 address**Concept of the State Machine**Computer Hardware = Datapath + Control Qualifiers Registers Combinational Functional Units (e.g., ALU) Busses FSM generating sequences of control signals Instructs datapath what to do next Control Control State Control Signal Outputs Qualifiers and Inputs Datapath**FSM of the Computer**• For this highly simplified computer, the controller can be described by a FSM Decode Fetch if (ADD) if (MOV) if (JNG) ADD MOV JNG ... Each state will generate certain control signalsto control the datapath**Internal Structure of Controller**Next state (state transition) State Register Combinationalcircuit Output To datapath Clk Controller EFLAGS Instruction Reg ALU zero,carry,overflow,...**Combinational & Sequential Logic**• Combinational logic does not have memory • It generates output solely according to the input and does not care about history • Often, we need a different reaction on the same input depending on the current state • E.g. • Current state is 7 and input is 1, the new state and output are 8 • Current state is 15 and input is 1, the new state and output are • To make the new state depends on the previous state, we need memory**Sequential Logic and FSM**• Computers are made of sequential logic blocks • Truth tables are used to design combinational logic, but can’t be used for sequential logic • Finite state machines (FSM) are used instead • FSM describes a sequential logic block in terms of: • Set of states, • State transition function (defined on current state and input), and • Output function (defined on current state and input, or on current state only)**Two Types of FSMs**Differ in how outputs are produced • Moore Machine: • Outputs are independent of the inputs, i.e. outputs are produced from within the state of the state machine • Mealy Machine: • Outputs can be determined by the present state alone, or by the present state and the present inputs, i.e. outputs are produced as the machine makes a transition from one state to another • Any Moore machine can be turned into a Mealy machine (and vice versa)**State 1**q,r State 2 x,y Moore Machine The Moore State Machine output remains the same as long as the state machine remains in that state. The output can be arbitrarily complex but must be the same every time the machine enters that state. a,b Input condition that must exist in order to execute these transitions from State 1 i,j Output condition that results from being in a particular present state**State 1**State 2 Mealy Machine • The Mealy State Machine generates outputs based on: • The Present State, and • The Inputs to the M/c. • So, same state can generate many different patterns of output signals, depending on the inputs. • Outputs are shown on transitions since they are determined in the same way as is the next state. a,b q,r Input condition that must exist in order to execute these transitions from State 1 i,j x,y Output condition that results from being in a particular present state**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**Safety Belt Control**• We want to design a controller for safety belt • If the seat is seated and the belt is not buckled within a set time, a buzzer will sound until the belt is buckled event driven • Inputs: seat sensor, timer, belt sensor • Output: buzzer, timer • System: specialized computer for reacting according to eventssensed by the sensors**FSM for Event-driven Systems**no seat/- no seat/ buzzer off idle seat/timer on no belt and no timer/- no seat/- buzzer seated Belt/buzzer on belt/- belt/ buzzer off belted no belt/timer on**C Code Structure**• Current state is kept in a variable • State table is implemented as a switch • Cases define states • States can test inputs and produce outputs while (TRUE) { switch (state) { case state1: … } } • Switch is repeatedly evaluated by while-loop**C Implementation**#define IDLE 0 #define SEATED 1 #define BELTED 2 #define BUZZER 3 switch (state) { case IDLE: if (seat) { state = SEATED; timer_on = TRUE; } break; case SEATED: if (belt) state = BELTED; else if (timer) state = BUZZER; break; … }**Another Example**in1=1/x=a A B r=0/out2=1 r=1/out1=0 in1=0/x=b C D s=0/out1=0 s=1/out1=1**C State Table**switch (state) { case A: if (in1==1) { x = a; state = B; } else { x = b; state = D; } break; case B: if (r==0) { out2 = 1; state = B; } else { out1 = 0; state = C; } break; case C: if (s==0) { out1 = 0; state = C; } else { out1 = 1; state = D; } break;**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**Moore Machine**Mealy Machine 0/0 0 Even Even Input Output [0] 1/1 1/0 1 1 Input Output Odd Odd [1] 0 0/1 FSM as Recognizer/Translator • Outputs a ‘0’ if an even # of 1’s is received and outputs a ‘1’ otherwise • What is state? • Two states: whether an even # of 1s have been received, or an odd # of 1s have been received**Finite State Machines**• Language recognizer (acceptor) • Problem solver recognizer of L yes, y in L y no problems strings (languages) answers machines**others/[word]**State 2 State 1 [a-z, A-Z] others [a-z, A-Z] FSM as Recognizer/Translator • How to recognize words in a document? • Assume characters in the document are input sequentially • What is state?**Example**• A FSM that outputs a ‘1’ whenever it receives a multiple of 3 # of 1’s on a serial input • Relevant information to solve the problem: (A) A multiple of 3 # is received (B) A non-multiple of 3 # is received • Questions to consider: (1) How do we go from (A)(B)Ans.: If a ‘0’ is received (2) How do we go from (B)(A)Ans.: Not clear; need to consider(B1): 3y+1 # of 1’s received.(B2): 3y+2 # of 1’s received. y is an integer 0**Example**• Transitions between 3 classes of information: (A) (B1) (B2) (A) • They can be considered states of the FSM: 1 received 1 received 1 received 0 Output 00 0/1 Reset Reset i=0 i=0 [1] Input 0/0 1 0 1/0 i=1 1/1 01 1 i=1 [0] i=2 1/0 i=2 [0] 10 1 0/0 0 Mealy Machine Moore Machine**Outline**• What is finite state machine? • Finite state machine for program modeling • The important concept of “state” • Finite state machine for circuit control • Finite state machine for embedded control • Finite state machine for recognition • Modeling with finite state machine**Scenario 1**入侵偵測系統： • 某一房間的麥克風偵測到有不正常的聲音，即提高偵測頻率，並通知房間內的攝影機追蹤聲音的來源，且將影像傳到管理員手機。 • 同時，通知走廊及其他房間內的麥克風提高偵測頻率，攝影機追蹤移動物體。 What are the states? What is the FSM?**Scenario 2**大富翁： • What are the states? • How states transit? Are the transitions deterministic? • What are the expected outcomes? Markov chain, stochastic process, ...