Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
1
Programming Languages (CS 550)
Mini Language Compiler
Jeremy R. Johnson
2
Introduction
• Objective: To illustrate how to map Mini Language instructions to RAL instructions. To do this in a systematic way that illustrates how to write a compiler to translate Mini Language programs to RAL programs. Show simple optimizations that can be used to reduce the number of instructions.
• Algorithm– Construct code for expressions, assignments, if, and while.– Concatenate statements in stmt-list– Allocate temporaries as needed– Keep track of variables, constants, and temporaries in Symbol Table– Use symbolic instructions and fill in absolute addresses (linking) when
complete code has been constructed
3
A Random Access Machine
Control Unit
AC
Program
1
2
3
4
5
6...
...
Memory
AC = accumulatorregister
4
Instruction Set
• LDA X; Load the AC with the contents of memory address X• LDI X; Load the AC indirectly with the contents of address X• STA X; Store the contents of the AC at memory address X• STI X; Store the contents of the AC indirectly at address X• ADD X; Add the contents of address X to the contents of the AC• SUB X; Subtract the contents of address X from the AC• MUL X; Multiply the contents of address X to the contents of the AC• JMP X; Jump to the instruction labeled X• JMZ X; Jump to the instruction labeled X if the AC contains 0• JMN X; Jump to the instruction labeled X if the contents of the AC
; is negative• HLT ; Halt execution
5
Memory Organization
Constants
Prog. Variables
Temp. Variables
• Num_Consts• Num_Vars• get_temp()• Num_Temps
6
Symbolic Instructions
• Addresses and labels can be symbolic names• Symbolic names are mapped to actual addresses during linking
• Example:– LD x– ST z– ADD y– JMP L
• Linked code with (x=100, y =110, z = 105, L = 20)– LD 100– ST 105– ADD 110– JMP 20
7
Symbol Table
• Map from identifiers → Symbol table entries
• Symbol table entries contain: address [may be unknown]
• Indicate whether entry is an constant, variable, temporary or label
8
Expressions
expr → expr1 op expr2
Code1 ; result stored in t1
Code2 ; result stored in t2
LD t1 ; load result of exp1
OP t2 ; apply op to result of exp2 and result of exp1
ST t3 ; store result of exp1 op exp2
9
Expressions
expr → NUMBER
; check to see if NUMBER in symbol table,; otherwise add to symbol table
LD NUMBER ; load constant from constant tableST tn ; next available temporary
10
Expressions
expr → IDENT
; check to see if IDENT in symbol table; otherwise add to symbol table
LD IDENT ; load constant from constant tableST tn ; next available temporary
11
Assignment
assign_stmt → IDENT = expr
; check to see if IDENT in symbol table; otherwise add to symbol table
CodeLD tST IDENT
12
Conditional Statements
if_stmt → if expr then S1 else S2 fi
Codee ; result stored in tLD t ;JMN L1Code1
JMP L2L1: Code2L2:
13
While Statements
while_stmt → while expr do S od
L1: Codee ; result stored in tLD t ;JMN L2CodeS
JMP L1L2:
14
Statement List
stmt-list → stmt; stmt-list | stmt
code1
code2
…coden
15
Example
n := 0 - 5; if n then i := n else i := 0 - n fi; fact := 1; while i do fact := fact * i; i := i - 1 od
16
Example
n := 0 - 5;
LD ZEROST T1LD FIVEST T2LD T1SUB T2ST T3LD T3ST n
17
Example
if n then i := n else i := 0 - n fi;
LD nST T4LD T4JMN L1LD nST T5LD T5ST iJMP L2
L1: LD ZEROST T6LD nST T7LD T6SUB T7ST T8LD T8ST i
L2:
18
Example
fact := 1;
LD ONEST T9LD T9ST fact
19
Example
while i do fact := fact * i; i := i - 1
od
L3: LD iST T10JMN L4LD factST T11LD iST T12LD T11MUL T12ST T13
LD T13ST factLD iST T14LD ONEST T15LD T14SUB T15ST T16LD T16ST iJMP L3
L4:
20
Complete Example(concatenate and append HLT)
LD ZEROST T1LD FIVEST T2LD T1SUB T2ST T3LD T3ST nLD nST T4LD T4JMN L1LD nST T5LD T5ST iJMP L2
L3: LD iST T10JMN L4LD factST T11LD iST T12LD T11MUL T12ST T13
LD T13ST fact
LD iST T14LD ONEST T15LD T14SUB T15ST T16LD T16ST iJMP L3
L4: HLT
L1: LD ZEROST T6LD nST T7LD T6SUB T7ST T8LD T8ST i
L2: LD ONEST T9LD T9ST fact
21
Symbol TableName Value Type addrZERO 0 const ?FIVE 5 const ?n u var ?T1 u temp ?T2 u temp ?T3 u temp ?T4 u temp ?T5 u temp ?i u var ?T6 u temp ?T7 u temp ?
T8 u temp ?ONE 1 const ?T9 u temp ?fact u var ?T10 u temp ?T11 u temp ?T12 u temp ?T13 u temp ?T14 u temp ?T15 u temp ?T16 u temp ?
22
Symbol Table and Label Summary
Num_Vars = 3Num_Consts = 3Num_Temps = 16
ConstantsZERO -> addr 1FIVE -> addr 2One -> addr 3
Variablesn -> addr 4i -> addr 5fact -> addr 6
TemporariesT1 -> addr 7T2 -> addr 8
…T16 -> addr 22
L1 = 19L2 = 28L3 = 32L4 = 54
23
Linked Example
LD 1ST 7LD 2ST 8LD 7SUB 8ST 9LD 9ST 4LD 4ST 10LD 10JMN 19LD 4ST 11LD 11ST 5JMP 28
L3: LD 5ST 16JMN 53LD 6ST 17LD 5ST 18LD 17MUL 18ST 19
LD 19ST 6
LD 5ST 20LD 3ST 21LD 20SUB 21ST 22LD 22ST 5JMP 32
L4: HLT
L1: LD 1ST 12LD 4ST 13LD 12SUB 13ST 14LD 14ST 5
L2: LD 3ST 15LD 15ST 6
24
Optimizations
• Peephole optimization – Remove ST immediately followed by LD
• Commute (expr1,expr2) in expr → expr1 op expr2 to allow additional peephole optimizations
• Constant folding
• Common subexpression elimination
25
Complete Example(after peephole optimization)
LD ZEROST T1LD FIVEST T2LD T1SUB T2ST T3LD T3ST nLD nST T4LD T4JMN L1LD nST T5LD T5ST iJMP L2
L3: LD iST T10JMN L4LD factST T11LD iST T12LD T11MUL T12ST T13
LD T13ST fact
LD iST T14LD ONEST T15LD T14SUB T15ST T16LD T16ST iJMP L3
L4: HLT
L1: LD ZEROST T6LD nST T7LD T6SUB T7ST T8LD T8ST i
L2: LD ONEST T9LD T9ST fact
26
Complete Example(after peephole optimization)
LD ZEROST T1LD FIVEST T2LD T1SUB T2JMN L1LD nST iJMP L2
L3: LD iST T10JMN L4LD factST T11LD iST T12LD T11MUL T12ST fact
LD iST T14LD ONEST T15LD T14SUB T15ST i
JMP L3L4: HLT
L1: LD ZEROST T6LD nST T7LD T6SUB T7ST i
L2: LD ONEST fact
38 vs. 54 instructions
27
Supporting Procedures
• Fully static environment – No recursion– Activation record
• Parameters• Local variables (keep count)• Return address (indirect jump needed)• Can be statically allocated
• Dynamic environment– Allow recursion– Call stack (dynamic allocation)– Indirect load and store needed
28
Memory Organization
Constants
GlobalProg. Variables
GlobalTemp. Variables
Activation Records
Constants
GlobalProg. Variables
GlobalTemp. Variables
Call Stack
29
Program Memory Organization
Procedures
P1
P2
P3
Main Program
Procedure Entry in Function Table
• Number of parameters• Number of local/temp variables• Starting address• Number of instructions
• Need to know starting address of main program
30
Activation Record
Parameters
Local Variables
Temp. Variables
Return Address
Frame Pointer
Stack Pointer
For call stack
31
Example: fact(n)
defineproc(n)i := n;fact := 1;
while i do fact := fact * i; i := i - 1
od;return := fact
end
32
fact(n)
LD iST T7LD ONEST T8LD T7SUB T8ST T9LD T9ST iJMP L1
L2: LD factST T10LD T10ST return
L1: LD iST T3JMN L2LD factST T4LD iST T5LD T4MUL T5ST T6LD T6ST fact
LD nST T1LD T1ST iLD ONEST T2LD T2ST fact
33
Activation Recordni
factT1T2T3T4T5T6T7T8T9
T10return
ret. addr.
FP
SP
Accessing ARLD n ⇔ LDI FPST n ⇔ STI FP
LD i ⇔ LD FPADD ONEST FPBLDI FPB
⇔ LDO FP[1]ST i ⇔ STO FP[1]LD Tj ⇔LDO FP[j+Num_Param+Num_Vars]
34
Calling SequenceInitiate call1. Create activation record
1. Update FP and SP
2. Store parameters in activation record3. Store return address (RA)4. Jump to starting address of procedure code
1. Introduce call instruction (can place RA relative to SP)
Return from call1. Store return value in activation record (when return is assigned)2. Jump to RA
1. Introduce ret instruction (jmp indirect)
3. Retrieve return value from activation record4. Update FP and SP