Download Compiler Optimizations: Techniques and Techniques for Improving Code Performance and more Assignments Computer Science in PDF only on Docsity!
Concepts Introduced in Chapter 9 introduction to compiler optimizations basic blocks and control flow graphs local optimizations global optimizations
Compiler Optimizations
^ Compiler optimization is a misnomer. ^ A code-improving transformation consists of asequence of changes that preserves the semanticbehavior (i.e. are safe). ^ A code-improving transformation attempts to makethe program^ ñ^ run faster^ ñ^ take up less space^ ñ^ consume less power ^ An optimization phase consists of a sequence ofcode-improving transformations of the same type.
Types of Compiler Optimizations Function call Loop Memory access Control flow Data flow Machine specific
Function Call Optimizations Procedure integration or inlining Procedure specialization or cloning Tail recursion elimination Function memoization
Loop Optimizations
^ Invariant code motion ^ Strength reduction ^ Induction variable elimination ^ Unrolling ^ Collapsing ^ Fusion ^ Software pipelining
Memory Access Optimizations Register allocation Memory hierarchy improvement ñ Array padding ñ Scalar replacement ñ Loop interchange ñ Prefetching
Control Flow Optimizations Jump elimination ñ Branch chaining ñ Reversing branches ñ Code positioning ñ Loop inversion ñ Useless jump elimination Unreachable code elimination
Data Flow Optimizations
^ Common subexpression elimination ^ Partial redundancy elimination ^ Dead assignment elimination ^ Evaluation order determination ^ Recurrence elimination
Optimizations after Code Generation
Instruction Selection
^ Accomplished by combining RTLs. ^ Data dependences (links) are detected betweenRTLs. ^ Pairs or triples of RTLs are symbolically merged. ^ Legality is checked via a machine description.
Combining a Pair of RTLs
r[1] = r[30]+i; 27 {26}
r[2] = M[r[1]]; r[1]: ^ r[2] = M[r[30]+i]; r[1] = r[30]+i; r[1]:orr[2] = M[r[30]+i];
r[1]:
Combining Three RTLs
r[2] = M[r[3]]; 32 {31}
r[2] = r[2]+1; 33 {32}
M[r[3]] = r[2]; r[2]: ^ M[r[3]] = M[r[3]]+1; r[2] = M[r[3]]+1; r[2]:orM[r[3]] = M[r[3]]+1; r[2]:
Cascading Instruction Selection Actual example on PDP-11 (2 address machine)
r[36]=r[5]; 39 {38}
r[36]=r[36]+i; 40
r[37]=r[5]; 41 {40}
r[37]=r[37]+i; 42 {41}
r[40]=M[r[37]];
r[37]:
r[41]=1; 44 {42}
r[42]=r[40];
r[40]:
r[42]=r[42]+r[41];
r[41]:
M[r[36]]=r[42];
r[42]:r[36]:
Cascading Instruction Selection
(cont.)
r[36]=r[5]; 39 {38}
r[36]=r[36]+i; 40
r[37]=r[5]; 42 {40}
r[40]=M[r[37]+i]];
r[37]:
r[41]=1; 44 {42}
r[42]=r[40];
r[40]:
r[42]=r[42]+r[41];
r[41]:
M[r[36]]=r[42];
r[42]:r[36]:
Cascading Instruction Selection
(cont.)
r[36]=r[5]; 39 {38}
r[36]=r[36]+i; 42
r[40]=M[r[5]+i]]; 43
r[41]=1; 44 {42}
r[42]=r[40];
r[40]:
r[42]=r[42]+r[41];
r[41]:
M[r[36]]=r[42];
r[42]:r[36]:
Cascading Instruction Selection
(cont.)
r[36]=r[5]; 39 {38}
r[36]=r[36]+i; 43
r[41]=1; 44
r[42]=M[r[5]+i]]; 45 {43,44}
r[42]=r[42]+r[41];
r[41]:
M[r[36]]=r[42];
r[42]:r[36]:
Example Sequence of Optimizations
for (sum=0, j = 0; j < n; j++)sum = sum + a[j]; ^ after instruction selectionM[r[13] + sum] = 0;M[r[13] + j] = 0;PC = L18;L19r[0] = M[r[13] + j] <<2;M[r[13] + sum] = M[r[13] + sum] + M[r[0] + _a];M[r[13] + j] = M[r[13] + j] + 1;L18IC = M[r[13] + j]? M[_n];PC = IC < 0
→L19;
Example Sequence of Optimizations
(cont.)
^ after register allocation^ r[2]^ = 0;^ r[1]^ = 0;PC = L18;L19r[0] =
r[1]^ << 2; r[2] =^ r[2]^ + M[r[0] + _a]; r[1] =^ r[1]^ + 1; L18IC =^
r[1]^? M[_n];PC = IC < 0^ →
L19;
Example Sequence of Optimizations
(cont.)
^ after code motionr[2] = 0;r[1] = 0;^ r[4] = M[_n];^ PC = L18L19r[0] = r[1] << 2;r[2] = r[2] + M[r[0] + _a];r[1] = r[1] + 1;L18IC = r[1]?
r[4] ; PC = IC < 0
→^ L19;
Example Sequence of Optimizations (cont.)^
after loop strength reductionr[2] = 0;r[1] = 0;r[4] = M[_n];^ r[3] = _a;^ PC = L18L19r[2] = r[2] + M[
r[3] ]; r[3] = r[3] + 4; r[1] = r[1] + 1;L18IC = r[1]? r[4];PC = IC < 0
→^ L19;
Example Sequence of Optimizations
(cont.)
^ after induction variable eliminationr[2] = 0;r[4] = M[_n] << 2;r[3] = _a;^ r[4] = r[4] + r[3];^ PC = L18;L19r[2] = r[2] + M[r[3]];r[3] = r[3] + 4;L18IC =^
r[3]^?^ r[4]
; PC = IC < 0
→^ L19;
Example of Common Subexpression
Elimination
r[1] = M[r[13] + i] << 2;r[1] = M[r[1] + _b];r[2] = M[r[13] + i] << 2;r[2] = M[r[2] + _b]; r[1] = M[r[13] + i] << 2;r[1] = M[r[1] + _b];r[2] = r[1];
Example of Unreachable Code
Elimination
PC = L12;r[1] = M[r[13] + i];r[1] = r[5] + r[1];M[r[13] + j] = r[1];L13 PC = L12;L13
Example of Branch Chaining
IF a THEN b
While c DO d
compiles into
compiles into
a^
L2:^ c
PC = IC != 0
→L1;^
PC = IC != 0
→L3;
b^
d
L1:^
PC = L2;L3:
