Code Optimization in Compiler Design

Code optimization is the process of improving a program to make it more efficient in terms of speed, memory, and resource usage, without changing its functionality. The key aspects of code optimization include:

• Improved performance: Optimized code executes faster and uses fewer resources.
• Reduced code size: Smaller code is easier to distribute and deploy.
• Increased portability: Optimized code can run efficiently on multiple platforms.
• Lower power consumption: Efficient code uses less energy, making it more energy-friendly.
• Better maintainability: Well-optimized code can be easier to read and maintain.
• Reasonable compilation time: Optimization should not make compilation excessively slow.
• Correctness assurance: Optimization must preserve the program’s original behavior and logic.
• Manageable complexity: Code optimization may increase complexity, but the program should remain easy to understand, maintain, and debug.

When to Optimize?

Optimization of the code is often performed at the end of the development stage since it reduces readability and adds code that is used to increase performance. 

Why Optimize?

Algorithm optimization is beyond the scope of the code optimization phase. Instead, the compiler optimizes the intermediate program code, which may also reduce the size of the generated code. Therefore, code optimization helps to:

• Reduce memory usage and code size.
• Improve program execution speed.
• Automatically optimize code using a compiler optimizer, since manual optimization is difficult and time-consuming.

Types of Code Optimization

• Machine Independent Optimization: Improves the intermediate code to produce more efficient target code. It is performed without considering the target machine architecture. The transformations do not involve CPU registers or absolute memory addresses.

• Machine Dependent Optimization: Performed after the target code is generated. It optimizes the code according to the specific architecture of the target machine. This type of optimization uses CPU registers and may involve absolute memory references to take better advantage of the memory hierarchy and hardware features.

Ways to Optimize Code

There are several ways to optimize code. Some of them are mentioned below.

1. Compile Time Evaluation:

Expressions evaluated during compilation.

2. Variable Propagation:

Replaces a variable with another variable having the same value.

3. Constant Propagation:  

Replaces variables with their constant values.

Example: 

4. Constant Folding: 

Computes constant expressions at compile time.

Example: 

5. Copy Propagation: 

• It is extension of constant propagation.
• Replaces copied variables with the original variable.

Example : 

6. Common Sub Expression Elimination: 

Avoids repeated calculation of the same expression.

7. Dead Code Elimination: 

Removes code that is never used.

Example: 

8. Unreachable Code Elimination:

Removes code that will never execute. 

9. Function Inlining: 

Replaces a function call with its body.

10. Function Cloning: 

• Here, specialized codes for a function are created for different calling parameters.
• Function Cloning is a compiler optimization technique where multiple specialized versions of a function are created based on different usage contexts. It is different from function overloading, which is a programming language feature.

11. Induction Variable :

Uses variables that change regularly in loops.

Examples: 

12. Strength Reduction:

Replaces costly operations with cheaper ones.

Loop Optimization Techniques

1. Code Motion (Frequency Reduction)

This technique reduces the number of times an expression is evaluated by moving loop-invariant statements outside the loop.

Example: 

2. Loop Jamming

Loop Jamming combines two or more loops into a single loop to reduce loop overhead and improve runtime performance.

Example: 

3. Loop Unrolling

This technique reduces the number of loop iterations by performing multiple operations in a single iteration, which decreases loop control overhead.

Example: 

Where to Apply Optimization?

Code optimization can be applied at different stages of program execution and compilation to improve performance and efficiency.

1. Source Program Level
Optimization at the source program level involves modifying the algorithm or program structure, such as improving loop structures or reducing unnecessary computations. Here, the programmer (user) performs the optimization.

2. Intermediate Code Level
At this stage, the compiler optimizes the intermediate representation of the program. This may include improving address calculations and optimizing procedure calls.

3. Target Code Level
Optimization is performed on the generated machine code. The compiler improves the use of CPU registers, instruction selection, and instruction movement to produce efficient target code.

Optimization Based on Program Scope