Let’s discuss control flow. We’ll be covering three powerful techniques for faster, cleaner, and more maintainable code:
- Push
if
statements up to centralize logic - Move
for
loops to the bottom to support batching and optimization - Move
switch
logic outside of hot loops.
We’ll also examine the Big O consequences and how compilers benefit from clean structure.
Move if
Statements to the Top
If you’re verifying nulls or guards within a function, ask yourself: should this logic be at the calling site instead?
Bad:
interface User {
email: string;
}
function processUser(user: User | null): void {
if (!user) return;
sendEmail(user.email);
}
Good:
function processUser(user: User): void {
sendEmail(user.email);
}
const maybeUser = getUser();
if (maybeUser !== null) {
processUser(maybeUser);
}
Why?
- Fewer branches within reusable logic
- Call site has context to inform behavior
- Enhances type safety and compiler inference
Big O:
Eliminating inner function branching decreases decisions at run-time from O(n) to O(1) per invocation in performance-critical code paths
Pull Down for
Loops
Rather than looping externally, bring iteration into the function to take advantage of batching.
Bad:
for (let task of taskList) {
runTask(task);
}
Good:
function runTasks(tasks: Task[]): void {
for (let task of tasks) {
runTask(task);
}
}
runTasks(taskList);
Better:
function runTasksOptimized(tasks: Task[]): void {
// Stage 1: Validate
for (let task of tasks) validate(task);
// Stage 2: Execute
for (let task of tasks) execute(task);
// Stage 3: Cleanup
for (let task of tasks) console.log(task);
}
Why?
- Batch-aware logic minimizes the overhead of calls
- Facilitates vectorization, caching, and stages of the pipeline
Big O:
O(n) remains O(n) but with less branching = better caching and better CPU efficiency
Combine: if
Outside, for
Inside
Bad:
for (let msg of messages) {
if (urgent) handleUrgent(msg);
else handleNormal(msg);
}
Good:
if (urgent) {
for (let msg of messages) handleUrgent(msg);
} else {
for (let msg of messages) handleNormal(msg);
}
Why?
- Conditional extracted outside the loop
- No repeated branching, assists CPU prediction
Big O:
- Same number of steps overall, but reduced branch misprediction and improved runtime
Loop-Switching: Refactor Your Dispatch
Typical Code:
for (let item of items) {
switch (item.kind) {
case 'text':
renderText(item);
break;
case 'image':
renderImage(item);
break;
}
}
Improved:
const textItems = items.filter(i => i.kind === 'text');
const imageItems = items.filter(i => i.kind === 'image');
for (let t of textItems) renderText(t);
for (let i of imageItems) renderImage(i);
Why?
- Fewer branches in the inner loop
- Supports multi-pass or parallel strategies
Big O:
Minor increment (2×O(n)), but enhanced locality and optimizable dispatching
Compile-Time Wins: What Compilers Understand
- Inner loops without branches allow compilers to employ SIMD and loop unrolling
- Split loops can be parallelized or pipelined
- Centralized conditions result in fewer jumps and improved CPU branch prediction
Final Thoughts
These aren’t tricks. These are coding models for what works with the machine, not against it.
- Organize information more effectively:
- Shun accidental complexity
- Improve performance (particularly in hot paths)
- Make debugging and testing simpler
Push your
if
s to the top, bring yourfor
s to the bottom, shift yourswitch
aside—and code to please the compiler.
TL;DR (Too Long? Don’t Refactor):
- ✅ Move
if
conditions out of inner logic - ✅ Batch your logic within
for
-friendly functions - ✅ Split
switch
into groups prior to looping
Small changes result in big wins. Now go fix some loops!