Go manages memory for you. This topic explains exactly how, because the difference between a service that spends 2% of its CPU on memory management and one that spends 30% is almost always a handful of decisions you can see once you know where to look.
| # | Lesson | The question it answers |
|---|---|---|
| 01 | Stack and Heap | Where do values live, and what does each place cost? |
| 02 | Escape Analysis | How does the compiler decide, and how do I read its decision? |
| 03 | The Allocator | What happens, step by step, when the heap is asked for memory? |
| 04 | The Garbage Collector | How is unused memory found and freed while the program keeps running? |
| 05 | Writing Allocation-Aware Code | How do I cut allocations where it matters? |
When you finish you can predict whether a line of code allocates, explain what GOGC and
GOMEMLIMIT do, read a GC trace, and remove the allocations from a hot path.
flowchart LR
V["A value is created"] --> EA{"Escape analysis:<br/>can it outlive the function?"}
EA -->|"no"| ST["Goroutine stack<br/>free: just move a pointer"]
EA -->|"yes"| AL["Allocator<br/>size class, span, per-thread cache"]
AL --> HP[("Heap")]
HP --> GC["Garbage collector<br/>mark what is reachable, sweep the rest"]
GC -->|"freed spans"| AL
class V neutral
class EA queue
class ST compute
class AL,HP,GC memory