---
id: wiki-2026-0508-concurrent-programming
title: Concurrent Programming
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Concurrency, Multithreading, Async Programming]
duplicate_of: none
source_trust_level: A
confidence_score: 0.9
verification_status: applied
tags: [programming, concurrency, parallelism, systems]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack:
  language: Go, Rust, Python, JavaScript
  framework: tokio, asyncio, goroutines
---

# Concurrent Programming

## 매 한 줄
> **"매 concurrency 매 task 의 interleaving — parallelism 의 simultaneous 와 별개"**. Hoare CSP, Hewitt Actor 의 lineage. 2026 매 Rust async, Go goroutines, Java virtual threads (Project Loom GA), Python free-threaded mode 매 mainstream.

## 매 핵심

### 매 Concurrency vs Parallelism
- **Concurrency**: 매 dealing-with-many things at once (composition).
- **Parallelism**: 매 doing-many things at once (execution).
- 매 concurrent 코드 매 single-core 매 still concurrent. 매 parallel 매 multi-core required.

### 매 Primitives
- **Threads / processes**: 매 OS-level. Heavyweight.
- **Coroutines / fibers / virtual threads**: 매 user-mode lightweight.
- **Async / await**: 매 cooperative scheduling.
- **Channels**: 매 message passing (Go, Rust mpsc).
- **Mutex / RWLock / Semaphore**: 매 shared-memory sync.
- **Atomic**: 매 lock-free primitive.

### 매 응용
1. Web server: 매 1 connection-per-virtual-thread (Loom) / async (tokio).
2. Pipeline: 매 channel-based fan-out / fan-in.
3. Actor system: 매 isolation per actor + supervision (Erlang/Elixir, Akka).

## 💻 패턴

### Go: goroutine + channel
```go
package main

import "fmt"

func worker(id int, jobs <-chan int, results chan<- int) {
    for j := range jobs {
        results <- j * 2
    }
    fmt.Println("worker", id, "done")
}

func main() {
    jobs := make(chan int, 100)
    results := make(chan int, 100)
    for w := 1; w <= 3; w++ {
        go worker(w, jobs, results)
    }
    for j := 1; j <= 9; j++ { jobs <- j }
    close(jobs)
    for r := 1; r <= 9; r++ { <-results }
}
```

### Rust: tokio + select!
```rust
use tokio::{select, time::{sleep, Duration}};

#[tokio::main]
async fn main() {
    let task1 = async { sleep(Duration::from_millis(100)).await; "fast" };
    let task2 = async { sleep(Duration::from_millis(200)).await; "slow" };
    select! {
        v = task1 => println!("got {v}"),
        v = task2 => println!("got {v}"),
    }
}
```

### Java 21+: virtual threads
```java
try (var executor = Executors.newVirtualThreadPerTaskExecutor()) {
    IntStream.range(0, 10_000).forEach(i ->
        executor.submit(() -> {
            Thread.sleep(Duration.ofSeconds(1));
            return i;
        }));
}  // 10k virtual threads, ~negligible memory
```

### Python: asyncio TaskGroup (3.11+)
```python
import asyncio
import httpx

async def fetch(client, url):
    r = await client.get(url)
    return r.status_code

async def main():
    async with httpx.AsyncClient() as client:
        async with asyncio.TaskGroup() as tg:
            tasks = [tg.create_task(fetch(client, u)) for u in URLS]
        results = [t.result() for t in tasks]
```

### Rust: structured concurrency (tokio JoinSet)
```rust
use tokio::task::JoinSet;
let mut set = JoinSet::new();
for url in urls { set.spawn(fetch(url)); }
while let Some(res) = set.join_next().await {
    println!("{:?}", res?);
}
```

### Lock-free atomic counter
```rust
use std::sync::atomic::{AtomicU64, Ordering};
static COUNTER: AtomicU64 = AtomicU64::new(0);

fn bump() -> u64 {
    COUNTER.fetch_add(1, Ordering::Relaxed)
}
```

### Actor pattern (Elixir)
```elixir
defmodule Counter do
  use GenServer
  def start_link(_), do: GenServer.start_link(__MODULE__, 0, name: __MODULE__)
  def init(s), do: {:ok, s}
  def handle_call(:inc, _, s), do: {:reply, s+1, s+1}
end
```

### Backpressure with bounded channel
```go
sem := make(chan struct{}, 10)  // max 10 concurrent
for _, url := range urls {
    sem <- struct{}{}
    go func(u string) { defer func(){ <-sem }(); fetch(u) }(url)
}
```

## 매 결정 기준
| 상황 | Approach |
|---|---|
| I/O-bound, 10k+ connections | Async (tokio, asyncio, Loom) |
| CPU-bound, multi-core | Threads / rayon (Rust) |
| Pipeline, multi-stage | Channels (Go, Rust mpsc) |
| Isolation + fault tolerance | Actors (Elixir, Akka) |
| Shared mutable state | Mutex + minimize critical section |
| Hot counter | Atomic, no lock |

**기본값**: structured concurrency + bounded channel + cancellation propagation.

## 🔗 Graph
- 변형: [[Async Programming]]
- 응용: [[Distributed Systems]]

## 🤖 LLM 활용
**언제**: 매 concurrent code review 의 race-condition spotting, 매 deadlock-pattern detection, 매 channel-pattern translation.
**언제 X**: 매 subtle memory-ordering bug — 매 formal verification (TLA+, loom) 의 사용.

## ❌ 안티패턴
- **Shared mutable state without sync**: 매 race condition. UB in C++/Rust.
- **Mutex in hot path**: 매 contention 의 serialization. 매 atomic / per-thread state.
- **Unbounded goroutine spawn**: 매 OOM. 매 bounded pool / semaphore.
- **Sync I/O in async function**: 매 single blocked task → 매 entire executor stall.
- **Lock ordering inconsistent**: 매 deadlock. 매 always same order.
- **Cancellation 의 ignore**: 매 dangling task / resource leak.

## 🧪 검증 / 중복
- Verified (Hoare _Communicating Sequential Processes_, Go Memory Model 2022, Rust Async Book, JEP 444 (Java Virtual Threads), Python PEP 703).
- 신뢰도 A.

## 🕓 Changelog
| 날짜 | 변경 |
|---|---|
| 2026-05-08 | Phase 1 |
| 2026-05-10 | Manual cleanup — primitives, Go/Rust/Java/Python pattern, anti-patterns |