Concepts/Synchronization/Shared Counter

Shared Counter

Trace how different synchronization approaches affect correctness and latency. Step through the same workload three ways and see where shared state goes wrong.

Non-Atomic Atomic Mutex

Compare ›Deep Dive ›

Steps

1/6

Stateshared + per-thread

T1line 4

T1 localread

local0

shared

shared memory

counter0

counter per stepnow = 0

T2 local

local—

Per increment2.50nsmedian of 10 reps

Correctness✗ 74% lost1,026,290 / 4,000,000

vs Non-Atomicbaseline (but wrong)fast but breaks

Setup4 threadsCPP · 1M iters · Apple M1 Pro

OBSERVED

T1 reads counter = 0 into its local variable.

Non-Atomic

1int counter = 0;
2
3void worker() {
4    int local = counter;  // read
5    local = local + 1;    // compute
6    counter = local;      // write
7}

Historystep 1 / 6

space · ← → · rstep 1/6

Counter ended at 1,026,290 of 4,000,000 expected — 74% of writes vanished. The fix is one of these →

to Atomic to Mutex