CPU scaling benchmark
workers
2 +1 main
iters total
500M
166666666/stream
elapsed
1470.1 ms
total CPU used
4340.57 ms
speedup
2.95×
vs serial
efficiency
98.3%
of 3× ideal
| stream | spawn ms | spawned@ | work start@ | work end@ | work ms | reap wait ms |
|---|---|---|---|---|---|---|
| 0 (main) | 0 | 3.67 | 3.68 | 1449.84 | 1446.16 | 0 |
| 1 | 1.986 | 2 | 19.4 | 1467.38 | 1447.98 | 18.99 |
| 2 | 1.602 | 3.64 | 19.4 | 1465.83 | 1446.43 | 16.13 |
main
w1
w2
fork+handshake
CPU work
parent reap wait
what this measures
Each stream runs a tight integer LCG loop — working set is one CPU register, no memory access,
no shared data. Speedup = sum(stream CPU time) / wall-clock elapsed. Efficiency = speedup / (workers+1).
100% efficiency means perfect linear scaling; less than 100% is the cost of serial fork setup,
reap tail, SMT/core contention.