CPU scaling benchmark
workers
2 +1 main
iters total
500M
166666666/stream
elapsed
1466 ms
total CPU used
4333.93 ms
speedup
2.96×
vs serial
efficiency
98.7%
of 3× ideal
| stream | spawn ms | spawned@ | work start@ | work end@ | work ms | reap wait ms |
|---|---|---|---|---|---|---|
| 0 (main) | 0 | 3.39 | 3.39 | 1443.48 | 1440.09 | 0 |
| 1 | 1.92 | 1.94 | 14.49 | 1460.35 | 1445.86 | 17 |
| 2 | 1.423 | 3.38 | 15.39 | 1463.37 | 1447.98 | 20 |
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.