CPU scaling benchmark
workers
2 +1 main
iters total
500M
166666666/stream
elapsed
1467.19 ms
total CPU used
4338.31 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 | 3.42 | 3.43 | 1445.04 | 1441.61 | 0 |
| 1 | 1.887 | 1.9 | 14.34 | 1462.32 | 1447.98 | 17.42 |
| 2 | 1.481 | 3.41 | 15.56 | 1464.28 | 1448.72 | 20.25 |
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.