CPU scaling benchmark
workers
3 +1 main
iters total
500M
125000000/stream
elapsed
1140.32 ms
total CPU used
4392.52 ms
speedup
3.85×
vs serial
efficiency
96.3%
of 4× ideal
| stream | spawn ms | spawned@ | work start@ | work end@ | work ms | reap wait ms |
|---|---|---|---|---|---|---|
| 0 (main) | 0 | 5.12 | 5.13 | 1088.62 | 1083.49 | 0 |
| 1 | 1.996 | 2.01 | 14.62 | 1104.87 | 1090.25 | 16.39 |
| 2 | 1.526 | 3.56 | 15.7 | 1114.13 | 1098.43 | 25.63 |
| 3 | 1.532 | 5.11 | 16.86 | 1137.21 | 1120.35 | 48.73 |
main
w1
w2
w3
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.