CPU scaling benchmark
workers
2 +1 main
iters total
500M
166666666/stream
elapsed
1471 ms
total CPU used
4348.3 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 | 3.35 | 3.36 | 1448.2 | 1444.84 | 0 |
| 1 | 1.846 | 1.87 | 15.7 | 1467.97 | 1452.27 | 22.13 |
| 2 | 1.446 | 3.34 | 16.02 | 1467.21 | 1451.19 | 19.18 |
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.