CPU scaling benchmark
workers
1 +1 main
iters total
100M
50000000/stream
elapsed
453.13 ms
total CPU used
867.88 ms
speedup
1.92×
vs serial
efficiency
96%
of 2× ideal
| stream | spawn ms | spawned@ | work start@ | work end@ | work ms | reap wait ms |
|---|---|---|---|---|---|---|
| 0 (main) | 0 | 1.97 | 1.98 | 434.68 | 432.7 | 0 |
| 1 | 1.941 | 1.96 | 14.82 | 450 | 435.18 | 15.47 |
main
w1
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.