CPU scaling benchmark
workers
1 +1 main
iters total
500M
250000000/stream
elapsed
2189.38 ms
total CPU used
4331.13 ms
speedup
1.98×
vs serial
efficiency
99%
of 2× ideal
| stream | spawn ms | spawned@ | work start@ | work end@ | work ms | reap wait ms |
|---|---|---|---|---|---|---|
| 0 (main) | 0 | 1.93 | 1.94 | 2161.99 | 2160.05 | 0 |
| 1 | 1.894 | 1.91 | 15.58 | 2186.66 | 2171.08 | 24.82 |
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.