CPU scaling benchmark

W=0 W=1 W=2 W=3 W=4 W=6 W=8 W=9 W=10 W=12 W=16 iters: 100M 500M 1B 2B
workers
8 +1 main
iters total
500M
55555555/stream
elapsed
1206.64 ms
total CPU used
8970.83 ms
speedup
7.43×
vs serial
efficiency
82.6%
of 9× ideal
stream spawn ms spawned@ work start@ work end@ work ms reap wait ms
0 (main) 0 20.97 20.99 982.6 961.61 0
1 2.042 2.06 20.63 664.58 643.95 0.13
2 1.498 3.58 21.93 1203.93 1182 221.46
3 1.43 5.03 23.36 920.8 897.44 0.19
4 5.239 10.29 43.96 1033.83 989.87 51.33
5 1.649 11.95 57.94 1070.63 1012.69 88.13
6 1.378 13.35 60.51 1157.97 1097.46 177.88
7 1.381 14.74 50.62 1159.27 1108.65 180.68
8 6.197 20.96 46.73 1123.89 1077.16 141.41
main
w1
w2
w3
w4
w5
w6
w7
w8
    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.