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
1145.14 ms
total CPU used
9397.19 ms
speedup
8.21×
vs serial
efficiency
91.2%
of 9× ideal
stream spawn ms spawned@ work start@ work end@ work ms reap wait ms
0 (main) 0 14.82 14.83 1079.86 1065.03 0
1 2.172 2.18 32.44 1119.47 1087.03 39.79
2 1.689 3.89 20.48 1110.23 1089.75 33.08
3 1.589 5.5 22.12 1012 989.88 0.14
4 1.6 7.11 33.81 1142.05 1108.24 62.31
5 1.616 8.74 43.06 1120.49 1077.43 42.65
6 1.633 10.39 41.77 1082.16 1040.39 7.54
7 2.714 13.12 43.03 909.77 866.74 0.19
8 1.666 14.8 63.05 1135.75 1072.7 56.04
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.