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
10 +1 main
iters total
500M
45454545/stream
elapsed
1172.67 ms
total CPU used
11693.34 ms
speedup
9.97×
vs serial
efficiency
90.6%
of 11× ideal
stream spawn ms spawned@ work start@ work end@ work ms reap wait ms
0 (main) 0 30.59 30.6 1067.48 1036.88 0
1 2.016 2.03 15.56 1124.44 1108.88 57.06
2 1.496 3.56 35.17 1069.51 1034.34 12.75
3 1.486 5.06 28.77 1099.94 1071.17 32.57
4 2.756 7.84 28.85 1101.19 1072.34 45.4
5 1.86 9.71 50.28 1169.72 1119.44 102.39
6 2.214 11.95 50.32 1137.03 1086.71 69.65
7 1.609 13.58 51.19 1110.43 1059.24 45.41
8 7.439 21.03 70.34 1125.48 1055.14 60
9 2.023 23.07 58.16 1081.41 1023.25 16.87
10 7.497 30.58 63.04 1088.99 1025.95 21.6
main
w1
w2
w3
w4
w5
w6
w7
w8
w9
w10
    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.