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
6 +1 main
iters total
500M
71428571/stream
elapsed
1209.95 ms
total CPU used
7006.22 ms
speedup
5.79×
vs serial
efficiency
82.7%
of 7× ideal
stream spawn ms spawned@ work start@ work end@ work ms reap wait ms
0 (main) 0 13.3 13.31 938.02 924.71 0
1 2.219 2.23 14.64 1055.3 1040.66 119.38
2 1.581 3.84 22.97 1123.25 1100.28 185.36
3 2.167 6.02 17.73 767.3 749.57 0.13
4 1.788 7.82 37.74 1076.24 1038.5 138.35
5 3.817 11.65 65.06 1207.01 1141.95 269.13
6 1.628 13.29 51.52 1062.07 1010.55 126.67
main
w1
w2
w3
w4
w5
w6
    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.