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
1184.52 ms
total CPU used
9247.66 ms
speedup
7.81×
vs serial
efficiency
86.8%
of 9× ideal
stream spawn ms spawned@ work start@ work end@ work ms reap wait ms
0 (main) 0 24.16 24.18 1062.48 1038.3 0
1 1.929 1.98 25.03 1108.29 1083.26 48.99
2 1.513 3.53 19.42 942.15 922.73 0.13
3 1.571 5.13 25.01 1133.74 1108.73 71.42
4 3.461 8.61 33.77 1102.23 1068.46 39.88
5 3.198 11.83 47.79 1107.89 1060.1 45.52
6 8.721 20.57 56.03 1181.55 1125.52 119.22
7 1.948 22.54 82.31 914.01 831.7 0.18
8 1.573 24.14 62.88 1071.74 1008.86 9.37
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.