Benchmarks

Generated by olab-study bench. Times exclude first-call compilation, which has its own row. Each is the median of five runs on one core's worth of work, on the machine below.

Intel64 Family 6 Model 183 Stepping 1, GenuineIntel, 32 logical cores, 127.7 GB RAM, Windows-11-10.0.26200-SP0, Python 3.12.10, Numba 0.68.1.

Requirement	Case	Seconds	Target (s)	Met
NFR-01	pbo_cscv, T = 1,260, N = 1,000, S = 16	0.396	10	yes
NFR-02	reality_check, T = 1,260, N = 1,000, B = 1,000	0.150	60	yes
NFR-02	spa_test, T = 1,260, N = 1,000, B = 1,000	0.230	60	yes
NFR-02	romano_wolf_stepdown, T = 1,260, N = 1,000, B = 1,000	0.228	120	yes
NFR-03	run_strategies, 1,034 variants, 5,000 bars	0.100	5	yes
NFR-03	first-call JIT compile	2.453	30	yes