Calculations run on this device. Scenario values are not sent anywhere — the engine is fully local.
system design calculator · free
Bloom Filter Tuner
Size the bit array and hash count for your error budget — and watch the trade curve.
How many bits buy the false-positive rate you need?
Engine 1.0.0 · calculation
5 (10^5 = 100,000)
1 %
117.0KiB of bitmap
- Bits required (m)
- 958,506
- Hash functions (k)
- 7
- Bits per item
- 9.59
- Actual FPR
- 1.004%
m = −n·ln(p) / (ln 2)²; k = (m/n)·ln 2. Halving the error target costs a constant multiplicative memory — there is no free precision.
Method
- Optimal bitmap size m = −n·ln(p)/(ln 2)² and hash count k = (m/n)·ln 2.
- Actual FPR computed back from (1 − e^(−kn/m))^k so you can check the theory.