I was expecting the False Positive rate to stay below the level the bloomfilter was initialized with until the bloomfilter reaches capacity, but instead it vastly exceeds it.
Create a bloomfilter with 400 million capacity, FPP of 0.001, put 100 million things into the bloomfilter, then test 1 million things that weren't put into the bloomfilter.
I expected to find 1,000 or less false positives, but instead I find 23,242.
I get the exact same false positive count when I initialize the bloomfilter with a capacity of 1.6 billion.
While experimenting I also tried creating a bloomfilter with capacity of 3.2 billion, but that resulted in a very different error that I'll open a separate issue for.
I have the same issue. I use in production bloomfilter created with b = Blex.new(300000000, 0.0001) and containing 250 million things. Just tested 1 million things and received 56784 false positives.
I was expecting the False Positive rate to stay below the level the bloomfilter was initialized with until the bloomfilter reaches capacity, but instead it vastly exceeds it.
Create a bloomfilter with 400 million capacity, FPP of 0.001, put 100 million things into the bloomfilter, then test 1 million things that weren't put into the bloomfilter.
I expected to find 1,000 or less false positives, but instead I find 23,242.
I get the exact same false positive count when I initialize the bloomfilter with a capacity of 1.6 billion.
While experimenting I also tried creating a bloomfilter with capacity of 3.2 billion, but that resulted in a very different error that I'll open a separate issue for.