Add a nTweak to bloom filters to tweak the seed.

This commit is contained in:
Matt Corallo
2012-11-02 18:33:50 -04:00
parent 4c8fc1a588
commit b1f99bed6f
3 changed files with 57 additions and 24 deletions

View File

@@ -15,7 +15,7 @@ using namespace std;
static const unsigned char bit_mask[8] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80};
CBloomFilter::CBloomFilter(unsigned int nElements, double nFPRate) :
CBloomFilter::CBloomFilter(unsigned int nElements, double nFPRate, unsigned int nTweakIn) :
// The ideal size for a bloom filter with a given number of elements and false positive rate is:
// - nElements * log(fp rate) / ln(2)^2
// We ignore filter parameters which will create a bloom filter larger than the protocol limits
@@ -23,14 +23,15 @@ vData(min((unsigned int)(-1 / LN2SQUARED * nElements * log(nFPRate)), MAX_BLOOM
// The ideal number of hash functions is filter size * ln(2) / number of elements
// Again, we ignore filter parameters which will create a bloom filter with more hash functions than the protocol limits
// See http://en.wikipedia.org/wiki/Bloom_filter for an explanation of these formulas
nHashFuncs(min((unsigned int)(vData.size() * 8 / nElements * LN2), MAX_HASH_FUNCS))
nHashFuncs(min((unsigned int)(vData.size() * 8 / nElements * LN2), MAX_HASH_FUNCS)),
nTweak(nTweakIn)
{
}
inline unsigned int CBloomFilter::Hash(unsigned int nHashNum, const std::vector<unsigned char>& vDataToHash) const
{
// 0xFBA4C795 chosen as it guarantees a reasonable bit difference between nHashNum values.
return MurmurHash3(nHashNum * 0xFBA4C795, vDataToHash) % (vData.size() * 8);
return MurmurHash3(nHashNum * 0xFBA4C795 + nTweak, vDataToHash) % (vData.size() * 8);
}
void CBloomFilter::insert(const vector<unsigned char>& vKey)