undefined | Better HN

0 pointsajross9y ago0 comments

Do the hash tables in use by the C/C++ entries use low-security hash functions? Seems like that needs evidence.

0 comments

The first-place Rust program uses this very simple low-security hash function:

    impl Hasher for NaiveHasher {
        fn write_u64(&mut self, i: u64) {
            self.0 = i ^ i >> 7;
        }
    }

The second-place C program uses the exact same hash function as the Rust program, except it also truncates the result to 32 bits:

    #define CUSTOM_HASH_FUNCTION(key) (khint32_t)((key) ^ (key)>>7)

The third-place C++ program uses the identity function as its hash function:

    struct hash{
        uint64_t operator()(const T& t)const{ return t.data; }
    };

Sources:

j / k navigate · click thread line to collapse

The first-place Rust program uses this very simple low-security hash function:

    impl Hasher for NaiveHasher {
        fn write_u64(&mut self, i: u64) {
            self.0 = i ^ i >> 7;
        }
    }

The second-place C program uses the exact same hash function as the Rust program, except it also truncates the result to 32 bits:

    #define CUSTOM_HASH_FUNCTION(key) (khint32_t)((key) ^ (key)>>7)

The third-place C++ program uses the identity function as its hash function:

    struct hash{
        uint64_t operator()(const T& t)const{ return t.data; }
    };

Sources:

j / k navigate · click thread line to collapse