murmur-delphi

Murmur hash for Delphi

MurmurHash is a fast, non-cryptographic, hash, suitable for hash tables.

It comes in multiple variants:

• hash result of 32 bits
• hash result of 64 bits
• hash result of 128 bits, optimized for x86 architecture
• hash result of 128 bits, optimized for x64 architecture

Sample Usage

MurMur supports any other kind of data:

var
  hash:    Cardinal;
  hash128: string;
  find:    WIN32_FIND_DATA;
  find2:   string;

hash := TMurmur3.Hash_x86_32(find, SizeOf(WIN32_FIND_DATA), $ba5eba11);
hash := TMurmur3.Hash_x86_32(Pointer(find2)^, Length(find2) * SizeOf(Char), 123);
// These are CPU target optimized
hash := TMurmur3.Hash(find, SizeOf(WIN32_FIND_DATA), $ba5eba11);
hash := TMurmur3.Hash(Pointer(find2)^, Length(find2) * SizeOf(Char), 123);

Test Setting

Input:  9084 ANSI characters from a Base64 encoded text (Found in MurMurHashVsTests.pas Ln: 260)
CPU:    Intel Core I5-8600K @ 4.2 GHz
Memory: 32GB DDR3 @ 2666 MHz in Dual-Channel

Test Results

x86 build results:

MurMur1.Hash > Hashed 1000x in 21.151 µs average

MurMur2.Hash > Hashed 1000x in 17.733 µs average

MurMur2.HashA > Hashed 1000x in 17.723 µs average

MurMur2.Hash64A > Hashed 1000x in 26.601 µs average

MurMur2.Hash64B > Hashed 1000x in 15.480 µs average

MurMur3.Hash > Hashed 1000x in 18.331 µs average

Adler32CRC > Hashed 1000x in 31.032 µs average

BobJenkins > Hashed 1000x in 24.993 µs average

x64 build results:

MurMur1.Hash > Hashed 1000x in 20.777 µs average

MurMur2.Hash > Hashed 1000x in 18.788 µs average

MurMur2.HashA > Hashed 1000x in 18.634 µs average

MurMur2.Hash64A > Hashed 1000x in 9.671 µs average

MurMur2.Hash64B > Hashed 1000x in 15.529 µs average

MurMur3.Hash > Hashed 1000x in 10.302 µs average

Adler32CRC > Hashed 1000x in 30.083 µs average

BobJenkins > Hashed 1000x in 24.766 µs average