O2 Hashing functions which is better to use

[Chris Chancellor]

Hello Charles

i dug thru' the O2 example codes and O2 forum and found 2 hashing functions which are listed below

normal O2 hash code

Code Select Expand


'===================================
'  O2  Hash
'  This gives a very huge number which is more difficult
'  for comparison purposes than O2 ASM Hash
Function O2Hash(string GivStr) as sys
if not GivStr then
'  exit if blank string
   return 0
end if

byte b at (strptr GivStr)
sys  LengS=len GivStr
sys  a=LengS*256
sys  i
for i=1 to LengS
  a = a xor 0x3a575a73 xor b
  rol a,18
  @b++
next
return a
End Function




O2 Hash  in assembly

Code Select Expand


'===================================
'  O2  ASM Hash
'  https://www.oxygenbasic.org/forum/index.php?topic=586.msg6461#msg6461
'  reply #4  this is the better Hashing as it gives a smaller number
Function O2HashASM(string GivStr) as Sys
  xor eax,eax
  mov edx, [GivStr]     ' name pointer
  mov ecx, [edx-4] ' length
  (
  cmp ecx,0
  jnz exit
  return 0 'EMPTY WORD WILL TRIGGER HASHCODE ERROR LATER
  )
  mov ah,  cl      ' assume max length 255
  (
    xor al,  [edx]
    xor eax,0x3a575a73
    rol eax,18
    inc edx
    dec ecx
    jnz repeat
  )
  return eax
  End Function



seemingly the O2 Hash in ASM is a better one amongst these 2 sets of codes as
it gives a hash some thing like  289133459   as compare to the normal hash which give huge results of  5.19993 E+18

What do you suggest to use  O2 hash in ASM  or  O2 hash normal ?

so far i have tested both functions out and found no problem or collision errors from small strings  ( 6 letters)  to very
huge strings ( 83 MB file strings )

[Charles Pegge]

Hi Chris,

The assembler version always generates a 32bit hash, whereas the basic version generates a 64bit hash in 64bit mode.

How unique does your hash need to be?

[Chris Chancellor]

Thanxx a lot Charles

so far from my tests results (over 100 strings and files) , the O2 Hash ASM did not fail  (ie. no collisions ) and pretty fast too
i will stick to    O2 Hash ASM  as it provides  unique  and manageable numbers.

i was using FNV hashing in PB which is not so good as strings with less than 7 characters encounter collisions


i wonder what method or algorithm that this O2 ASM hash was based on such as SHA256 , RC4 or AES  ?

[Charles Pegge]

It's just cheap and cheerful hashing for data storage and id. Not suitable for cryptography.

[Chris Chancellor]

Nevermind, my work does not need sopshicated cryptography, i just need to have a simple and
normal check on some file sizes  using normal hashing method as effected by O2 ASM Hash.

not for those server or database logins and password authentication methods which i don't use at all
so this hash function will do