soccer jersey forums.silverfrost.com :: View topic - Errors in some bit operations with 8-byte integers
forums.silverfrost.com Forum Index forums.silverfrost.com
Welcome to the Silverfrost forums
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Errors in some bit operations with 8-byte integers

 
Post new topic   Reply to topic    forums.silverfrost.com Forum Index -> Support
View previous topic :: View next topic  
Author Message
mecej4



Joined: 31 Oct 2006
Posts: 1895

PostPosted: Tue Jan 16, 2024 1:33 pm    Post subject: Errors in some bit operations with 8-byte integers Reply with quote

The late George Marsaglia posted his SuperKiss-64 RNG code in the "General" forum in 2009 ( http://forums.silverfrost.com/viewtopic.php?t=1480 ). At the time (and, perhaps, even now), 32-bit code generated by FTN95 used the FPU for 8-byte integer operations, which leads to lots of FPU exceptions, so we use only /64 for this code.

Marsaglia's test program generates a billion 8-byte random integers and compares the last of those to a known value. With FTN95 8.972 and 9.0, the test fails. With 8.9, 8.92, 8.95, the test passes. Here is a very short program that contains a small portion of Marsaglia's code, just enough to display the bug.

Code:
! Suitable only for 64-bit compilation
! Some of the last digits are wrong with Versions 8.972, 9.0
! Correct output with Versions 8.9, 8.92, 8.95, as follows
!   i  h          z                 carry            q(i)
!  1  1 -2945433577893571930   863862222379  5890867155787143858
!  2  1  6661552707265476373  1293535672257  5123638659178598868
!  3  1 -4642211112981496864    93621267173 -9162321847746557890
!  4  1 -8101377823381285518  1808705431682 -2243988426946980582
!
program S64BITOPS
   implicit None
   Integer, Parameter :: QSIZ = 4
   Integer, Parameter :: i8 = selected_int_kind(18)
   Integer (Kind=i8) :: i, z, h, q(QSIZ)
   Integer (Kind=i8), Parameter :: CMUL = 6906969069_i8, COFFS = 123_i8
   Integer (Kind=i8) :: carry = 36243678541_i8, &
        xcng = 12367890123456_i8, xs = 521288629546311_i8

   Do i = 1, qsiz                 ! fill Q with Congruential + Xorshift
      xcng = xcng*CMUL + COFFS    ! ignore integer overflow here
      xs = ieor(xs, ishft(xs,13))
      xs = ieor(xs, ishft(xs,-17))
      xs = ieor(xs, ishft(xs,43))
      q(i) = xcng + xs
   End Do
   print *,' i  h          z                 carry            q(i)'
   Do i = 1, QSIZ
      h = iand(carry, 1_i8)
      z = ishft(ishft(q(i),41), -1) + &
          ishft(ishft(q(i),39), -1) + &
          ishft(carry, -1)
      carry = ishft(q(i), -23) + ishft(q(i), -25) + ishft(z, -63)
      q(i) = not(ishft(z,1)+h)
      print '(2i3,i21,i15,i21)',i,h,z,carry,q(i)
   End Do
end program


Last edited by mecej4 on Sat Feb 03, 2024 8:53 am; edited 1 time in total
Back to top
View user's profile Send private message
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 7941
Location: Salford, UK

PostPosted: Tue Jan 16, 2024 1:55 pm    Post subject: Reply with quote

mecej4

Thanks for the feedback which I have logged for investigation.
Back to top
View user's profile Send private message AIM Address
mecej4



Joined: 31 Oct 2006
Posts: 1895

PostPosted: Tue Jan 16, 2024 4:26 pm    Post subject: Reply with quote

Paul, I did not mean to post this in the FTN77 Support forum!

I can repost in the 64-bit forum, where it belongs, but I am afraid that if I did so your reply may be left behind in the FTN77 forum.
Back to top
View user's profile Send private message
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 7941
Location: Salford, UK

PostPosted: Tue Jan 16, 2024 7:40 pm    Post subject: Reply with quote

No problem either way.
Back to top
View user's profile Send private message AIM Address
PaulLaidler
Site Admin


Joined: 21 Feb 2005
Posts: 7941
Location: Salford, UK

PostPosted: Fri Feb 02, 2024 5:20 pm    Post subject: Reply with quote

This regression (for 64 bit FTN95) has been fixed for the next release of FTN95. It concerns ishft(z, -63) and ishft(z, 63) for INTEGER*8 and ishft(z, -31) and ishft(z, 31) for INTEGER*4.
Back to top
View user's profile Send private message AIM Address
Display posts from previous:   
Post new topic   Reply to topic    forums.silverfrost.com Forum Index -> Support All times are GMT + 1 Hour
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group