HBO043%DJUKFA11.BITNET@cunyvm.cuny.edu (Christoph van Wuellen) (06/22/91)
I have a modified version of the 68k code generator which knows about singles and doubles (32 and 64 bits). It calls its own internal library functions. C functions which return a float return is in D0, they return a double in D0/D1, so it should be compatible with other compilers. The internal support functions use a pointer technique, but I can take any choice w.r.t internal functions. If there is interest, I could post this gen68k version. I have no support functions yet, but with some small wrappers it should be possible to use the ACK FP support functions. C.v.W.
breure@tno.nl (Frank Breure) (06/24/91)
From: Frank Breure <breure@itihp1.tno.nl> > > The internal support functions use a pointer technique, but I can take > any choice w.r.t internal functions. If there is interest, I could post > this gen68k version. I have no support functions yet, but with some small > wrappers it should be possible to use the ACK FP support functions. > I am interested in this new 'gen68k', if you don't send it out to the newsgroup/mailing list please mail me one (see adress in .signature, sometimes some hosts really mess up the reply and/or from fields in my mail). I am willing to try the port for the ACK FP package. > C.v.W. > -- Frank Breure (breure@itihp1.tno.nl) TNO Institute of Applied Computer Science Delft, The Netherlands ---------------------------------------------------------------------- My opinions may be subject to change without notice.
v892544@si.hhs.nl (Witmond) (06/25/91)
Hello world, I have a few questions: 1 How much memory is needed for C86? 2 Where can I get it? 3 Do I get the sources? 4 Can I load object files from C86 and ACK? (to load math parts with other parts). Now my opinions: 1 If there is a standard, USE it. 2 If there are two standards take the less restricting, although it may give more or slower code in the programs as newer computers will be able to handle those standards. 3 If there is no standard at all define one or more and reread thes opinions. Guido M. Witmond Internet: v892544@si.hhs.NL
Joerg Conradt <jac@unlisys.in-berlin.de> (06/26/91)
> > I have a modified version of the 68k code generator which knows about > singles and doubles (32 and 64 bits). > It calls its own internal library functions. > C functions which return a float return is in D0, they return a double > in D0/D1, so it should be compatible with other compilers. > > The internal support functions use a pointer technique, but I can take > any choice w.r.t internal functions. If there is interest, I could post > this gen68k version. I have no support functions yet, but with some small > wrappers it should be possible to use the ACK FP support functions. > > C.v.W. > hello out there, i would very much like you to post it, so many st-user will use it. thanx joerg -- Joerg Conradt Berlin, Germany || UUCP: jac@unlisys.in-berlin.de
HBO043%DJUKFA11.BITNET@cunyvm.cuny.edu (Christoph van Wuellen) (06/26/91)
I posted c68 here in this group, you get the full sources. ACK and c68 objects can be mixed except for a different floating point format (THIS RESTRICTION TO BE REMOVED WITHIN THE NEAR FUTURE). c68 and ACK use different support functions (for ACK specialists: why does .mli calculate a 64-bit result - slow and never needed) C.v.W.
HBO043%DJUKFA11.BITNET@cunyvm.cuny.edu (Christoph van Wuellen) (06/26/91)
will post it soon... had to prepeare a seminar lecture for today. C.v.W.