I do a lot of real number problems which are extreemly comput intensive (up
to 6 hours run time using FFP) and am looking for a compiler for a
structured, modular language like Ada or MODULA2 which supports the 68882
coprocessor for math opperations. I have not been able to find any
information about which compilers do or do not have this support. Can
anybody tell me if M2S or Benchmark M2 support the copprocessor of my 2500?
Neither compiler offers direct math chip support but it is my understanding
that if you use the LONGREAL type in either compiler (M2Sprint I know for
sure) it will call IEEE libraries in libs: which DO use the 68881/68882 if
present. — Art
Gary:
Benchmark does not directly support the 68882 however using an Amiga 2500
the FFP code compares favorably with LONGREAL code but without the extra
precision. If you are currently using an interpretive language (i.e. BASIC)
this could be most of the delay you are experiencing.
>> Leon Frenkel <<