• Re: Fortran, old and slow base and bounds, Why I've Dropped In

    From John Levine@21:1/5 to All on Thu Jun 19 18:03:52 2025
    According to Thomas Koenig <tkoenig@netcologne.de>:
    EricP <ThatWouldBeTelling@thevillage.com> schrieb:

    Note also that 360 index register was not scaled and so not
    directly usable from array index value for other than byte arrays.

    There is always strength reduction. It seems the original FORTRAN
    compiler did a lot of that for the 704, but I'm not sure that the
    /360 compilers did - from what I read, they regressed in code
    generation quality.

    IBM had two FORTRAN compilers for the /360, G and H, named after the amount
    of memory they needed, G needed a 128K system, H needed 256K.

    The code from FORTRAN G was pretty bad, but it compiled quickly so it's
    what you used for debugging. FORTRAN H compiled much slower but its
    code was very good, considered state of the art at the time, I presume
    aided by people who'd worked on the original FORTRAN. They used the
    same calling sequences and data formats so you could combine code from
    the two, e.g., library routines compiled by H along with your debug
    runs using G.

    There was also WATFOR/WATFIV for student load and go programs, and
    subset FORTRAN E (32K) for small systems.

    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)