22 Oct
2019
22 Oct
'19
10:29 a.m.
[Redirecting to COFF]
On 2019-Oct-20 00:02:56 +0530, Abhinav Rajagopalan <abhinavrajagopalan(a)gmail.com>
wrote:
Forgive me for both hijacking this thread, and to
address my amateurish
gnawing concern, but how was it be possible to write differential/integral
equations at an assembly/machine level at the time, especially in machines
such as the PDP-7 and such which had IIRC just 16 instructions and operated
on the basis of mere words, especially the floating point math being done.
My 1st edition Wilkes, Wheeler, Gill[1] documents that, by 1951, EDSAC[2]
had a floating-point library that supported addition, subtraction and
multiplication (no division) of numbers with 23-27 bits of precision and a
range of 1e-63 to 1e63. EDSAC was much less powerful than a PDP-7.
Writing a floating-point library is not that difficult, though getting
the rounding correct for all the edge cases is tricky. Actually using
floating-point and avoiding the pitfalls can be harder - see (eg)
https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html (though
https://floating-point-gui.de/ may be more approachable).
[1] https://archive.org/details/programsforelect00wilk
[2] https://en.wikipedia.org/wiki/EDSAC
--
Peter Jeremy