[clean-list] Re: (clean-list digest:) Functions that look the same (smithll)

[John van Groningen]

I tried the same on my Macintosh (PowerPC 603e RISC).
>
>In *both* cases I got as answer: 9.88131291682493e-324

Of course. Because you will only get different results if computations
in floating point registers are more accurate than computations on 64-bit
values in memory. So this may happen on ia32 or 68k processors, but not on
PowerPC processors.

>I made a small C program with MPW (the native C/C++ Apple-compiler):
>
>...
>
>That program delivers as results (notice %1.14e in printf):
>
>...
>9.88131291682493e-323
>9.88131291682493e-324
>0.00000000000000e+00
>
>One thing bothers me is the observation that Clean turns over the
>double-results with only 14 significant places (any documentation somewhere in
>order to clarify this?).

Actually I count 15 significant digits, one before and 14 after the dot.
There are 52 significant bits in a 64-bit floating point value, so there are
15 significant digits.

Regards,
John van Groningen