floating-point number - hardware or lisp dependent?
Subject: floating-point number - hardware or lisp dependent?
From: Raymond Toy
Date: Thu, 04 Apr 2013 13:47:56 -0700
>>>>> "Volker" == Volker van Nek <volkervannek at gmail.com> writes:
Volker> build_info();
Volker> gives the information you want. Below there are some examples.
Volker> If I am not mistaken Maxima on Windows is compiled with GCL, Ubuntu uses a SBCL-compiled version and
Volker> I believe MAC uses SBCL too.
Volker> Surprisingly CMUCL doesn't like 5e-324, 4e-324, 3e-324.
Neat. CMUCL itself can handle 5d-324. 4d-324 is smaller than the
smallest representable float so an error is produced instead of
silently returning 0. However, for 5d-324, it looks like a bug
printing. Maxima tries (format t "~ve" 21 5d-324), which causes the
error.
Ray