[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gcl-devel] gcl vs sbcl vs ccl
From: |
Raymond Toy |
Subject: |
Re: [Gcl-devel] gcl vs sbcl vs ccl |
Date: |
Mon, 12 Oct 2015 08:21:24 -0700 |
User-agent: |
Gnus/5.101 (Gnus v5.10.10) XEmacs/21.5-b34 (linux) |
>>>>> "Gunter" == Gunter Königsmann <address@hidden> writes:
Gunter> The things I have solved by switching to a 64-bit linux so far are:
Gunter> - misusing wxMaxima as a replacement of excel that isn't
Gunter> limited to a million cells or so.
A million cells doesn't seem so big, even for a 32-bit lisp.
Gunter> - worst-case-tolerance calculations: I frequently use
Gunter> maxima to find out which equation describes the system I
Gunter> am working with. Then I make a list of all circuit element
Gunter> values including the tolerance band they can lie
Gunter> within. Then I use a package I have written myself to
Gunter> iterate all possible combinations of element values and
Gunter> try to determine the minimum and maximum value I can
Gunter> possibly get this way. If this involves an "if" it is
Gunter> 64-bits for sure.
This seems more easily solved by doing interval arithmetic or even
Monte-Carlo simulations.
Gunter> - plotting a set of solutions for differential equations
Gunter> that are barely within the limits desolve finds solutions
Gunter> in
Numerical solutions don't work?
Gunter> One could argue if most of this falls into the categories
Gunter> "misuse of maxima" and "ugly workaround for
Gunter> bugs". Determining where an oscillogram comes from and
Gunter> tolerance calculations most probably don't. But - since
Gunter> 64-bit get me rid of most of the problems I might
Gunter> encounter I would still be grateful if they existed.
I wouldn't call them misuse.
--
Ray