[Prev][Next][Index][Thread]
Re: A85: Why do anything and another question.
The TI-ROM is not very optimized ! I have seen disassembled version of
both the TI82,83 and 85 ROM, and none of them are optimized. The FP math
instructions might be optimized (I have not checked), but the rest of the
ROM is not. I guess TI assumed that no one would ever see the ROM so they
did not bother to optimize it.
Dines
BTW To give an exampel of how well written the ROM is, ithought I would
mention a thing I found in a variable function.
; A hold length of var name
CP Max_Var_name_Length
loop:
JR C,loop ; If length is too long then CRASH !
-----Original Message-----
From: Ben Sferrazza <bens@eden.rutgers.edu>
To: assembly-85@lists.ticalc.org <assembly-85@lists.ticalc.org>
Date: 12. marts 1998 00:38
Subject: Re: A85: Why do anything and another question.
>
>Erik L Gillespie wrote:
>>
>> The ROM square root would be helpful but I'd like one as optimized as
>> possible and the one in ROM definitely is not.
>>
>> A 32-bit floating point can't be really fast but an optimized ASM
version
>> could be fastER than the default one on the calc. Simple things like
>> plotting a point have even been proven to be faster with optimized asm
in
>> something such as ZShell.
>
>What do think the ROM code is written in? The ROM code is written in
>the Z80's native machine code, so barring incompetence, TI's routines
>are probably as good as any other. Considering square root routines are
>useful for many platforms, TI probably modified an existing square root
>routine to work with its specific floating-point number format. So if
>you use TI's routines, you not only don't have to write them yourself,
>but you can also store floating-point variabes in your program in TI's
>format.
>
>Ben
>
_______________________________________
Dines Justesen
Email: dines@post1.com or
c958362@student.dtu.dk
WWW : http://www.gbar.dtu.dk/~c958362/
_______________________________________
Follow-Ups: