A89: Re: Re: Back to arrays... =(


[Prev][Next][Index][Thread]

A89: Re: Re: Back to arrays... =(




Hehe, another question? =)

Where do you get the info on the timing of each op code and register access
and all?

-Miles Raymond      EML: m_rayman@bigfoot.com
ICQ: 13217756       IRC: Killer2        AIM: KilIer2 (kilier2)
http://www.bigfoot.com/~m_rayman/

----- Original Message -----
From: Zoltan Kocsi <zoltan@bendor.com.au>
To: <assembly-89@lists.ticalc.org>
Sent: Wednesday, July 28, 1999 2:37 AM
Subject: A89: Re: Back to arrays... =(


>  > Now, one more question: Is the 'array+2' meathod better than the
'2(a6)'
>  > meathod?  It is faster on the calc since it doesn't have to add it
while
>  > running, correct?
>
> Actually, no. The n(ax) form has an extension word which contains the
> offset. It is one word. The absolute address is stored in 2 extension
> words. On the 68000 reading a word from memory takes 4 clock cycles (min).
> Therefore, reading the instruction itself is 4 clocks more in case of
> the absolute addres version. The penalty that you pay for the adding
> in the n(ax) addressing mode is IIRC 2 clocks (I don't have the 68000
> insn timing handy now). All in all, you are 2 clocks and 1 word better
> off using the register. Of course, you had to load that register but if
> you use the address often, then it definitelly worth to use a register.
>
> Regards,
>
> Zoltan



Follow-Ups: References: