On Sun, 23 Apr 2000 12:57:15 +0200, Laurens Holst wrote:
>UNTRUE!!!
>Those are the same as the Z80 versions, but work different in another mode
>(or havea so-called Data Directive (DDIR) opcode).
What I've tried to say (but I think I had failed) that is:
most instructions NEW on Z380 will doesn't make any sense on Z180.
This is why they are not there.
>> 1 time : 32 bit opcodes for the additional registers.
>Those are very useful. Included are NOT ONLY instructions for handling the
>additional register banks (you are greatly mistaking here, the register
>banks are handled by internal I/O). New instructions feature relative calls,
>EX almost all register sets, multiplication and dividing, and much more.
"new" Relative calls are about the 32 bit addresses. It doesn't make sense
on a 8 bit processor (such as Z180).
"new" EX are relative to new registers. It doens't make sense on a processor
that doesn't have such registers.
Mult and Div is something new. So, the good features of Z380 are the
Mult, Div, New registers and a new address space. The "other" new
functions are depending on those enhancements... (and would be a
shame if they don't exist!)
>> I was told it had... so if hadn't... it's sad... (((-;
>You should really (again) read the Z380 docs!!!
I know Z380. (-; You just had not understood what I was trying to
say. (-; Of course I know Z380 is very good (and this is why I fought
to use it in the past), because it's easier to program due to
lots of registers and a big address space. More instructions are
just consequence. (-;
What I really think on instructions enhancements on Z380 is the
FPU part. BTW, It was designed to be FPU, and this is why it
has a large address apace (working with large matrix) and a lot
of registers (speed up math processing).
>(and maybe some Z180 too for comparisation)
These I really need to read.
>No. The multiplication instructions are only faster when both values are
>unknown, because that would require a quite slow routine, which can be
>replaced by an instruction. However, when one of the values is known (in
>this case, the 9), you can do it faster by doing something like:
>ld b,a
>add a,a
>add a,a
>add a,a
>add a,b
>At least, on the R800 that's true.
I think it's not on Z180.
>I just reviewed the docs, and a MLT would take 7 cycly, while the above
>would take 10 cycly. But then again, I never said it was unuseful (in fact,
>I referred to MLT and TST as being the only useful instructions).
(((-;
>> Yes, it has, but almost all of them refers to new resiters...
>> (like I said a above).
>WRONG!!!
I'm not wrong. (-; Z380 has lots of new things that cannot be used
and remains compatible, like it's DMA (but even Z80 had DMA, and
was not used) and some other things. I'm just putting my "enthusiasm"
by side and looking what we will really need and what will be throw
processor throught the window.
>Indeed. It will be slow, but at least it will work... Haha stupid Intel!!!
>If they thought of this (which is really easy to implement in a chip), then
>they could have released software-patches for MMX, so that all software
>could be released for MMX instead of using MMX-specific parts of code...
I think x86 (starting with 286) does TRAP. They just doesn't had implemented
the patchs. On PC, when a trap occours (non-existent opcode) the computer hangs,
AFAIK.
>> If they cause a trap, we can make a workaround.
>On the Z180 they do, on the Z380 there are other instructions there.
This is a problem. No Z180 instruction conflicts with R800 instructions?
>It doesn't appear that you read the Z380 specs... Read them again please...
>(thoroughly!).
I had read, but I don't know where you found so many "new actions" Z380,
that are not realted to the new address space and/or new registers. I'M NOT
SAYING THAT IT DOESN'T HAVE NEW instructions, and they are not usefull. They
are! With the new registers and new address space, they are amazing! But
not on Z180. The only I really miss are mult and div, FOR MSX, of course.
On Educar the things will be a lot different... (-;
>> Z180 is more "compatible" with R800 throught trapping. (-;
>That is not an advantage.
>The R800 instructions are hardly used, and it prevents the ability to
>'upgrade'.
That's point of view. (-;
>> Well... use what you want. ACE002 will be, at first, in Z180. (-;
>Certainly.
>But the 'new MSX' should feature a Z380, because it's way superior.
But it's a way not MSX... (-; And you will not convince me. Just
because a new processor has a lot of enhancements doesn't make it
a better solution for upgrade MSX. I mentioned x86 as a powerful
processor, better than Z380. You said "oh, but there is only 7
registers...!"... Well... then let's use PowerPC, on a new board.
PowerPC is a looooooooooong way better than Z380. And has lots of
registers also.
> (((-; I'm sure that for programs designes specially for Z380, Z380 will be
>*very* faster than Z180. On Z380 you can do LOTS of thing only using
>registers, which of course increase A LOT the speed the apps run. But we
>are thinking on "legacy" applications... We want to support than in
>the "fastest way". (-;
Of course...! This is why Ademir will use Z380 for EduCar and not Z180!
>'The fastest way' doesn't matter that much. It's 10.5x compared to 12x.
>And when using it to its full potential, according to Zilog:
>Relative Performance Z80: 2.5, Z180: 11, Z380: 33
I don't know how this comparison was made.
>By the way, the Z180 still has got an adress range of 64kB. However, they do
>some tricky stuff with some sort of mapping, which enables the user to use
>1MB. But there are no new instructions to cope with the extended adress
>range. And the Z180 has non 'extended' mode either, by the way. The 'legacy'
>mode is the only mode, with a few added instructions.
I never said there was an extended mode on Z180. I said THE OPOSITE.
And this is one of the reasons I like it to a "new" MSX. We just need
a MSX2+ that will be BETTER than TurboR. In every single aspect.
ACE002 doesn't need even be compatible with TR, I think.
>If it's already there, they must (otherwise they can't run new European
>software anymore).
I think they're not "that worried" about this. (-;
AbraçOS/2, Daniel Caetano ([EMAIL PROTECTED])
...!m.tag
OS/2 Sites: http://www.quasarbbs.com/daniel/
http://www.geocities.com/SiliconValley/8752/os2hp/os2index.html
MSX Sites: http://www.fudeba.cjb.net/
Drawings: http://www.djgallery.tsx.org/
****
MSX Mailinglist. To unsubscribe, send an email to [EMAIL PROTECTED]
and put "unsubscribe msx [EMAIL PROTECTED]" (without the quotes) in
the body (not the subject) of the message.
Problems? contact [EMAIL PROTECTED]
More information on MSX can be found in the following places:
The MSX faq: http://www.faq.msxnet.org/
The MSX newsgroup: comp.sys.msx
The MSX IRC channel: #MSX on Undernet
****
