Laurens Holst wrote:
>If I remember correctly SLL is handled differently on a R800 than on a Z80.
Yes, that was exactly my point. ;)
>I just don't know the exact difference between them. To me, the Z380 seems
>the best candidate, because AFAIK (in capitals) the Z380 is only a little
>bit slower in Z80-emulation mode than the Z180 is. But if the difference is
>very large... Ah, whatever I already said this!
The difference isn't large IMO:
Z80 @ 3.57Mhz, 4 T-states/opcode = 1x speed
Z180 @ 33 Mhz, 3 T-states/opcode = ~9.2x and ~1.3x = ~12x speed
Z380 @ 14 Mhz, 2 T-states/opcode = 4x and 2x = 8x speed
Z380 is 800% the speed of Z80, while Z180 is only 50% faster than the Z380.
This is ofcourse not a real-world calculation. In a real-world the Z380
would perform closer to the Z180 because its a generally faster processor
and the Z180 is only faster because of the 2.5x greater clockspeed.
Z380 doesn't seem to be faster than R800 @ 7 MHz, but in reality will
probably run about twice as fast as turboR, since I've learned (from Alex
Wulm's article) R800 is often slowed down by memory.
>But afaik LPE's design hardly increases the speed of existing software.
See above, it does.
> > You're overreacting a bit aren't you?
>
>Nope.
I wasn't talking to you! ;)
>I do not dislike the Z380, but I do dislike the things like putting your
>processor in an expansion slot and not having direct access to old hardware.
So you'd rather want a complete new motherboard that will have to be
inserted in a PC-case?!
>PC is an expandable system, with all its drivers etc. and having every
>component pluggable.
>MSX is hardly.
BY DESIGN MSX is just as expandable, perhaps even more. It's also
standardized, so if you like having to go through SEVERAL software layers,
wasting CPU power along the way, before you can access the VDP and even
then not use its full potential, go right ahead and call PC expandable.
>I think the engine should automatically slowdown too fast access to I/O
>ports. The way LPE's design does it is, if not 'bad', then at least 'not
>nice'.
Problem with slowing down is that you get a slower system ;)
The way the Z80/R800 becomes a slave processor for accessing slow hardware
is perfect in my opinion. It does not slow down the main program at all.
>I partly too. I have no objection against the Z380 unless there is a really
>large difference in speed.
See above, if you think the an extra 50% is gonna make 'a really large
difference' over the already 800% increase, vote for Z180. But Z380 is
still president in my book :)
> > The power of the Z380, especially with the new instructions, is great
> > enough to run C programs very fast (SP-relative addressing!). That means
> > many more programmers will be tempted to program for MSX, since they do
>not
> > have to bother with assembly.
>
>I think not.
Why not? I could name several people that stopped developing MSX programs,
or didn't START developing just because they didn't want to bother to use
or learn Z80 assembly.
>I am not delighted by those new instructions because they cause
>incompatibility with not Z380-expanded MSX computers. I want to run my
>programs to run better, faster, smoother on a Z380. But I don't want my
>programs not to work at all on normal MSX systems...
If Konami used that same idea, MSX2 would never have been popular! How many
games don't do something extra on MSX2+ or turboR (or what about Nemesis3
that uses the MSX2 palette for nicer colors)? You can program stuff and do
something extra on Z380. Or write seperate versions (Dragonslayer 4, King's
Valley 2)...
How many games don't have MSX-MUSIC/AUDIO music and ALSO MoonSound?
Heh, why don't new games have PSG support anymore? Now I can't enjoy music
on a standard MSX2! BWAH!
>I will not program Z380 specific. Most people won't I think. As said before,
>it's a bit like the Gfx9000 (only 'worse'). Only a few people have it, and
>if you write Gfx9000 specific soft, everybody without it can't run it. And
>why the 'worse'? Well, the Gfx9000 is a real great expansion which adds alot
>of extra features.
The same counts for FM-PAC, MSX2, hell even DISKDRIVES! It's the same as
what I said above... If everyone sticks with the old, there will never be
any new!
>The Z380-specific instructions are nice, but you can also
>use Z80 code which might be a bit slower, but works on all MSX computers.
>Surely, they make programming easier. But they can be worked around. Which
>is also a reason why some people program Z80, because of the limits which
>oppose a challenge.
Z380 still has enough limits, believe me.. If you are afraid of becoming
lazy, just stay away from x86, that's all. I mean:
MOV [8000h+6*EAX+EBX],12345678h
>This does not mean I plead against the Z380. Only against Z380-specific
>programming, and why we should rather look at speedup with the old
>instruction set instead of a higher speed with a new instruction set. And
>then always programs can be made which execute Z380 (or Z180)-specific code
>when it is detected.
Exactly, just like MMX or 3D-Now/SSE on PC... But I don't understand why
you would be against Z380 specific programming?!
Some things require high processing power, for instance MP3 replayers :)
I bet a Z380 could do it! But not in Z80-legacy code!
>But no really, it's useful. Then a Z180 / Z380 / eZ80-specific version of
>GEM could be made.
>Instead of only having benefit of the increased speed, it could also gain
>extra speed by using the specific instructions and extended adress range.
Z180 nah... Only Z380, and to a less extent eZ80, allows great speedups
when run in 'full' mode. The Z180 just doesn't have a lot of nice speedy
opcodes!
In other words there's only a minimal difference between programming Z180
in legacy code or in full Z180 assembly.
Patriek
****
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
****
