Arcade / Console --> Amiga conversion discussions

[roondar]

Dragon Breed runs in 16 colours. Other than that it offers quite a bit over the ST version, such as smooth scrolling with parallax effects, better colours in some places and oh, twice the framerate.

I don't mind ST ports if they improve that much on the original

[touko]

Originally Posted by dlfrsilver

The CPU is slightly faster on the ST because the Atari Engineers had a nasty surprise on the prototype they were testing before the ST went mass builded in factory : The quartz they choose at first had the same speed as the one used on the Amiga. Unfortunately, the ST to the contrary of the Amiga had no hardware acceleration, and the prototype was too slow, too slow to the point that it could not be sold. They changed the quartz initially used for a 8mhz one, so that the ST speed could be sufficient to be used. Then they gave green light for mass production of the ST.

If the ST was too slow with a 7.16 mhz CPU, i don't think that a 8 mhz change so many things .
Arcade games need to have many and flexible hardware sprites with variable sizes, a bunch of palettes and a big amount of memory.

[roondar]

The ST crystal being swapped with a faster one for speed reasons does sound very unlikely - especially since the reason for the Amiga's 7.16/7.09MHz clock was the custom chip needing exact timing with NTSC/PAL.

Even more so since the standard 68000 ratings at the time where 8, 10 and 12MHz.

Seems to me that Atari (which probably didn't need the same exact timing due to lack of Copper etc) used the 8MHz clock because that was what the 68000 and standard RAM were designed for, whereas the Amiga clocked the CPU down a bit to accommodate the exact timings needed for PAL/NTSC based video effects.

[sandruzzo]

Commodore did lazy project, since you can have CPU That run at some MHZ and chip set to another

[roondar]

It's not all bad!

A synchronous design like this does allow the CPU and chipset to run without wait states (exluding DMA contention obviously). Having the CPU at a different speed would have only really made sense if both chipset and CPU had their own memory, which the original Amiga didn't have (otherwise you would get wait states).

[sandruzzo]

Amiga had some issue that i made me think that it wasn't finished project...

[frank_b]

Originally Posted by sandruzzo

Amiga had some issue that i made me think that it wasn't finished project...

The only major oddity was that it wasn't fully compatible with the 68k. TAS <cough>

[sandruzzo]

Originally Posted by frank_b

The only major oddity was that it wasn't fully compatible with the 68k. TAS <cough>

It lack of a little bit of fast mem, even 32k would be great. Not Ham6 in hires, and so forth

[frank_b]

Originally Posted by sandruzzo

It lack of a little bit of fast mem, even 32k would be great. Not Ham6 in hires, and so forth

Not enough chip bandwidth for six bit high res ham.

[touko]

The real mistake IMO (but it has a cost) it's the slowness of chip mem .
The 2 cycles memory access have a big impact on the amiga's performances .

[sandruzzo]

Originally Posted by frank_b

Not enough chip bandwidth for six bit high res ham.

I know. That's why they should have done do better. Amiga was a computer not a pure consolle.

In my humble opinion with some tweak, Amiga would be still here..

If we add 32k of fast and do "demo" with 30 to 50% less screen width, we can see what should had been Amiga power

[chb]

Originally Posted by sandruzzo

I know. That's why they should have done do better. Amiga was a computer not a pure consolle.

Well, if you have the DRAM technology of that time and a 16 bit wide bus, you do not get higher memory bandwidth without considerable effort, like multiple memory subsystems... And 32k of fastram would be rather worthless for general applications, much more suited for games and demos, so contradicting the "computer" argument. In the end it's 1983/84 technology, I do not know any affordable system from that time that managed to get a higher bandwidth.

[sandruzzo]

Originally Posted by chb

Well, if you have the DRAM technology of that time and a 16 bit wide bus, you do not get higher memory bandwidth without considerable effort, like multiple memory subsystems... And 32k of fastram would be rather worthless for general applications, much more suited for games and demos, so contradicting the "computer" argument. In the end it's 1983/84 technology, I do not know any affordable system from that time that managed to get a higher bandwidth.

The points is that even only 32k would do the difference even thought you're gonna use Amiga 500 as computer. Try it and tell if it's worthless...

Amiga laked of 50% of bandwidth, that's my opinion.

[chb]

Originally Posted by sandruzzo

The points is that even only 32k would do the difference even thought you're gonna use Amiga 500 as computer. Try it and tell if it's worthless...

How do you want to do that? How do you ensure that applications use that 32k for their most time critical routines? How do you handle multitasking? Swapping between those 32k and chipram at task switches? Nice concept for a single-tasking game machine, but IMHO worthless for a general purpose computer. And ram wasn't that expensive at the time anyhow to justify such a small amount of fast ram.

Originally Posted by sandruzzo

Amiga laked of 50% of bandwidth, that's my opinion.

Where do you want to get it from? Fast ram alone does not give you HAM6-hires...

[sandruzzo]

Originally Posted by chb

How do you want to do that? How do you ensure that applications use that 32k for their most time critical routines? How do you handle multitasking? Swapping between those 32k and chipram at task switches? Nice concept for a single-tasking game machine, but IMHO worthless for a general purpose computer. And ram wasn't that expensive at the time anyhow to justify such a small amount of fast ram.

Table accessing, for examples

Originally Posted by chb

Where do you want to get it from? Fast ram alone does not give you HAM6-hires...

I was talking about chip-ram

[chb]

Originally Posted by sandruzzo

Table accessing, for examples

Again: How do you handle this in a multitasking environment? Do you expect application programmers to tailor their code (that's maybe written in C or Pascal) to this 32k ram?

Originally Posted by sandruzzo

I was talking about chip-ram

That's clear - but how do you improve bandwidth? Wider bus (32 bit) in 1984?

[Galahad/FLT]

Originally Posted by zero

Galahad is right, the gulf between games designed for the Amiga and those that are ports is huge.

Having said that, not all ports are terrible. Midnight Resistance was a pretty good arcade port. The original arcade system has a lot of slow-down anyway, and they did a good job keeping the port running at a more consistent frame rate. All the levels and enemies are there, it plays really well...

The music is probably the weakest aspect. Great tunes, and the conversions aren't bad or anything, but not the best on the Amiga either. Since most of the music is in-game they were limited by the available channels and RAM though, rather than lack of skill.

Special FX were different. When they could, they did do improvements to the Amiga version.

Robocop 2 on Amiga is far superior to the ST in every way imaginable, even their earlier stuff like Untouchables has noticeable improvements over the ST version.

They were the exception though, the likes of Tiertex and Teque largely didnt bother, the latter did later on, but most of their early stuff was a better title tune and not much else.

[sandruzzo]

@Chb

Do a config on winuae with only 32k of fast as see for yourdself. What about faster ram to have more cycles?

[idrougge]

Those 32 k would be allocated by the first program loaded, rendering all other programs as slow as usual. It'd only be useful for games and demos.

[touko]

Where do you want to get it from? Fast ram alone does not give you HAM6-hires...

No, but 140ns DRAM yes,but more expansive .