Click Here
home features news forums classifieds faqs links search
6052 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
Home
Features
News
Forums
Classifieds
Links
Downloads
Extras
OS4 Zone
IRC Network
AmigaWorld Radio
Newsfeed
Top Members
Amiga Dealers
Information
About Us
FAQs
Advertise
Polls
Terms of Service
Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
15 crawler(s) on-line.
 30 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 zipper:  59 mins ago
 DiscreetFX:  1 hr 13 mins ago
 Kronos:  1 hr 23 mins ago
 clint:  1 hr 23 mins ago
 pixie:  1 hr 25 mins ago
 Karlos:  1 hr 32 mins ago
 Futaura:  2 hrs 3 mins ago
 ppcamiga1:  2 hrs 41 mins ago
 agami:  3 hrs 21 mins ago
 Maijestro:  3 hrs 25 mins ago

/  Forum Index
   /  Amiga General Chat
      /  How good or bad was the AGA chipset in 1992/1993.
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 Next Page )
Poll : How good or bad was the AGA chipset in 1992/1993.
10p Excellent (Best at 2D/3D, colors, and resolution, frame rate etc.)
5p Good / better than most computer.
0p Barely hanging in there.
-5p Below average / slow but usable
-10p useless / horrible
 
PosterThread
Hammer 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 3:20:04
#461 ]
Elite Member
Joined: 9-Mar-2003
Posts: 4656
From: Australia

@cdimauro

Quote:

Depends on what you need to achieve with that.

For 2D games I've already stated that the additional Fast RAM isn't needed. 128kB at most were ok, but nothing more.

Having 1MB+ Fast RAM would have mean that most of it should have been moved to Chip-Mem when needed (because then the Fast RAM is used to store a consistent part of the assets). And we know that Commodore engineers were so smart to cripple the Chip-Mem access, by forcing the CPU to wait an additional two cycles. So, moving data from Fast to Chip-Mem was already computationally expensive and thanks to them it's double that expensive!

Many new Amiga 2D games are requiring or recommended Fast RAM e.g. Turrcian AGA, Jump! AGA


From Turrican AGA's developer, Muzza


It requires Fast RAM to run at 50hz on an A1200. From testing I found it can run at 50hz on a base A1200 but only if I reduce the number of bitplanes. I may do this eventually, but for now I'd rather get it working without modifications to any of the source artwork. Currently it uses 8 bitplanes.



From Jump!'s developer, Santander Bank

Hardware requirements:

Amiga 1200 14MHz with at least 6MB of fast ram. This configuration ensures full gameplay smoothness.


These are the full 256-color (8-bit plane) games.

CPU's workload that ping-pongs with external memory shouldn't disturb Amiga chipset operations i.e. CPU also consumes memory access cycle slot.

Both 68000 and 68020 doesn't have data cache.

Unlike SNES Wing Commander end-user experience, Amiga 1200 wasn't configured for a smooth frame rate Wing Commander experience.

Last edited by Hammer on 28-Oct-2022 at 04:20 AM.
Last edited by Hammer on 28-Oct-2022 at 04:04 AM.
Last edited by Hammer on 28-Oct-2022 at 03:52 AM.
Last edited by Hammer on 28-Oct-2022 at 03:32 AM.
Last edited by Hammer on 28-Oct-2022 at 03:30 AM.
Last edited by Hammer on 28-Oct-2022 at 03:26 AM.
Last edited by Hammer on 28-Oct-2022 at 03:23 AM.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 5:51:08
#462 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@Hammer

Quote:

Hammer wrote:
@cdimauro

https://miharinne.artstation.com/projects/OoOxlv



Elfmania's unreleased art pack shows its object size budgeting.

It's a balancing act between the two player object sizes and gameworld effects.

Exactly. And as the artist said:

The "body parts" library we used to save time on making animations. But they were also recycled by code to save memory! The way characters were assembled and parts recycled on the fly to make the game fit in mere 512K was insane.

With Fightin' Spirit we split the graphic parts in more, smaller pieces. And we used 1MB of memory (512kB Chip + 512kB Slow/Fast RAM). Because we had many more colors and animations.
Quote:
Fightin' Spirit's Neo Geo influence followed a less successful franchise on the Neo Geo platform.

https://www.youtube.com/watch?v=2NvhhZPvaVQ
CD32's Fightin' Spirit AGA VS. NEO GEO's Fatal Fury. Posters preferred Shadow Fighter AGA to compete against Fatal Fury, not Fightin' Spirit AGA.

Again, two completely different graphic engines.

Shadow Fighter uses the Dual Playfield because they wanted to have two screens, one for the background and one for the characters, and they wanted to implement the parallax. This also allowed to have much more space for the animations (8 colors when talking about background + 7 colors for players for the OCS version; 16 + 15 for AGA). However the price to pay is to have a poor graphic palette.

Fightin' Spirits used 64 colors for the background and 31 colors for the players. On OCS (AGA had much more). That's why it's so colorful and doesn't look bad compared to the consoles. If you also consider that the Amiga has 1985 hardware and had only the sprites, whereas the Neo Geo was released 5 years after and had a spectacular chipset devoted only for games.

BTW the Na.P.S. guys are Sicilians like me and my team (their name is a Sicilian sentence, which I will not reveal here), leaving at around 100Km from our site. They also visited us when we both were developing our games and they shown us a WiP version of Shadow Fighter (OCS).
Quote:

Hammer wrote:
@cdimauro

Quote:

Depends on what you need to achieve with that.

For 2D games I've already stated that the additional Fast RAM isn't needed. 128kB at most were ok, but nothing more.

Having 1MB+ Fast RAM would have mean that most of it should have been moved to Chip-Mem when needed (because then the Fast RAM is used to store a consistent part of the assets). And we know that Commodore engineers were so smart to cripple the Chip-Mem access, by forcing the CPU to wait an additional two cycles. So, moving data from Fast to Chip-Mem was already computationally expensive and thanks to them it's double that expensive!

Many new Amiga 2D games are requiring or recommended Fast RAM e.g. Turrcian AGA, Jump! AGA


From Turrican AGA's developer, Muzza


It requires Fast RAM to run at 50hz on an A1200. From testing I found it can run at 50hz on a base A1200 but only if I reduce the number of bitplanes. I may do this eventually, but for now I'd rather get it working without modifications to any of the source artwork. Currently it uses 8 bitplanes.



From Jump!'s developer, Santander Bank

Hardware requirements:

Amiga 1200 14MHz with at least 6MB of fast ram. This configuration ensures full gameplay smoothness.


These are the full 256-color (8-bit plane) games.

CPU's workload that ping-pongs with external memory shouldn't disturb Amiga chipset operations i.e. CPU also consumes memory access cycle slot.

Both 68000 and 68020 doesn't have data cache.

Unlike SNES Wing Commander end-user experience, Amiga 1200 wasn't configured for a smooth frame rate Wing Commander experience.

Again, you didn't got the point. The discussion here was about what was desirable to have on Amiga for developing 2D games. And I've expressed my opinion due to my knowledge and experience on the field.

So, just to highlight it again: it's about hypothetical Amiga with a better chipset from this PoV.

The games that you talk about are realized for the current machines, so the real hardware and you can do nothing else other than adding memory and/or faster processors. And as you can see yourself they required A LOT of additional memory to overcame the limits of the original machines.

But not only that: Turrican AGA developer's accelerator card has a 28Mhz 68020 with 4MB of Fast Mem. So, it's not only a matter of having more memory.

Whereas Jump! is a simple 3D game where the graphic isn't that complicated. But it requires 6MB of Fast RAM! Graphic games didn't require so much memory because they should... render the scene on the fly, and the scene is very simple here, as you can see.
It means that the 3D graphic is pre-rendered, so the game is simulating the 3D effect by drawing 2D graphic. It's obvious why so much Fast RAM is used.

Hammer, you have no experience with game coding and you don't see those details.

 Status: Offline
Profile     Report this post  
Massi 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 6:08:08
#463 ]
Cult Member
Joined: 2-Feb-2011
Posts: 627
From: Rome, Italy

@cdimauro

A game tailored for A1200 should be written in Assembly 68020 rather than 68000 , to get the most out of the machine. Otherwise it is just yet another port from OCS / ECS with better graphics.

Was it the case for Fighting Spirit for A1200 ?

_________________
SAM440EP-FLEX @ 733 Mhz, AmigaOS 4.1 Update 1

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 8:05:26
#464 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@Massi

Quote:

Massi wrote:
@cdimauro

A game tailored for A1200 should be written in Assembly 68020 rather than 68000 , to get the most out of the machine.

This isn't necessarily true. The game should work well on that machine, and this is and should be the only requisite.
Quote:
Otherwise it is just yet another port from OCS / ECS with better graphics.

This is a completely different thing. And it's not a logical implication of the previous sentence.
Quote:
Was it the case for Fighting Spirit for A1200 ?

I haven't worked at it, but I don't think that using the 68020 was necessary for this specific games and according to how we've written the OCS version.

There's nothing that I recall about the 68020 which could have really helped pushing more the AGA version.

 Status: Offline
Profile     Report this post  
Karlos 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 10:35:22
#465 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3565
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hammer

Quote:
From Jump!'s developer, Santander Bank





Wow, they really do have a diverse portfolio.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
MEGA_RJ_MICAL 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 10:39:42
#466 ]
Super Member
Joined: 13-Dec-2019
Posts: 1200
From: AMIGAWORLD.NET WAS ORIGINALLY FOUNDED BY DAVID DOYLE

@Karlos

After all, it anagrams to

SATAN NERD

Nothing could better describe a company that is involved in both gaming and banking.


/M!

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

 Status: Offline
Profile     Report this post  
Karlos 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 10:59:10
#467 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3565
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@MEGA_RJ_MICAL

(((SANTANDER))) == (((NERD SATAN)))

o_O


Actually maybe just Nerd Santa, bringing new games for a 30 year old machine.

Last edited by Karlos on 28-Oct-2022 at 11:02 AM.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Bosanac 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 11:03:40
#468 ]
Regular Member
Joined: 10-May-2022
Posts: 242
From: Unknown

@Karlos

Quote:
@MEGA_RJ_MICAL

(((SANTANDER))) == (((NERD SATAN)))

o_O


Did (((they))) make you write that?

 Status: Offline
Profile     Report this post  
Karlos 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 12:08:22
#469 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3565
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Bosanac

I felt a strange compulsion, starting close to where I had my last COVID vaccination.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
bhabbott 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 14:44:41
#470 ]
Regular Member
Joined: 6-Jun-2018
Posts: 251
From: Aotearoa

Quote:

cdimauro wrote:

Having 1MB+ Fast RAM would have mean that most of it should have been moved to Chip-Mem when needed (because then the Fast RAM is used to store a consistent part of the assets). And we know that Commodore engineers were so smart to cripple the Chip-Mem access, by forcing the CPU to wait an additional two cycles. So, moving data from Fast to Chip-Mem was already computationally expensive and thanks to them it's double that expensive!

So, before adding Fast RAM this ridiculous handicap should have been removed first.

I agree with you on the Chip vs Fast RAM allocation. Better to have more Chip RAM where the custom chips access it directly and cost is reduced. Users who want higher CPU performance can simply add a RAM board or accelerator card.

However your suggestion that Commodore's engineers were idiots for keeping the 2 cycle CPU interface is off-base. It was a fundamental part of how the system worked, so bypassing it would require big changes to the design that could compromise compatibility and break other stuff. It would be a real headache to get right and delay the introduction of AGA even longer. It's easy to look at some design limitation and imagine waving a magic wand to make it disappear, but actually doing it is a different story.

But as you say, having data in Fast RAM means it has to be copied to Chip RAM before use - which is something you don't want to do anyway. So the only issue is that the 2 cycle Chip RAM interface slows the CPU down, but that also leaves more slots free for the essential display DMA and desirable blitter operation. Ultimately it's a balance between CPU speed and custom chip performance. Which would you rather have?

The Amiga's philosophy was to offload graphics functions to the custom chips, freeing up the CPU for other stuff. This make sense in a multitasking environment, and also when the CPU is less efficient than custom hardware. Having the CPU do it all was the bane of 8 bit systems, and hardware graphics functions were cherished. However as CPUs became more powerful there was a shift towards software rendering which had opportunities for optimization by programmers, particularly on the PC which had relatively 'dumb' graphics hardware. But then 3D happened and once again it became obvious that using the CPU to do it all would hamper performance.

All modern systems get their insane graphics performance by offloading rendering to GPUs that work independently of the CPU, just like the Amiga did. In modern systems the CPU doesn't touch the frame buffer directly at all, because that would be a waste of its capabilities. And everybody knows the answer to slow graphics performance is not a faster CPU, but faster graphics hardware.

Having the same chipset between Amiga models resulted in consistent performance that made it more inclusive. Sure it meant that faster CPUs were 'handicapped', but it also forced developers to make their products work on all models - not just those with the fastest CPUs. For those of us who don't have infinite funds or don't want to 'upgrade' our system just to play a new game, this is a good thing. It's good for developers and vendors too, as they have a wider market and fewer configurations to support.

Developers may wish for a magical machine that does whatever they can think of as fast as theoretically possible (or even impossible), but the constraints of real hardware actually make their job easier. Compare the effort required to develop a typical 8 bit game vs the latest 3D production. In the retro scene it also increases the satisfaction of squeezing more out of limited hardware, rather than 'cheating' by just specifying a faster CPU. This is particularly important now that CPUs are being developed with ridiculously high performance. We don't want a situation where 'Amiga' games require a Vampire or PiStorm because developers are too lazy to work within the constraints of retro hardware.

Quote:
However and as I game developer what I really missed from the chipset, and this since the beginning, was the possibility to flip the sprite and Blitter graphic. AT LEAST horizontally (which is the most important).

Why? Flipping is a cheap and lazy way to reverse the orientation of a character. It's useful when you have limited memory, but doesn't give as realistic results as using separate graphics for each direction.

TV tropes: Ambidextrous Sprite
Quote:
On earlier platforms, there were also memory size concerns, so it was often more efficient to mirror the sprite than to store the opposite poses, especially with graphics hardware that made horizontal flipping as simple as changing one bit of sprite data.

However, sometimes characters or objects don't lend themselves well to symmetry. A character might have an object in one hand, an eye patch or scar on one side, or some other form of Fashionable Asymmetry. An object might have text written on it, or asymmetrical pieces such as a missing headlight on a car. Whatever it is, simply flipping the sprite causes the feature to "change sides". In the most blatant examples, the sprite will have letters or numbers on it which flip with the sprite.



Quote:
IMO this is one, if not The, biggest mistake made by Jay Miner (besides using planar graphics, of course. BTW, tomorrow my article is finally published).

I disagree. It's not a 'mistake' at all. The same effect can be achieved without any extra display hardware by simply having more shapes in memory. It's a lot easier to increase RAM than custom chip hardware. Agnus started with 512K, then 1MB, then 2MB, without much effort - even before AGA was developed. The next generation was intended to have up to 8MB. With that much Chip RAM you don't need to use crude tricks.

Quote:
Instead of using that big part of the chip for implementing the HAM mode, the graphic flipping should have been the best usage possible for that silicon.

Perhaps, or perhaps HAM mode gave the Amiga a 'wow' factor that set it apart from other machines which were solely designed for 2D games. In 1985 nobody had anything that could touch it for 'photo-realistic' images.

There are a ton of other things that could have been added that IMO were potentially more important. A 'character' or 'tiled' background would have improved performance in text-based applications and games designed to use the technique. Packed pixels would have made porting and emulation easier, and prepared for future expansion to 'chunky' modes. These are features that many other home computers had which were dropped in favor of a much simpler '1 bit per pixel' bitplane system.

When Jay miner designed the Amiga no other home computer had sprite or tile flipping, and I'm not aware of any system that had blitter flipping. It's only real use is for fighting games, which have large objects that 'need' to be mirrored quickly (they really shouldn't, but that's another story). You may be disappointed that he didn't include this feature just for you, but that doesn't make it a 'mistake'.

Last edited by bhabbott on 28-Oct-2022 at 02:46 PM.

 Status: Offline
Profile     Report this post  
BigD 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 15:14:34
#471 ]
Elite Member
Joined: 11-Aug-2005
Posts: 6915
From: UK

@Hammer

Quote:
Unlike SNES Wing Commander end-user experience, Amiga 1200 wasn't configured for a smooth frame rate Wing Commander experience.


True, it seems to have still been focused on platformers IMHO! Fast Ram out of the box was the only way to make up for the poor kludge of cheap 020 chip and lack of many upgraded custom chips other than Alice and Lisa brought to you from HP rather than CSG/MOS!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

 Status: Offline
Profile     Report this post  
Hammer 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 16:29:20
#472 ]
Elite Member
Joined: 9-Mar-2003
Posts: 4656
From: Australia

@BigD

Quote:

BigD wrote:
@Hammer

Quote:
Unlike SNES Wing Commander end-user experience, Amiga 1200 wasn't configured for a smooth frame rate Wing Commander experience.


True, it seems to have still been focused on platformers IMHO! Fast Ram out of the box was the only way to make up for the poor kludge of cheap 020 chip and lack of many upgraded custom chips other than Alice and Lisa brought to you from HP rather than CSG/MOS!

68EC020/68020 has a barrel shifter @ 14 Mhz.

For modern CPUs on desktop computers, there is very little need for DSP when modern CPUs have SIMD extensions.

Due to Alice's 68000 memory access pattern, it takes 68EC020 two clock cycles to access Chip RAM.

CSG/MOS fabs weren't up to date, hence the contracting HP fabs.

Dec 1990 era C65's 256 color display chipset was limited to 4096 color palette.

In 1992, HP can fab PA-RISC PA-7100 CPUs starting from 33 Mhz to 100 Mhz.





_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

 Status: Offline
Profile     Report this post  
Hammer 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 16:34:23
#473 ]
Elite Member
Joined: 9-Mar-2003
Posts: 4656
From: Australia

@Karlos

I referring to the developer, was it Kefir_Union? https://www.indieretronews.com/2022/05/jump-amiga-aga-version-of-yoomp-gets.html

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

 Status: Offline
Profile     Report this post  
Hammer 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 17:01:13
#474 ]
Elite Member
Joined: 9-Mar-2003
Posts: 4656
From: Australia

@cdimauro

Quote:

Again, two completely different graphic engines.

Shadow Fighter uses the Dual Playfield because they wanted to have two screens, one for the background and one for the characters, and they wanted to implement the parallax. This also allowed to have much more space for the animations (8 colors when talking about background + 7 colors for players for the OCS version; 16 + 15 for AGA). However the price to pay is to have a poor graphic palette.

Fightin' Spirits used 64 colors for the background and 31 colors for the players. On OCS (AGA had much more). That's why it's so colorful and doesn't look bad compared to the consoles. If you also consider that the Amiga has 1985 hardware and had only the sprites, whereas the Neo Geo was released 5 years after and had a spectacular chipset devoted only for games.

BTW the Na.P.S. guys are Sicilians like me and my team (their name is a Sicilian sentence, which I will not reveal here), leaving at around 100Km from our site. They also visited us when we both were developing our games and they shown us a WiP version of Shadow Fighter (OCS

My point is the preference for game world effects that adds dynamism instead of a static and dead game world.

I'm aware of Shadow Fighter OCS/ECS was using dual play fIelds hardware feature which is also a 6-bit plane mode.

Shadow Fighter's developers prioritize game world dynamism.

Each time some Amiga fanboys show Fight'n Sprit example to the SNES owners and it was mocked for the dead game world.


Quote:

Again, you didn't got the point. The discussion here was about what was desirable to have on Amiga for developing 2D games. And I've expressed my opinion due to my knowledge and experience on the field.

FYI, Wing Commander is 2.5D which is not truly 3D like Quake or Wing Commander 3.

You missed NutsAboutAmiga's "10p Excellent (Best at 2D/3D , colors, and resolution, frame rate etc.)" context.



Quote:

So, just to highlight it again: it's about hypothetical Amiga with a better chipset from this PoV.

The games that you talk about are realized for the current machines, so the real hardware and you can do nothing else other than adding memory and/or faster processors. And as you can see yourself they required A LOT of additional memory to overcame the limits of the original machines.


Psygnosis, Gremlin Interactive, and Ex-Commodore UK MD wanted accelerated A1200 bundle deals and I support these viewpoints.


Quote:

But not only that: Turrican AGA developer's accelerator card has a 28Mhz 68020 with 4MB of Fast Mem. So, it's not only a matter of having more memory.

FALSE.

Turrican AGA only needs base A1200 with Fast Memory.

https://eab.abime.net/showthread.php?t=106735

From modrobert, Just tested, runs great on my stock A1200 with 4mb fast RAM!

I'm not going to waste my time on dismantling my A1200 with TF1260 and swapping it out for AmigaKit's A1200 8 MB Fast RAM memory expansion.


Quote:

Whereas Jump! is a simple 3D game where the graphic isn't that complicated. But it requires 6MB of Fast RAM! Graphic games didn't require so much memory because they should... render the scene on the fly, and the scene is very simple here, as you can see.
It means that the 3D graphic is pre-rendered, so the game is simulating the 3D effect by drawing 2D graphic. It's obvious why so much Fast RAM is used.

Hammer, you have no experience with game coding and you don't see those details.

Don't assume I didn't know about Shadow Fighter's dual-play fIelds and that's your mistake.

I did state about the balance between game world effects and player objects, hence I'm aware of Amiga's hardware constraints. I prefer ElfMania's game world effects and player object balance.

Jump! follows Stardust's tunnel scenes and Reshoot Proxima 3 pre-rendered video methods that fall into Amiga's strength in video playback, hence Amiga's chipset is ideal for Japanese laser-disc quick-time event (QTE) games.

I like Reshoot Proxima 3's game world dynamism with multiple objects balance.

Don't defend Amiga's inferior hardware progress. Your statement "Hammer, you have no experience with game coding" is shit.



Last edited by Hammer on 28-Oct-2022 at 05:33 PM.
Last edited by Hammer on 28-Oct-2022 at 05:30 PM.
Last edited by Hammer on 28-Oct-2022 at 05:25 PM.
Last edited by Hammer on 28-Oct-2022 at 05:21 PM.
Last edited by Hammer on 28-Oct-2022 at 05:20 PM.
Last edited by Hammer on 28-Oct-2022 at 05:10 PM.

_________________
Ryzen 9 7900X, DDR5-6000 32 GB RAM, GeForce RTX 4080
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 21:05:40
#475 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@bhabbott

Quote:

bhabbott wrote:
Quote:

cdimauro wrote:

Having 1MB+ Fast RAM would have mean that most of it should have been moved to Chip-Mem when needed (because then the Fast RAM is used to store a consistent part of the assets). And we know that Commodore engineers were so smart to cripple the Chip-Mem access, by forcing the CPU to wait an additional two cycles. So, moving data from Fast to Chip-Mem was already computationally expensive and thanks to them it's double that expensive!

So, before adding Fast RAM this ridiculous handicap should have been removed first.

I agree with you on the Chip vs Fast RAM allocation. Better to have more Chip RAM where the custom chips access it directly and cost is reduced. Users who want higher CPU performance can simply add a RAM board or accelerator card.

Exactly.
Quote:
However your suggestion that Commodore's engineers were idiots for keeping the 2 cycle CPU interface is off-base.

No. I've said that they weren't smart.
Quote:
It was a fundamental part of how the system worked, so bypassing it would require big changes to the design that could compromise compatibility and break other stuff. It would be a real headache to get right and delay the introduction of AGA even longer. It's easy to look at some design limitation and imagine waving a magic wand to make it disappear, but actually doing it is a different story.

Bypassing it is much simpler.

Actually the chipset has to keep track of odd and even color clocks = DMA slots. When the CC is odd it completely blocks the CPU access. When the CC is even it checks if there's no other peripheral requiring access to the memory and checks if the CPU requested it: if both conditions are true, then it grants access to the CPU.

This worked fine with the 68000, because this (usually) processor accesses memory exactly following this pattern (first two cycles are kept for internal processing. On other twos it accesses memory; if required).

But this mechanism could have simply be removed with 68020+. So, basically only leaving the logic for the even CCs. This would have simplified the logic, requiring less transistors to be used for its implementation.

As you can see, it's exactly the opposite of what you said. BTW, this should have already be fixed on the ECS (with the Amiga 3000): AGA was already too late.
Quote:
But as you say, having data in Fast RAM means it has to be copied to Chip RAM before use - which is something you don't want to do anyway.

Exactly. No sane coder would have liked to copy some data (BOBs, sprites, audio samples) from Fast RAM to Chip-Mem (so, wasting both space in Chip-Mem and its bandwidth) and THEN asking the chipset to use it.

Just give more Chip-Mem and let the chipset immediately use it!
Quote:
So the only issue is that the 2 cycle Chip RAM interface slows the CPU down, but that also leaves more slots free for the essential display DMA and desirable blitter operation. Ultimately it's a balance between CPU speed and custom chip performance. Which would you rather have?

As I've said before, currently it requires to split the Chip-Mem accesses between the CPU and the Blitter to use as much as possible of them.

As you could imagine, this is a nightmare for the developers. And it's still inefficient, because you cannot use all available memory slots at their full potential.
Quote:
The Amiga's philosophy was to offload graphics functions to the custom chips, freeing up the CPU for other stuff. This make sense in a multitasking environment, and also when the CPU is less efficient than custom hardware. Having the CPU do it all was the bane of 8 bit systems, and hardware graphics functions were cherished.

For 2D games the CPU has a very limited role, whatever is the system: 8-bit, 16-bit, 32-bit.
Quote:
Having the same chipset between Amiga models resulted in consistent performance that made it more inclusive. Sure it meant that faster CPUs were 'handicapped', but it also forced developers to make their products work on all models - not just those with the fastest CPUs. For those of us who don't have infinite funds or don't want to 'upgrade' our system just to play a new game, this is a good thing. It's good for developers and vendors too, as they have a wider market and fewer configurations to support.

This is not incompatible with having the chipset better supporting 68020+ CPUs.
Quote:
Developers may wish for a magical machine that does whatever they can think of as fast as theoretically possible (or even impossible), but the constraints of real hardware actually make their job easier. Compare the effort required to develop a typical 8 bit game vs the latest 3D production. In the retro scene it also increases the satisfaction of squeezing more out of limited hardware, rather than 'cheating' by just specifying a faster CPU. This is particularly important now that CPUs are being developed with ridiculously high performance. We don't want a situation where 'Amiga' games require a Vampire or PiStorm because developers are too lazy to work within the constraints of retro hardware.

I don't see any problem on that, with the current discussion.
Quote:
Quote:
However and as I game developer what I really missed from the chipset, and this since the beginning, was the possibility to flip the sprite and Blitter graphic. AT LEAST horizontally (which is the most important).

Why? Flipping is a cheap and lazy way to reverse the orientation of a character.

Hardware flipping is the SMARTEST way to do this operation. Lazy can only applied to whom has NOT implemented it in hardware.
Quote:
It's useful when you have limited memory,

Which is exactly the case!
Quote:
but doesn't give as realistic results as using separate graphics for each direction.
Quote:
TV tropes: Ambidextrous Sprite
[quote] On earlier platforms, there were also memory size concerns, so it was often more efficient to mirror the sprite than to store the opposite poses, especially with graphics hardware that made horizontal flipping as simple as changing one bit of sprite data.

However, sometimes characters or objects don't lend themselves well to symmetry. A character might have an object in one hand, an eye patch or scar on one side, or some other form of Fashionable Asymmetry. An object might have text written on it, or asymmetrical pieces such as a missing headlight on a car. Whatever it is, simply flipping the sprite causes the feature to "change sides". In the most blatant examples, the sprite will have letters or numbers on it which flip with the sprite.

And what's the problem? On THIS specific case proper graphic could be provided.

Bu you cannot take SOME EXCEPTIONS and pretending to shut down the GENERAL case.
Quote:
Quote:
IMO this is one, if not The, biggest mistake made by Jay Miner (besides using planar graphics, of course. BTW, tomorrow my article is finally published).

I disagree. It's not a 'mistake' at all. The same effect can be achieved without any extra display hardware by simply having more shapes in memory.

But this requires more memory, which was very expensive at the time.
Quote:
It's a lot easier to increase RAM than custom chip hardware.

It's exactly the opposite: RAM is much more expensive, whereas hardware flipping required a bunch of transistors.
Quote:
Agnus started with 512K, then 1MB, then 2MB, without much effort - even before AGA was developed. The next generation was intended to have up to 8MB. With that much Chip RAM you don't need to use crude tricks.

And you should know how expensive was memory at the time.
Quote:
Quote:
Instead of using that big part of the chip for implementing the HAM mode, the graphic flipping should have been the best usage possible for that silicon.

Perhaps, or perhaps HAM mode gave the Amiga a 'wow' factor that set it apart from other machines which were solely designed for 2D games. In 1985 nobody had anything that could touch it for 'photo-realistic' images.

How many times people enjoyed such images and how many times they enjoyed videogames? Think about it and draw your conclusions...
Quote:
There are a ton of other things that could have been added that IMO were potentially more important. A 'character' or 'tiled' background would have improved performance in text-based applications and games designed to use the technique.

This would have required much more transistors, since it's a completely different hardware implementation.
Quote:
Packed pixels would have made porting and emulation easier, and prepared for future expansion to 'chunky' modes. These are features that many other home computers had which were dropped in favor of a much simpler '1 bit per pixel' bitplane system.

Actually packed/chunky pixels would have required less resources to be implemented.
Quote:
When Jay miner designed the Amiga no other home computer had sprite or tile flipping, and I'm not aware of any system that had blitter flipping.

Why you have to wait for others to implement it?
Quote:
It's only real use is for fighting games, which have large objects that 'need' to be mirrored quickly (they really shouldn't, but that's another story). You may be disappointed that he didn't include this feature just for you, but that doesn't make it a 'mistake'.

Hardware flipping can be used on very different games: car games, platform games, even adventures!

You haven't developed games, so you don't know the challenges (and space constraints!). Hence, you don't know what's better needed for developers.

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 28-Oct-2022 21:28:07
#476 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@Hammer

Quote:

Hammer wrote:
@cdimauro

Quote:

Again, two completely different graphic engines.

Shadow Fighter uses the Dual Playfield because they wanted to have two screens, one for the background and one for the characters, and they wanted to implement the parallax. This also allowed to have much more space for the animations (8 colors when talking about background + 7 colors for players for the OCS version; 16 + 15 for AGA). However the price to pay is to have a poor graphic palette.

Fightin' Spirits used 64 colors for the background and 31 colors for the players. On OCS (AGA had much more). That's why it's so colorful and doesn't look bad compared to the consoles. If you also consider that the Amiga has 1985 hardware and had only the sprites, whereas the Neo Geo was released 5 years after and had a spectacular chipset devoted only for games.

BTW the Na.P.S. guys are Sicilians like me and my team (their name is a Sicilian sentence, which I will not reveal here), leaving at around 100Km from our site. They also visited us when we both were developing our games and they shown us a WiP version of Shadow Fighter (OCS

My point is the preference for game world effects that adds dynamism instead of a static and dead game world.

This contradicts your previous statement about your preference for Elfmania, which has even much less dynamism...
Quote:
I'm aware of Shadow Fighter OCS/ECS was using dual play fIelds hardware feature which is also a 6-bit plane mode.

Shadow Fighter's developers prioritize game world dynamism.

And Fightin' Spirit is a balance between it and more colored graphics.
Quote:
Each time some Amiga fanboys show Fight'n Sprit example to the SNES owners and it was mocked for the dead game world.

ROFL Do you understand the differences between the Amiga and SNES hardware?!?

Besides that, why don't you show other beat'em ups and ask the SNES owners a judgement?
Quote:
Quote:

Again, you didn't got the point. The discussion here was about what was desirable to have on Amiga for developing 2D games. And I've expressed my opinion due to my knowledge and experience on the field.

FYI, Wing Commander is 2.5D which is not truly 3D like Quake or Wing Commander 3.

You missed NutsAboutAmiga's "10p Excellent (Best at 2D/3D , colors, and resolution, frame rate etc.)" context.

And I've said nothing about that, right?

Whereas the discussion on this part was talking about something different. Hence, my statements.
Quote:
Quote:

So, just to highlight it again: it's about hypothetical Amiga with a better chipset from this PoV.

The games that you talk about are realized for the current machines, so the real hardware and you can do nothing else other than adding memory and/or faster processors. And as you can see yourself they required A LOT of additional memory to overcame the limits of the original machines.


Psygnosis, Gremlin Interactive, and Ex-Commodore UK MD wanted accelerated A1200 bundle deals and I support these viewpoints.

Which is a completely different thing...
Quote:
Quote:

But not only that: Turrican AGA developer's accelerator card has a 28Mhz 68020 with 4MB of Fast Mem. So, it's not only a matter of having more memory.

FALSE.

Turrican AGA only needs base A1200 with Fast Memory.

https://eab.abime.net/showthread.php?t=106735

From modrobert, Just tested, runs great on my stock A1200 with 4mb fast RAM!

OK. Then the game should make some consistent differences when running only with Chip-Mem and when using the additional Fast-Mem.

Because and as I've already discussed, using the Fast-Mem required to copy the assets from there to the Chip-Mem, before that the chipset could have used them. Which means a performance drop.
Quote:
I'm not going to waste my time on dismantling my A1200 with TF1260 and swapping it out for AmigaKit's A1200 8 MB Fast RAM memory expansion.

Not needed. It's more important to understand why those differences between Chip-Mem-only and with Fast-Mem.
Quote:
Quote:

Whereas Jump! is a simple 3D game where the graphic isn't that complicated. But it requires 6MB of Fast RAM! Graphic games didn't require so much memory because they should... render the scene on the fly, and the scene is very simple here, as you can see.
It means that the 3D graphic is pre-rendered, so the game is simulating the 3D effect by drawing 2D graphic. It's obvious why so much Fast RAM is used.

Hammer, you have no experience with game coding and you don't see those details.

Don't assume I didn't know about Shadow Fighter's dual-play fIelds and that's your mistake.

Do you understand that here you talked about Jump! and I've replied on that? See the above quote!

Why you talk about Shadow Fighter here?

And where I've said that you don't know how it works? I NEVER stated it! You're inventing things now!
Quote:
I did state about the balance between game world effects and player objects, hence I'm aware of Amiga's hardware constraints. I prefer ElfMania's game world effects and player object balance.

Which contradicts your statement at the top of your post...
Quote:
Jump! follows Stardust's tunnel scenes and Reshoot Proxima 3 pre-rendered video methods that fall into Amiga's strength in video playback,

Which require so much additional Fast-Mem...
Quote:
hence Amiga's chipset is ideal for Japanese laser-disc quick-time event (QTE) games.

It was already used for laser games.
Quote:
I like Reshoot Proxima 3's game world dynamism with multiple objects balance.

Your taste. I don't care.
Quote:
Don't defend Amiga's inferior hardware progress.

Care to SHOW me WHERE I've stated it?!? You're lying! Again!
Quote:
Your statement "Hammer, you have no experience with game coding" is shit.

And here you go personal: clear signal that you aren't able to sustain the discussion. Again.

Not a news with you: it's your typical pattern...

 Status: Offline
Profile     Report this post  
Massi 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 29-Oct-2022 5:10:51
#477 ]
Cult Member
Joined: 2-Feb-2011
Posts: 627
From: Rome, Italy

@cdimauro

Come on, on A1200 you write Assembly 68000 instead of 68020 ?

No surprise Fighting Spirit seems to run in "slow motion".

And then in this thread you say that AGA was not good for its time ... what was not good is the use of the A1200 for Fighting Spirit.



_________________
SAM440EP-FLEX @ 733 Mhz, AmigaOS 4.1 Update 1

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 29-Oct-2022 5:24:51
#478 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@Massi

Quote:

Massi wrote:
@cdimauro

Come on, on A1200 you write Assembly 68000 instead of 68020 ?

No, I've said something different. Don't put on my mouth words that I've NOT written, please.
Quote:
No surprise Fighting Spirit seems to run in "slow motion".

This has nothing to do with the above: the two things are logically different. So, this is a logical fallacy.

As well as you cannot draw any conclusion because you don't know how Fightin' Spirit was written.
Quote:
And then in this thread you say that AGA was not good for its time ...

Which, again, is something completely different.

Another logical fallacy.

Double logical fallacy, considered that first you talk about the processor (68000 vs 68020) and now you talk about the chipset (AGA) and you mixed everything together.
Quote:
what was not good is the use of the A1200 for Fighting Spirit.

As I've said, you don't know how the game was written.

So, again another logical fallacy.

You started accumulating logical fallacies, Massi, and also inventing things that people haven't written: clear signals that you aren't able to sustain the discussion anymore.

 Status: Offline
Profile     Report this post  
Massi 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 29-Oct-2022 6:00:02
#479 ]
Cult Member
Joined: 2-Feb-2011
Posts: 627
From: Rome, Italy

@cdimauro

That Figthing Spirit for A1200 runs in "slow motion" is a fact.

The A1200 is a beast in the hands of the good developers, certainly not in the hands of opinionists that can only bla bla bla ...

_________________
SAM440EP-FLEX @ 733 Mhz, AmigaOS 4.1 Update 1

 Status: Offline
Profile     Report this post  
cdimauro 
Re: How good or bad was the AGA chipset in 1992/1993.
Posted on 29-Oct-2022 6:07:30
#480 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3104
From: Germany

@Massi

Quote:

Massi wrote:
@cdimauro

That Figthing Spirit for A1200 runs in "slow motion" is a fact.

Then you can prove this fact and show it.
Quote:
The A1200 is a beast

This is the only thing which I agree on.
Quote:
in the hands of the good developers, certainly not in the hands of opinionists that can only bla bla bla ...

Ah, ok. Now you, that never created a single game for Amiga, started your personal attacks.

In fact, actually it is exactly the opposite: YOU are the opinionists (not even able to sustain your "arguments") whereas I'm the one which DID something for the Amiga.

Spot the difference, incompetent.

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle