Click Here
home features news forums classifieds faqs links search
6071 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
0 crawler(s) on-line.
 104 guest(s) on-line.
 1 member(s) on-line.


 Dragster

You are an anonymous user.
Register Now!
 Dragster:  2 mins ago
 VooDoo:  21 mins ago
 mailman:  38 mins ago
 BillE:  42 mins ago
 matthey:  50 mins ago
 Hypex:  55 mins ago
 pixie:  1 hr 24 mins ago
 saimo:  1 hr 47 mins ago
 DiscreetFX:  1 hr 50 mins ago
 Lou:  2 hrs ago

/  Forum Index
   /  Amiga General Chat
      /  Could lack of parallel bitplane writes crippled the Amiga?
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 Next Page )
PosterThread
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 4:52:24
#221 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@kamelito

Quote:

kamelito wrote:
@Hammer

SNES resolution is 256x224 compared to 320x256 so if Amiga developpers reduce the resolution to the SNES level they could put more stuff on screen.

Super NES's internal resolution is between 256x224 (8:7 ratio, NTSC) or 256 × 240 (PAL) and 512x448.

The resolution difference between 320x200 and 256x224 is about 11 percent.

In most cases, Nintendo made sure its arcade-ported games for SNES were of good quality.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 5:46:20
#222 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Hammer:
Quote:
Quote: 64-bit Blitter on 32-bit bus would need two memory access cycles.


Gunnar:
Quote:
What exactly is the problem?
What do you think how AGA works?
The memory chips that Commodore uses in A1200 support burst mode.
ALICA the AMIGA AGA DMA controller can do a 64bit bursts access consisting of two reads each 32bit in one memory cycle.
This is how AGA works.


Hammer:
Quote:
Your "64-bit" is meaningless without context. 64-bit payload over several clock cycles is different from 64-bit per clock cycle.
Your "64 bit" marketing doesn't work on me.



Fact: AGA chipset runs 100% exactly the same clockrate as OCS and ECS.
Fact: AGA Alice can do in one Amiga chipset clock a 64bit memory read.



Hammer, it saddens me to see that in all the 22 years that Amiga AGA chipset is out on the market now -
you post in rand here about Commodore and all this with absolutely ZERO understanding and knowledge how Amiga chipsets work.


 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 5:58:11
#223 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:
Unlike the Lisa chip, Alice's 16-bit Blitter design is unable to exploit 4X memory bandwidth improvements on stock A1200.


Alice delivers the 64bit memory to Lisa.
Alice can do a 64bit transfer in one clock cycle.

The AGA Blitter was not improved.
Simple because the deadline did not allow it.

Gunnar:
Quote:






(OCS AGNUS) (AGA ALICE) (Super-AGA/ANNI)
GFX display 16 bit 64 bit 256 bit
Blitter DMA 16 bit 16 bit 256 bit
Sprites DMA 16 bit 64 bit 256 bit


* AGA did significantly improve Display and Sprites x4 times
* It is correct that AGA did not improve Blitter.
It goes without saying that if AGA would have improved Blitter too then this would have made AGA stronger.



We all agree AGA would have been stronger if the Blitter would have been 64bit too.
But we all know that the team did not had the development time for this in the AGA release.
We all know this and we understand the reason.

Saying that the Blitter is 16bit in AGA is correct.
Everybody knows this.



What exactly is your point?

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:10:29
#224 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:

Gunnar wrote:

Fact: AGA chipset runs 100% exactly the same clockrate as OCS and ECS.

I already stated this information.

Quote:

Fact: AGA Alice can do in one Amiga chipset clock a 64bit memory read.


Fact: Stock A1200's Chip RAM has a 32-bit data memory bus. Your "64-bit" marketing is flawed.

Fact: AGA Blitter is still 16-bit. Prove AGA Blitter has 28 MB/s usage.

Fact: Index Information Ltd.'s Access incorporates an improved Chip RAM interface which yields twice as fast memory read performance as of a standard A1200.

Access incorporates the core Amiga chips only: Alice, Paula, Denise, and 8520 CIAs. Other chips such as Gayle and Budgie are replaced by then new chip design, offering increased CPU access speed to Chip RAM.

Quote:

Hammer, it saddens me to see that in all the 22 years that Amiga AGA chipset is out on the market now -
you post in rand here about Commodore and all this with absolutely ZERO understanding and knowledge how Amiga chipsets work.

Bullshit.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:16:17
#225 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

GUNNAR:
Quote:
Why do you always need to blame someone?


HAMMER:
Quote:

Refer to Dave Haynie's argument position. David Pleasance (Commodore UK MD) has other information on Commodore's Hong Kong to the Philippines factory transfer debacle, Commodore International CEO(1993)'s rejecting Commodore UK/Bull Frog (associated with EA)/Team 17/Psygnosis's CPU accelerated A1200 bundles.



I have the feeling you not understand,
how project management works and what people job were in Commodore.

David Pleasance: = was UK sales manager.
David is a nice guy and his job was to discuss bulk bargains discounts with retailers.
He did in his job good. His job was NOT to decide anything about Amiga development.


Dave Haynie
Dave worked as engineer then later lead engineer in several projects.
Dave did excellent development and engineering work.
His job was to develop new Amiga models and chipsets.

In his positions as engineer Dave of wanted to make the best possible - no matter what the costs, no matter
how long it takes. This perfectly matches his role as good engineer!


The opposite work position has a project manager.
A project Manager get from his boss a list of "minimal requirements" the project needs to full fill.
He gets a certain budget and a timeframe in which this project needs to be finished.
The job of the project manager is to keep the project finished in time, and in the budget.

You complain a lot that no new chips were done for the Amiga 3000.

But what does this means besides that the Amiga 3000 project manager did a good job?
His job was to guard thas the project is not getting over the timeline
or over the budget. And for an Office Amiga having more gaming power was certainly not on the requirement list.


Do you understand this?

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:26:58
#226 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:

Alice delivers the 64bit memory to Lisa.
Alice can do a 64bit transfer in one clock cycle.

Notice Alice to Lisa instead of factoring the entire graphics pipeline from 32-bit Chip RAM.

Fact: AGA has a 32-bit memory bus. Prove AGA has a 64-bit memory bus.

Prove Alice Blitter can use 28 MB/s memory bandwidth.

Quote:

The AGA Blitter was not improved.
Simple because the deadline did not allow it.

Blame Commodore management for the R&D schedule.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:38:04
#227 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:

Prove Alice Blitter can use 28 MB/s memory bandwidth.



Hammer what is wrong with you? Why should I do this?

I posted all times that AGA blitter is 16bit.

Did you not read what I wrote?
Let me repeat one more time for you:


Quote:

The Amiga has very many DMA channels.
Some are more hungry than others in terms of bandwidth.

The top 3 hungry users are:
- GFX display DMA
- Blitter DMA
- Sprites DMA

AGA did improve some of them but not all
Lets look the improvements:






(OCS AGNUS) (AGA ALICE) (Super-AGA/ANNI)
GFX display 16 bit 64 bit 256 bit
Blitter DMA 16 bit 16 bit 256 bit
Sprites DMA 16 bit 64 bit 256 bit


* AGA did significantly improve Display and Sprites x4 times
* It is correct that AGA did not improve Blitter.
It goes without saying that if AGA would have improved Blitter too then this would have made AGA stronger.

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:39:31
#228 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:
Blame Commodore management for the R&D schedule.


I don't want to blame anyone. Blaming people is childish.

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 6:52:55
#229 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

GUNNAR:
Quote:
Hammer, it saddens me to see that in all the 22 years that Amiga AGA chipset is out on the market now - you post in rand here about Commodore and all this with absolutely ZERO understanding and knowledge how Amiga chipsets work.


Hammer:
Quote:
Bullshit.



yes you posted a lot bullshit.


You looked at the CPU memory benchmark BUSTEST.
Based on the CPU benchmark you argued about AMIGA DMA bandwidth.
= Was this bullshit?


Everyone saw how many of your posts were bullshit
- you not understandd how AGA DMA works,
- you not understand how DRAM memory works
- and because you not knew that a CPU in the 80th needed several clocks for a memory access.

Not knowing or not understanding technical details is in itself not a problem.

While tried to explainl you how these things work.
What I not get is your motivation to continue to constantly argue without knowledge?

What do you want to proof, except your lack of Amiga knowledge?

Last edited by Gunnar on 11-Mar-2024 at 07:01 AM.

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 7:11:26
#230 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:

I have the feeling you not understand,
how project management works and what people job were in Commodore.

Your defense of Commodore management's debacle is something.

Quote:

David Pleasance: = was UK sales manager.
David is a nice guy and his job was to discuss bulk bargains discounts with retailers.
He did in his job good. His job was NOT to decide anything about Amiga development.

David Pleasance has established relations with 3rd party game companies and timed exclusive Amiga game bundles e.g. Batman pack. Reminder, exclusive game software sells the hardware.

Commodore UK can reduce the risk for the small 3rd party CPU accelerator vendors with secured CPU-accelerated A1200 bundles.

In 1993, David Pleasance led three major 3rd party game developers in a meeting with Commodore International's CEO. Commodore International's CEO told the game developers to fukoff in the meeting.

After the disastrous meeting,
Electronics Arts/Bullfrog's 3D game tiles focused on platforms like 3DO and PC and canceled Amiga's Magic Carpet development. Faster Amigas was ignored.

For "32-bit" 3D game titles, Psygnosis decided to go full dive with Sony's PlayStation. Faster Amigas was ignored.

Team 17 still released Alien Breed 3D in a comical A6 letter-size render window.

This is on top of John Carmark's damaging PR remarks against the Amiga.

Commodore International's CEO's AAA 3rd party game developer relations are trash which is opposite from Sony PlayStation's and Microsoft DirectX (precursor to Xbox) leadership teams.

You underestimated 3rd party game developer relations.

Quote:

Dave Haynie
Dave worked as engineer then later lead engineer in several projects.
Dave did excellent development and engineering work.
His job was to develop new Amiga models and chipsets.

In his positions as engineer Dave of wanted to make the best possible - no matter what the costs, no matter
how long it takes. This perfectly matches his role as good engineer!

The opposite work position has a project manager.
A project Manager get from his boss a list of "minimal requirements" the project needs to full fill.
He gets a certain budget and a timeframe in which this project needs to be finished.
The job of the project manager is to keep the project finished in time, and in the budget.

You complain a lot that no new chips were done for the Amiga 3000.

But what does this means besides that the Amiga 3000 project manager did a good job?
His job was to guard thas the project is not getting over the timeline
or over the budget. And for an Office Amiga having more gaming power was certainly not on the requirement list.

You didn't factor in the time wasted by Commodore's management and the loss of key original Amiga ASIC engineers exiting from Commodore. Refer to 3DO arrived in late 1993.







_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 7:36:25
#231 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:
You looked at the CPU memory benchmark BUSTEST.
Based on the CPU benchmark you argued about AMIGA DMA bandwidth.
= Was this bullshit?

You didn't factor in the gotchas with AGA's higher fetch rates.

1. Your "64-bit" marketing is a red herring since they are not general-purpose enough.

2. 1985 era low color display 64-bit wide sprites is NOT a "32-bit 2.5D/3D" next-gen gaming experience.

I prefer two 16-bit Blitters over 1985 era low color display 64-bit wide sprites.

3. 1985 era very low color display from four 64-bit wide sprites will not preserve 256-color artwork from VGA platforms without compromise.

4. Your Lisa to Alice link argument is useless when the primary composition engine besides the CPU is the AGA's 16-bit Blitter e.g. draw, fill, move, copy, and 'etc'.

I gave you an example of a 256-color artwork VGA ported game maxing the stock A1200's bus bandwidth.

What's the point of a raster with 4X memory bandwidth improvements with a slow 1985 Blitter? Hint: 640x400p 256 color productivity mode is slow!

Quote:

Everyone saw how many of your posts were bullshit
- you not understand how AGA DMA works,
- you not understand how DRAM memory works
- and because you not knew that a CPU in the 80th needed several clocks for a memory access.
[quote]
I already told you AGA and ECS/OCS timings weren't changed.

My argument is a full graphics pipeline from memory to the endpoint. You focused on just the Alice to Lisa. Lisa's raster doesn't complete the graphics solution!

I'm already aware of AGA's higher fetch modes with 4X memory bandwidth improvement, but they did NOT address the 16-bit Blitter!


[quote]
Not knowing or not understanding technical details is in itself not a problem.

While tried to explainl you how these things work.
What I not get is your motivation to continue to constantly argue without knowledge?

What do you want to proof, except your lack of Amiga knowledge?

Bullshit.

Last edited by Hammer on 11-Mar-2024 at 07:45 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
bhabbott 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 12:14:40
#232 ]
Regular Member
Joined: 6-Jun-2018
Posts: 349
From: Aotearoa

@Hammer

Quote:

Hammer wrote:

In 1993, David Pleasance led three major 3rd party game developers in a meeting with Commodore International's CEO. Commodore International's CEO told the game developers to fukoff in the meeting.

Can you provide a source for this story?


Quote:
This is on top of John Carmark's damaging PR remarks against the Amiga.

This one too.

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 12:41:12
#233 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:
David Pleasance has established relations with 3rd party game companies and timed exclusive Amiga game bundles e.g. Batman pack. Reminder, exclusive game software sells the hardware.


I know David Pleasance very well. We know each other since many years.
I also know all the war-time stories - that you refered.



I think understand know what you point is.
Its 100% emotional for you.

You want to hate someone for Amiga not ruling the world.
If this makes you happy, then go ahead.

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 13:20:14
#234 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@bhabbott

Quote:

bhabbott wrote:

Can you provide a source for this story?


Commodore, The Inside Story, the untold tale of a computer-giant by David-John Pleasance
Page 129 of 414,


Ian explained to Mehdi that with a few seemingly quite
modest design changes, the CD32 could have an
incredible boost in performance at very marginal
additional cost.
He also pointed out the benefits it would
give developers like Psygnosis and other major players in
the industry, who would find it easier to produce even
better-quality products and enhance the reputation of the
CD32 and the games publishers – a genuine ‘win-win’. Ian
had not requested any financial reward for this – it
seemed he simply wanted to offer considerably improved
games performance and to be credited for his
contribution.

Well, it went exactly as expected. Mehdi was rude and
ignorant, and clearly had no idea what Ian was talking
about.
But instead of just admitting that, he more or less
turned on Ian, as though he ‘must be crazy telling us how
to design our computers!’ I ushered Mehdi out of the
building feeling very ashamed, and it was quite a while
before I plucked up the courage to talk to Ian again.
Luckily for me, Ian had realised what kind of a person
Mehdi Ali was and held no bad feelings towards me.


The real sting to this story is that Psygnosis
subsequently sold their company to Sony Computer
Entertainment Europe, with Ian Hetherington being
made head of Sony PlayStation Europe – and I often
wonder to myself: ‘If Mehdi Ali had not been such an
obnoxious prick, would Commodore have had that
technology?’


Commodore's upper management are ignorant fools.

During other shows, David Pleasance repeated the above story with minor variations.

Facts: 3DO has more than 2 million units which is more than AGA's under a million install base.

Quote:

This one too.

For this topic.

From jo...@idcube.idsoftware.com Sun Sep 4 02:52 EST 1994
From: John Carmack
Date: Sat, 3 Sep 94 11:50:23 -0600
To: G.San...@ais.gu.edu.au
Subject: amiga doom

The amiga is not powerfull enough to run DOOM. It takes the full
speed of a 68040 to play the game properly even if you have a chunky
pixel mode in hardware. Having to convert to bit planes would kill
it even on the fastest amiga hardware, not to mention the effect it
would have on the majority of the amiga base.

John Carmack

Reference
https://groups.google.com/g/comp.sys.amiga.games/c/MZb9cC0FMhw

Amiga 4000/040 @ 25 Mhz (on 030 bus) Doom performance is not 486SX-33. WarpEngine 040 needs to replace the A3640 card. WarpEngine 040 turns the same 68040 @ 25Mhz into 28 Mhz slight overclock and with proper local 32-bit Fast RAM.

Amiga's 3rd party 68040 accelerator cards are usually superior when compared to Commodore's A3640 card.

During 1993, the problem was price vs performance when compared to the PC clone's 486SX/DX and Apple 68LC040/68040 competition.

PC's multiple 2.5D/3D AAA games sold the gaming PC platform.

Last edited by Hammer on 11-Mar-2024 at 02:03 PM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 13:47:19
#235 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:

Gunnar wrote:
@Hammer

Hammer what is wrong with you? Why should I do this?

I posted all times that AGA blitter is 16bit.

Did you not read what I wrote?
Let me repeat one more time for you:


Quote:

The Amiga has very many DMA channels.
Some are more hungry than others in terms of bandwidth.

The top 3 hungry users are:
- GFX display DMA
- Blitter DMA
- Sprites DMA

AGA did improve some of them but not all
Lets look the improvements:






(OCS AGNUS) (AGA ALICE) (Super-AGA/ANNI)
GFX display 16 bit 64 bit 256 bit
Blitter DMA 16 bit 16 bit 256 bit
Sprites DMA 16 bit 64 bit 256 bit


* AGA did significantly improve Display and Sprites x4 times
* It is correct that AGA did not improve Blitter.
It goes without saying that if AGA would have improved Blitter too then this would have made AGA stronger.


FYI, Lisa's data pins are numbered up to 32 i.e. D0 to D31. Where are the 64-bit data pins?

LISA's 14 Mhz clock is sync to ALICE's 28 Mhz clock generator.

LISA's "64-bit" refers to eight 64-bit Bitplane Registers and 2 planes wide.

Classic Pentium has a 64-bit front-side data bus. My "64-bit" POV is data I/O lines NOT SRAM registers! Your "64-bit" argument is a red herring.

Each of LISA sprites is 2 planes i.e. 4 colors. Combined four sprites have 16 colors.

ALICE is modified for 2 cycle Fast Page DRAM. There's a peak of 28 MB/s bandwidth with bitplanes fetch. ALICE Blitter is 16-bit wide, hence wasting 32-bit data bus improvements.

68EC020 CPU has its share of this bandwidth and it's bottlenecked by mentioned I/O control chips. 68EC020 CPU's hardware barrel shifter is wasted on this I/O bottleneck.

Dave Haynie wanted AT&T DSP3210 (13.3 MIPS, 25 MFLOPS FP32 @ 50Mhz) to be the graphics compute power.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 13:58:55
#236 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:
ALICE is modified for 2 cycle Fast Page DRAM. There's a peak of 28 MB/s bandwidth with bitplanes fetch. ALICE Blitter is 16-bit wide, hence wasting 32-bit data bus improvements.




Maybe it would help you to look at the goals for the AGA amiga chipset.


1) Being able to play OCS/A500 game.
his is a very important goal.

To reach this goal it was very clever to keep all the DMA cycle timing 100% the time.
Do you agree?


2) To improve system performance the engineers added this clever 64bit fetch feature.
This allowed to give the system 400% bandwidth while keeping the DMA slot design the same.

Was this clever?
I think yes

3) all the dream features were planned for 'AAA' chipset.
AGA on purpose limited the features to what they had working and debugged quickly.


Everyone has different wishes and ideas.
Personally I see a lot of sense in AGA.
I would rather have a debugged and working chipset with some nice improvements,
than a not debugged and not working chipset.

But everyone is free to have your own opinion.

 Status: Offline
Profile     Report this post  
bhabbott 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 14:15:34
#237 ]
Regular Member
Joined: 6-Jun-2018
Posts: 349
From: Aotearoa

@Hammer

Quote:

Hammer wrote:

From jo...@idcube.idsoftware.com Sun Sep 4 02:52 EST 1994
From: John Carmack
Date: Sat, 3 Sep 94 11:50:23 -0600
To: G.San...@ais.gu.edu.au
Subject: amiga doom

The amiga is not powerfull enough to run DOOM. It takes the full
speed of a 68040 to play the game properly even if you have a chunky
pixel mode in hardware. Having to convert to bit planes would kill
it even on the fastest amiga hardware, not to mention the effect it
would have on the majority of the amiga base.

This was a newsgroup post in response to an Amiga fan pointing out that there were Amigas powerful enough to run Doom. Carmack's response indicates that he wasn't interested in porting Doom to such a small market (the 'majority of the amiga base' being A500s which obviously couldn't run it).

We know this because he did allow porting it to other lesser platforms such as the SNES, which needed an SFX Chip in the cartridge to get acceptable performance. It also had no floor or ceiling textures and the enemies always faced forwards. Despite these simplifications and the 'Super' FX chip it still only achieved a frame rate of around 10fps. But this was OK because there was a huge captive market for Doom on the SNES (can't pirate that cartridge, have to buy one!). So it was really all about potential sales, not whether or not the hardware was fast enough.

BTW note the date on his post - 3 Sep 94 - well after Commodore was no more. Bit late for a 'damaging PR remark' against the Amiga, you think?

 Status: Offline
Profile     Report this post  
Hammer 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 14:44:45
#238 ]
Elite Member
Joined: 9-Mar-2003
Posts: 5394
From: Australia

@Gunnar

Quote:

Gunnar wrote:
@Hammer

Maybe it would help you to look at the goals for the AGA amiga chipset.

1) Being able to play OCS/A500 game.
his is a very important goal.

To reach this goal it was very clever to keep all the DMA cycle timing 100% the time.
Do you agree?

2) To improve system performance the engineers added this clever 64bit fetch feature.
This allowed to give the system 400% bandwidth while keeping the DMA slot design the same.

Was this clever?
I think yes

Your SAGA has backward compatibility.

Quote:

3) all the dream features were planned for 'AAA' chipset.
AGA on purpose limited the features to what they had working and debugged quickly.


Everyone has different wishes and ideas.
Personally I see a lot of sense in AGA.
I would rather have a debugged and working chipset with some nice improvements,
than a not debugged and not working chipset.

But everyone is free to have your own opinion.

Lisa's fix function hardware can exploit AGA's 4X memory bandwidth improvement with gotchas limitations while ALICE's 16-bit object manipulator hardware (Blitter) can't exploit 4X bandwidth improvements.

A dual 16-bit Blitter would have maintained backward compatibility when legacy Amiga software would use one of the 16-bit Blitters.

When scrolling a 640x200x2 screen, AAA Blitter is 6X faster than AGA Blitter.

In the 1987 year range, the Amiga Ranger chipset was designed around faster VRAM, and this R&D was aborted, and lost key original Amiga engineers.

Commodore's B-team and management can't design "3DO" in a timely manner.

Note why I compared Commodore's AAA/Hombre to 3DO's, NVIDIA's and 3DFX's R&D phase to product shipping time interval. Employees proven skills matter and company-supported skill development is important. SGI has produced excellent GPU engineers who founded 3DFX and ArtX.

What's required from 1992 to 1994 is a fast VGA with 486SX-33 compute performance levels.


Quote:

AGA on purpose limited the features to what they had working and debugged quickly.

Commodore management wasted time and resources, hence the bandaid solution known as AGA along with the intended AT&T DSP3210.

Amiga's technological lead was wasted.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A ECS, KS 3.2, PiStorm/RPi 3A+/Emu68)

 Status: Offline
Profile     Report this post  
Gunnar 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 14:54:19
#239 ]
Cult Member
Joined: 25-Sep-2022
Posts: 511
From: Unknown

@Hammer

Quote:
Your SAGA has backward compatibility.


Yes this is correct. And after successful doing this I really know how Amiga chipset works internally.

The developing of SAGA was very difficult because: SAGA keeps backward compatibility - while at the same time it uses modern memory with different timing attributes.

I was only able to do this development because I had the advantage
that I use an FPGA for development and I could "retry" my work as often as I wanted to get it perfect.


This means I was able to "retry" like in Groundhog day.
Mind that the Commodore team did NOT had this possible at their time.
They needed to get the chipset done without retrying.
This made their work a lot more difficult.

I have the utmost respect for their work.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Could lack of parallel bitplane writes crippled the Amiga?
Posted on 11-Mar-2024 16:17:47
#240 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4415
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Gunnar

Quote:
I have the utmost respect for their work.


+1

You know, we're all (well most of us) armchair engineers beating down on Commodore's mistakes. Personally I don't get the angst over AGA. It was meant to be a temporary solution and it was meant to be affordable. My A1200 was a big jump over my A600 back in 1993. I have absolutely no regrets. My A1200 is the computer I have had the most fun with.

My only criticism of the A1200 was the that it didn't have some fast ram out of the box. There are various ways this could have been done and all of them would have increased the cost, but even the basic 020 with fast ram was just so much more capable the stock configuration.

_________________
Doing stupid things for fun...

 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 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