Unreleased Amiga chipset better than anything today?

[hammer]

Originally Posted by idrougge

The C128 was released in 1985. It was a technological dead end that probably shouldn't have been released, but released it was.

It did in several ways. It could run Wing Commander in 256 colours. It could run Wolfenstein 3D. It still had an operating system from the 70s, but that disk operating system could use HD floppies and hard drives were very cheap.

It depends on the VGA solution and PC model.

My Dad had an IBM PS/2 Model 55SX's slow IBM VGA in the early 1990s before it was traded for a PC clone with 1992 era 386DX-33 with an ET4000AX SVGA card. Our 1992-era 386DX-33 PC clone has an onboard SRAM cache and full 32-bit system RAM.

IBM PS/2 Model 55SX was the lesson on IBM tax e.g. its MCA add-on cards weren't price competitive compared to PC clone counterparts.

With action games, the Amiga 500 can blow away IBM VGA, but fast VGA clones have different results.

[hammer]

Originally Posted by pandy71

Don't get me wrong but everything listed by you is not Amiga - emulators based on UAE and WinUAE are just Amiga emulators.

AmigaForever and A500Mini are licensed products from the current Amiga IP holder i.e. Amiga Corporation. They are realistic enough to recognize hardware 68K lost the CPU war.

FreeSync/GSync covers 50Hz content and Drawbridge USB covers Amiga floppy drive support including physical floppy sound e.g. Battle Squadron, Hybris and 'etc'.

Originally Posted by pandy71

AROS is not Amiga is a Amiga OS recreation independently from HW.

No shit Sherlock.

Originally Posted by pandy71

Hombre was not planned with x86 as CPU.

Modern X86 CPUs don't directly implement X86 CISC microarchitecture.

One of the major requirements for Commodore's Hombre is big endian support.

Originally Posted by pandy71

And you can achieve this in many ways - but this is not CF issue.

Real-life implementation is an issue since they followed the "Apple way."

Emu68's supervisor 68K emulation can be implemented on other big-endian CPUs like PowerPC, not just ARM. The major problem is the hardware cost.

Fully superscalar ColdFire V5 design didn't exist during Commodore's Hombre design phase. Read https://en.wikipedia.org/wiki/NXP_ColdFire


There are five generations or versions of the ColdFire available from Freescale:

v1: Intended to support migration from 8-bit microcontrollers, it is a cut-down version of the v2 processor-wise. It was launched in 2006, 12 years after the original ColdFire. It is designed to easily replace the 8-bit Freescale 68HC08 microcontrollers and compete with low-end ARM chips.
v2: The original ColdFire core launched in 1994. Single-issue pipeline, no MMU, no FPU. Versions are also available with MAC and enhanced MAC units.
v3: Added an optional MAC unit.
v4: Limited superscalar core.
v4e (or eV4 in some documents): Enhanced version of the v4, launched in 2000. Adds optional MMU, FPU, and enhanced MAC unit to the architecture.
v5: Fully superscalar core.

The ColdFire family is uncompetitive trash for desktop markets.

Back on the topic: Commodore engineers designed Amiga's Ramsey memory controller with a 32-bit bus, up to 25 Mhz, and 68030 bus protocols, but Commodore management had a "read my lips, no new chips" directive during A3000's R&D.

The Ramsey memory controller and Super DMac (DMA address generation) designs could have been the foundation for the new 25 Mhz 32bit Agnus/Alice for the Amiga 3000. A display raster engine on Amiga 3000's 32-bit Fast RAM @ 25 Mhz would be pretty good in 1990 and can rival 25Mhz 32bit VLB SVGA cards. The next evolution after a 25 Mhz 32-bit controller can be a PA RISC era ~50Mhz 64-bit bus design. The foundation for 3DFX's Voodoo 1's performance is memory bandwidth. In the late 1990s, NVIDIA rapidly increased the raw memory bandwidth with the aid of Samsung.

1996-era Voodoo 1 has three 64-bit buses at 50 Mhz. NVIDIA focused on a 128-bit bus with high clock speed for RiVA 128 and RiVA TNT series.

[pandy71]

Originally Posted by hammer

AmigaForever and A500Mini are licensed products from the current Amiga IP holder i.e. Amiga Corporation. They are realistic enough to recognize hardware 68K lost the CPU war.

FreeSync/GSync covers 50Hz content and Drawbridge USB covers Amiga floppy drive support including physical floppy sound e.g. Battle Squadron, Hybris and 'etc'.

And?

Originally Posted by hammer

Modern X86 CPUs don't directly implement X86 CISC microarchitecture.

No shit Sherlock.

Originally Posted by hammer

One of the major requirements for Commodore's Hombre is big endian support.

And?

Originally Posted by hammer

Real-life implementation is an issue since they followed the "Apple way."

Emu68's supervisor 68K emulation can be implemented on other big-endian CPUs like PowerPC, not just ARM. The major problem is the hardware cost.

Fully superscalar ColdFire V5 design didn't exist during Commodore's Hombre design phase. Read https://en.wikipedia.org/wiki/NXP_ColdFire

And?

Originally Posted by hammer

Back on the topic: Commodore engineers designed Amiga's Ramsey memory controller with a 32-bit bus, up to 25 Mhz, and 68030 bus protocols, but Commodore management had a "read my lips, no new chips" directive during A3000's R&D.

The Ramsey memory controller and Super DMac (DMA address generation) designs could have been the foundation for the new 25 Mhz 32bit Agnus/Alice for the Amiga 3000. A display raster engine on Amiga 3000's 32-bit Fast RAM @ 25 Mhz would be pretty good in 1990 and can rival 25Mhz 32bit VLB SVGA cards. The next evolution after a 25 Mhz 32-bit controller can be a PA RISC era ~50Mhz 64-bit bus design. The foundation for 3DFX's Voodoo 1's performance is memory bandwidth. In the late 1990s, NVIDIA rapidly increased the raw memory bandwidth with the aid of Samsung.

1996-era Voodoo 1 has three 64-bit buses at 50 Mhz. NVIDIA focused on a 128-bit bus with high clock speed for RiVA 128 and RiVA TNT series.

And?

[hammer]

Originally Posted by pandy71

And?

1. Hardware 680x0 lost the war.

2. No modern X86 CPUs directly implement the CISC microarchitecture.

Originally Posted by pandy71

No shit Sherlock.

It seems emulation is a taboo for you.

Originally Posted by pandy71

And?

1. Commodore's one of few CPU selection criteria for Amiga Hombre.

2. Read https://amitopia.com/the-amithlon-x86-amiga/
The Amithlon Debacle. After Bill McEwen's Amiga Inc promoted "AmigaOS X86", billyboy sent a "cease and desist" letter against Haage & Partner's AmigaOS XL and Amithlon.

Bill McEwen's Amiga Inc's "cease and desist" killed AmigaOS 3.9 being bundled for X86-based Amithlon. FlipFlop Bill McEwen killed "AmigaOS X86" stone dead.

Originally Posted by pandy71

And?

1. Any ColdFire argument for this topic is useless when ColdFire's development and product release timeline is factored in. Any ColdFire argument is useless until fully superscalar Cold Fire V.

2. Apple's CPU migration example is useless for Amiga legacy games. FireTOS's Apple CPU migration approach is rubbish for the Amiga.

Originally Posted by pandy71

And?

This topic is a "What IF" when Commodore management wasn't shit e.g. remove "read my lips, no new chips" directive during Amiga 3000's development.

[pandy71]

Originally Posted by hammer

1. Hardware 680x0 lost the war.

Oh, i had impression it is widely used architecture in US nuclear and conventional missile guidance systems and most of them was never employed in target (global war) conditions but ok, i f you say so then i must be true...

Originally Posted by hammer

2. No modern X86 CPUs directly implement the CISC microarchitecture.


It seems emulation is a taboo for you.

Well... at first i'm not aware about implied "emulation taboo", secondly seem facts settled in chronology are taboo for you... 10 years seem to be unimportant from your perspective also bending facts seem to be fine from your perspective...

Originally Posted by hammer

1. Commodore's one of few CPU selection criteria for Amiga Hombre.

Please provide source about Commodore x86 choice as next CPU ISA for Hombre chipset.

Originally Posted by hammer

2. Read https://amitopia.com/the-amithlon-x86-amiga/
The Amithlon Debacle. After Bill McEwen's Amiga Inc promoted "AmigaOS X86", billyboy sent a "cease and desist" letter against Haage & Partner's AmigaOS XL and Amithlon.

Bill McEwen's Amiga Inc's "cease and desist" killed AmigaOS 3.9 being bundled for X86-based Amithlon. FlipFlop Bill McEwen killed "AmigaOS X86" stone dead.

My point about AROS is valid for any similar projects.

Originally Posted by hammer

1. Any ColdFire argument for this topic is useless when ColdFire's development and product release timeline is factored in. Any ColdFire argument is useless until fully superscalar Cold Fire V.

Brave statement but as i've wrote previously - please argue with Motorola/NXP on this, not with me - CF is seen as direct 68k line successor compatible at binary level with legacy 68k code (but with suboptimal performance). If you don't like this please redirect your comments to Motorola/NXP.

Originally Posted by hammer

2. Apple's CPU migration example is useless for Amiga legacy games. FireTOS's Apple CPU migration approach is rubbish for the Amiga.


This topic is a "What IF" when Commodore management wasn't shit e.g. remove "read my lips, no new chips" directive during Amiga 3000's development.

Legacy software, mostly games are serious issue for any modern platform - they are written usually in a way requiring special approach (perhaps something like DOSBox on PC).

I fully agree with you, AGA should be released way earlier, probably way before 1990, AGA should be also way better especially that J. Miner finished RANGER before his departure from Commodore.

[hammer]

Originally Posted by pandy71

Oh, i had impression it is widely used architecture in US nuclear and conventional missile guidance systems and most of them was never employed in target (global war) conditions but ok, i f you say so then i must be true...

Refer to CPU wars.

Originally Posted by pandy71

Well... at first i'm not aware about implied "emulation taboo", secondly seem facts settled in chronology are taboo for you... 10 years seem to be unimportant from your perspective also bending facts seem to be fine from your perspective...

Cold Fire is useless for this topic. For Hombre, Commodore's legacy support is A1200 on a chip.

Originally Posted by pandy71

Please provide source about Commodore x86 choice as next CPU ISA for Hombre chipset.

1. X86 is a little-endian CPU family and Commodore's Hombre project has big-endian requirements.

Commodore couldn't avoid Motorola's dead-end 68K problem i.e. Apple switched to PowerPC in 1994.

Apple maintained price vs performance" competitive against PC's 486SX-25/ 486SX-33 PC clones in Xmas Q4 1993 time period.

2. In the real timeline, "AmigaOS X86" was promoted by Amiga Inc. until the Amithlon and Haage & Partner's AmigaOS 3.9 licensing issue.

Amithlon has a fast big endian 68K to little endian X86 JIT translator software. Bernd Mayer has stated He didn't care about Amiga games.

Originally Posted by pandy71

My point about AROS is valid for any similar projects.

AROS on the non-68K platform is largely a failure just like Windows NT PowerPC and MIPS editions.

Originally Posted by pandy71

Brave statement but as i've wrote previously - please argue with Motorola/NXP on this, not with me - CF is seen as direct 68k line successor compatible at binary level with legacy 68k code (but with suboptimal performance). If you don't like this please redirect your comments to Motorola/NXP.

That's a bullshit propaganda. Furthermore, Cold Fire wasn't "fully superscalar" until Cold Fire 5.

FireTOS uses Motorola's CF68KLib which doesn't survive kick-the-OS game use cases. Hypervisor level Emu68 survives kick-the-OS game use cases.

Like Apple's 68K-to-PPC approach, FireTOS can only run OS-friendly 68K Atari TOS programs.

Apollo-Core's AC68080 V2/V4 doesn't need CF68KLib.

Originally Posted by pandy71

Legacy software, mostly games are serious issue for any modern platform - they are written usually in a way requiring special approach (perhaps something like DOSBox on PC).

To run legacy 16-bit software, the modern PC has many methods such as emulation, hardware accelerated virtualization, sandbox, and native hardware.

[ Show youtube player ]
Running 16-bit Windows programs on 64-bit Windows 11 via WINEVDM.

Modern PCs have compute power to run brute force DosBox.

Originally Posted by pandy71

I fully agree with you, AGA should be released way earlier, probably way before 1990, AGA should be also way better especially that J. Miner finished RANGER before his departure from Commodore.

From https://en.wikipedia.org/wiki/Amiga_Ranger_Chipset

1987 era Ranger has 128 color (7-bit color depth) 1024×1024 display from the 4096 color palette. VRAM provides the memory bandwidth.

Ranger has 2 MB Chip RAM address capability like on A3000's ECS and AGA.

It's a relatively minor update since the later Amiga OCS revision has a 6-bitplane 64-color EHB. Early A1000 didn't have 6-bit 64-color EHB mode. EHB mode's 5-bit color register is a cost-cut measure since proper 6-bit color registers consume additional transistors. There's a reason why Commodore outsourced AGA Lisa's fabrication to 3rd parties.

Lisa's raster and post-Alice Blitter on the Ramsey memory controller's 32-bit @ 25 Mhz bus would have improved pixel performance, but 3D requires compute power.

IBM's 1984 PGA can do 640x480 256 colors from a 4096 color palette.

IBM's 1984 release PGA was replaced by the 1987 released 8514.
IBM's 1984 release EGA was replaced by the 1987 released VGA.
IBM's VGA and 8514 were replaced by XGA which competed against PC cloner's SVGA/VESA. NEC's driven VESA made sure IBM XGA was dead as a revenge tactic for IBM's killing of NEC's PC-98. VESA BIOS standard was largely killed off by Microsoft's "Designed For Windows" 2D accelerators and Intel's UEFI Graphics Output Protocol (GOP).

[pandy71]

Originally Posted by hammer

Refer to CPU wars.

AFAIK there was no war between x86 and 68k - AFAIR there was war in x86 camp between many vendors providing own variation of the x86 CPU's.

Originally Posted by hammer

Cold Fire is useless for this topic. For Hombre, Commodore's legacy support is A1200 on a chip.

This is your point and it was already expressed multiple times but you didn't provided any hard fact to support it.

Originally Posted by hammer

1. X86 is a little-endian CPU family and Commodore's Hombre project has big-endian requirements.

Once again provide hard proof that Commodore considered x86 for Hombre.

Originally Posted by hammer

Commodore couldn't avoid Motorola's dead-end 68K problem i.e. Apple switched to PowerPC in 1994.

Apple maintained price vs performance" competitive against PC's 486SX-25/ 486SX-33 PC clones in Xmas Q4 1993 time period.

2. In the real timeline, "AmigaOS X86" was promoted by Amiga Inc. until the Amithlon and Haage & Partner's AmigaOS 3.9 licensing issue.

Amithlon has a fast big endian 68K to little endian X86 JIT translator software. Bernd Mayer has stated He didn't care about Amiga games.



AROS on the non-68K platform is largely a failure just like Windows NT PowerPC and MIPS editions.

Once again - don't link Commodore with all that happened 10 years after Commodore bankruptcy.

Originally Posted by hammer

That's a bullshit propaganda. Furthermore, Cold Fire wasn't "fully superscalar" until Cold Fire 5.

FireTOS uses Motorola's CF68KLib which doesn't survive kick-the-OS game use cases. Hypervisor level Emu68 survives kick-the-OS game use cases.

Like Apple's 68K-to-PPC approach, FireTOS can only run OS-friendly 68K Atari TOS programs.

Apollo-Core's AC68080 V2/V4 doesn't need CF68KLib.


To run legacy 16-bit software, the modern PC has many methods such as emulation, hardware accelerated virtualization, sandbox, and native hardware.

[ Show youtube player ]
Running 16-bit Windows programs on 64-bit Windows 11 via WINEVDM.

Modern PCs have compute power to run brute force DosBox.

Once again please argue not with me but with Motorola/NXP about CF being or not successor to 68k line.


With respect to all you point (sometimes valid) - perhaps nobody give a shit about old incompatible games if you can have 100MIPS+ at lower price than 68060. This is price of progress.

Originally Posted by hammer

From https://en.wikipedia.org/wiki/Amiga_Ranger_Chipset

1987 era Ranger has 128 color (7-bit color depth) 1024×1024 display from the 4096 color palette. VRAM provides the memory bandwidth.

Ranger has 2 MB Chip RAM address capability like on A3000's ECS and AGA.

It's a relatively minor update since the later Amiga OCS revision has a 6-bitplane 64-color EHB. Early A1000 didn't have 6-bit 64-color EHB mode. EHB mode's 5-bit color register is a cost-cut measure since proper 6-bit color registers consume additional transistors. There's a reason why Commodore outsourced AGA Lisa's fabrication to 3rd parties.

Lisa's raster and post-Alice Blitter on the Ramsey memory controller's 32-bit @ 25 Mhz bus would have improved pixel performance, but 3D requires compute power.

LOL - i can read wiki - no need to c&p.

Not minor as Ranger for example introduced upgrades to Paula (never upgraded since ICS times)

Yes, ICS Denise without EHB was in very first Amiga NTSC, quickly it was replaced by OCS and Commodore offered to all Amiga owners upgrade from ICS to OCS Denise.

And of course there is reason for Lisa outsourcing - simply CSG was unable to deliver required speed of the CLUT DPRAM - even on ECS they need to introduce kludge for CLUT by interleaving and efficiently reducing amount of colors from 4096 to 64 in productivity and SHires.
Lisa demanded complexity and at the same time speed (DPRAM with access time around 12..15ns)

Originally Posted by hammer

IBM's 1984 PGA can do 640x480 256 colors from a 4096 color palette.

IBM's 1984 release PGA was replaced by the 1987 released 8514.
IBM's 1984 release EGA was replaced by the 1987 released VGA.
IBM's VGA and 8514 were replaced by XGA which competed against PC cloner's SVGA/VESA. NEC's driven VESA made sure IBM XGA was dead as a revenge tactic for IBM's killing of NEC's PC-98. VESA BIOS standard was largely killed off by Microsoft's "Designed For Windows" 2D accelerators and Intel's UEFI Graphics Output Protocol (GOP).

How neatly you are manipulating - time, price and technology are completely irrelevant from your perspective, PGA cost was around 3000$ in 1984 - this was higher than cost of PC, add to this display.
EGA nad VGA has no HW acceleration - they are dumb framebuffer.
8514 and XGA providing HW acceleration at relatively acceptable for professional users price. VESA is SOFTWARE standard - BIOS extension - you could create VESA for other cards (like TIGA, XGA, 8514).
And in US and Europe nobody give a shit about NEC PC-98 - sorry.

[hammer]

Originally Posted by pandy71

AFAIK there was no war between x86 and 68k -

You're a naive fool.

Originally Posted by pandy71

AFAIR there was war in x86 camp between many vendors providing own variation of the x86 CPU's.

There are licensed 68000 clones. Licensed 68000 clone vendors weren't as capable as cloner AMD's CPU R&D.

Licensed 68000 clone vendors could not continue 68K's development and evolution.

X86 CPU clones and PC vendor clones enabled direct competition and allowed the PC platform to survive when the originators go anti-legacy or anticompetitive e.g. Intel Itanium IA-64 and IBM's PowerPC/MCA/XGA initiatives.

Itanium was Intel's anticompetitive move against X86 CPU cloners.

Originally Posted by pandy71

This is your point and it was already expressed multiple times but you didn't provided any hard fact to support it.

1. Prove the original Coldfire can beat the 1994 released 68060 rev 1 or HP PA-RISC 7150.

The original ColdFire core was launched in 1994 with a single-issue pipeline, no MMU, no FPU. Versions are also available with MAC and enhanced MAC units.

Coldfire being useless for Commodore's Hombre project is based on the real timeline of Motorola/Freescale's product release schedule.



2. https://en.wikipedia.org/wiki/Amiga_Hombre_chipset

According to Hombre designer Dr. Ed Hepler, Commodore intended to produce an AGA Amiga upon a single chip to solve the backward compatibility issues.

Originally Posted by pandy71

Once again provide hard proof that Commodore considered x86 for Hombre.

Again, Commodore's Hombre requirement specs have big-endian and good code density requirements.

X86 is a little-endian CPU family and Commodore already selling X86 PC clones up to 1994 Pentium class PCs.

Originally Posted by pandy71

Once again please argue not with me but with Motorola/NXP about CF being or not successor to 68k line.

Don't repeat Motorola's BS.

Originally Posted by pandy71

With respect to all you point (sometimes valid) - perhaps nobody give a shit about old incompatible games if you can have 100MIPS+ at lower price than 68060. This is price of progress.

It depends on many factors. Nintendo Switch doesn't have hardware superiority when compared to Xbox and Playstation, yet Switch's sales are very good.

RPi has 3rd party support superiority when compared to superior hardware wannabe Pi clones.

Originally Posted by pandy71

Not minor as Ranger for example introduced upgrades to Paula (never upgraded since ICS times)

Paula has 56 kHz improvement when paired with ECS or AGA.

Paula has minor refinments e.g. 8364R4 vs 8364R7.

Originally Posted by pandy71

Yes, ICS Denise without EHB was in very first Amiga NTSC, quickly it was replaced by OCS and Commodore offered to all Amiga owners upgrade from ICS to OCS Denise.

That's a pro-consumer move by Commodore.

Renee "Buffee" Cousins is designing AGA-on-ECS drop-in replacement chips.

Originally Posted by pandy71

And of course there is reason for Lisa outsourcing - simply CSG was unable to deliver required speed of the CLUT DPRAM - even on ECS they need to introduce kludge for CLUT by interleaving and efficiently reducing amount of colors from 4096 to 64 in productivity and SHires.
Lisa demanded complexity and at the same time speed (DPRAM with access time around 12..15ns)

CSG fabricated the Lisa chip in 1992 as seen in this photo
https://www.amibay.com/attachments/lisa-jpg.2461400/
This 1992 fabrication markings would be too late for the retail 1992 timeline.

There was a time when CSG wasn't capable of producing the Lisa chip.

The mainstream gaming resolution from 1990 to 1995 was about 320x240 level, hence productivity 640x480p mode was pointless for action games.

Originally Posted by pandy71

How neatly you are manipulating - time, price and technology are completely irrelevant from your perspective, PGA cost was around 3000$ in 1984 - this was higher than cost of PC, add to this display.

The technology standard exists for cost-reduced clones.

1987 IBM 8514 was expensive and the cloners lowered the cost e.g. ET4000AX was released in 1989.

https://archive.org/details/bub_gb_h.../n603/mode/2up
PC Mag 1992-08, page 604 of 664,
Diamond Speedstar 24 (ET4000AX ISA) has $169 USD retail.

There are other IBM 8514 clones such as
(https://en.wikipedia.org/wiki/IBM_8514)
ATI Technologies: the Mach8, Mach32,Graphics Vantage and 8514/Ultra
Chips and Technologies: F82C480 B EIZO - AA40 and F82C481 Miro Magic Plus
Matrox: MG-108
Paradise Systems: Plus-A, Renaissance Rendition II
Desktop Computing: AGA 1024 (also capable of emulating TIGA standards)
NEC: Multisync Graphics Engine
IIT AGX and Tseng Labs ET4000 are also referenced as being IBM 8514 compatible.


IBM PGA was discontinued in 1987 with the arrival of VGA and 8514.

IBM PGA standard was cloned:
Matrox PG-640, PG-1280 and QG-640 (for the DEC MicroVAX)
Dell NEC MVA-1024 card
Everex EPGA
Orchid Technology TurboPGA
Vermont Microsystems IM-640, IM-1024

Gaming PCs rose around the early 1990s, but the PC clone market has been building its full 32-bit CPU and VGA standard install base since 1987 which benefits IDsoftware's Doom 1/2's 4 million PC DOS sold copies. Any full 32-bit CPU-equipped desktop PC has gaming PC potential.
PC's full 32-bit CPU and VGA standard install base buildup has about 7 years from 1987 to the end of 1993.

Full 32-bit 68020/68030 CPU accelerated Amiga with OCS/ECS weren't enabled to participate in the AGA era games since an Amiga owner needs a full AGA machine upgrade. "Only the Amiga Makes it Possible" with the dumping of a full 32-bit 68020/68030 CPU (386DX class) with 32-bit Fast RAM equipped Amiga machine into a non-gaming role. Q4 1992 Amiga AGA is effectively "ground zero" for 256 color-enabled Amiga games.

Originally Posted by pandy71

EGA nad VGA has no HW acceleration - they are dumb framebuffer.

VGA software emulation is not wise. Ask Rendition about their mistake.

Originally Posted by pandy71

8514 and XGA providing HW acceleration at relatively acceptable for professional users price.

Reminder, ET4000AX is one of many lower-cost 8514 clones.

Originally Posted by pandy71

VESA is SOFTWARE standard - BIOS extension - you could create VESA for other cards (like TIGA, XGA, 8514).

VESA BIOS includes a firmware-level API abstraction.

Originally Posted by pandy71

And in US and Europe nobody give a shit about NEC PC-98 - sorry.

The US didn't give a shit about the Amiga gaming i.e. it's Nintendo game console and gaming PC land.

[pandy71]

Originally Posted by hammer

You're a naive fool.

LOL - if you need help then don't talk with mirror, go to shrink.

There was no war between 68k and x86 - they made own ecosystems - x86 mostly PC and later embedded, 68k Unix and from beginning serious embedded (VME) - literally 68k dominated automotive, industrial automation, military - many critical applications...

Originally Posted by hammer

There are licensed 68000 clones. Licensed 68000 clone vendors weren't as capable as cloner AMD's CPU R&D.

Licensed 68000 clone vendors could not continue 68K's development and evolution.

X86 CPU clones and PC vendor clones enabled direct competition and allowed the PC platform to survive when the originators go anti-legacy or anticompetitive e.g. Intel Itanium IA-64 and IBM's PowerPC/MCA/XGA initiatives.

There was no 68k clones - it was officially licensed - x86 is opposite - one of many examples is NEC v20/v30 - better than Intel - with 386 it was even more different implementations...

Originally Posted by hammer

1. Prove the original Coldfire can beat the 1994 released 68060 rev 1 or HP PA-RISC 7150.

Price vs performance... obviously you are ignoring one of the most important factors... Commodore never offered PA or 68060 based machines...

Originally Posted by hammer

The original ColdFire core was launched in 1994 with a single-issue pipeline, no MMU, no FPU. Versions are also available with MAC and enhanced MAC units.

Coldfire being useless for Commodore's Hombre project is based on the real timeline of Motorola/Freescale's product release schedule.

2. https://en.wikipedia.org/wiki/Amiga_Hombre_chipset

According to Hombre designer Dr. Ed Hepler, Commodore intended to produce an AGA Amiga upon a single chip to solve the backward compatibility issues.

Again, Commodore's Hombre requirement specs have big-endian and good code density requirements.

X86 is a little-endian CPU family and Commodore already selling X86 PC clones up to 1994 Pentium class PCs.

Don't repeat Motorola's BS.

All above - not sure why are you arguing with me not with Motorola or other sources? Use time machine, go back to 1994 and argue with Motorola, Commodore, IBM, Intel etc...

Originally Posted by hammer

It depends on many factors. Nintendo Switch doesn't have hardware superiority when compared to Xbox and Playstation, yet Switch's sales are very good.

RPi has 3rd party support superiority when compared to superior hardware wannabe Pi clones.

Switch was never targeted to gamers but to people willing to play from time to time in particular type of games.

Please don't mix various incompatible HW SoC's only because products based on them share some Pi in name...

Originally Posted by hammer

Paula has 56 kHz improvement when paired with ECS or AGA.

Paula has minor refinments e.g. 8364R4 vs 8364R7.

Are you working in marketing? As a used cars seller...?
btw please provide list of "minor refinments" for 8364R4 vs 8364R7

Originally Posted by hammer

Renee "Buffee" Cousins is designing AGA-on-ECS drop-in replacement chips.

LoL - used cars seller you are truly...

Originally Posted by hammer

CSG fabricated the Lisa chip in 1992 as seen in this photo
https://www.amibay.com/attachments/lisa-jpg.2461400/
This 1992 fabrication markings would be too late for the retail 1992 timeline.

There was a time when CSG wasn't capable of producing the Lisa chip.

Commodore outsourced Lisa to HP and VLSI - VLSI in 1993 - obviously there must be reason why Lisa from CSG was usually not populated on Amiga boards...

Originally Posted by hammer

The mainstream gaming resolution from 1990 to 1995 was about 320x240 level, hence productivity 640x480p mode was pointless for action games.

But with remark that in PC 320x240 is in fact 320x480 in 60Hz...

Originally Posted by hammer

1987 IBM 8514 was expensive and the cloners lowered the cost e.g. ET4000AX was released in 1989.

https://archive.org/details/bub_gb_h.../n603/mode/2up
PC Mag 1992-08, page 604 of 664,
Diamond Speedstar 24 (ET4000AX ISA) has $169 USD retail.

There are other IBM 8514 clones such as
(https://en.wikipedia.org/wiki/IBM_8514)
ATI Technologies: the Mach8, Mach32,Graphics Vantage and 8514/Ultra
Chips and Technologies: F82C480 B EIZO - AA40 and F82C481 Miro Magic Plus
Matrox: MG-108
Paradise Systems: Plus-A, Renaissance Rendition II
Desktop Computing: AGA 1024 (also capable of emulating TIGA standards)
NEC: Multisync Graphics Engine
IIT AGX and Tseng Labs ET4000 are also referenced as being IBM 8514 compatible.


IBM PGA was discontinued in 1987 with the arrival of VGA and 8514.

IBM PGA standard was cloned:
Matrox PG-640, PG-1280 and QG-640 (for the DEC MicroVAX)
Dell NEC MVA-1024 card
Everex EPGA
Orchid Technology TurboPGA
Vermont Microsystems IM-640, IM-1024

Gaming PCs rose around the early 1990s, but the PC clone market has been building its full 32-bit CPU and VGA standard install base since 1987 which benefits IDsoftware's Doom 1/2's 4 million PC DOS sold copies. Any full 32-bit CPU-equipped desktop PC has gaming PC potential.
PC's full 32-bit CPU and VGA standard install base buildup has about 7 years from 1987 to the end of 1993.

Full 32-bit 68020/68030 CPU accelerated Amiga with OCS/ECS weren't enabled to participate in the AGA era games since an Amiga owner needs a full AGA machine upgrade. "Only the Amiga Makes it Possible" with the dumping of a full 32-bit 68020/68030 CPU (386DX class) with 32-bit Fast RAM equipped Amiga machine into a non-gaming role. Q4 1992 Amiga AGA is effectively "ground zero" for 256 color-enabled Amiga games.


VGA software emulation is not wise. Ask Rendition about their mistake.


Reminder, ET4000AX is one of many lower-cost 8514 clones.



VESA BIOS includes a firmware-level API abstraction.

Once again - Tseng ET4000 is a dumb framebuffer without HW acceleration - this is official Tseng information provided in ET4000 datasheet - if you can please provide alternate datasheet where Tseng provide information about HW acceleration implemented in ET4000 (not VLB ET4000Wxxxx).

Of course you could create PGA clone - it was based on Intel 8088 CPU so it was not difficult task... - you could even made it on 8086 (NEC v30) so it could be faster than IBM PGA...

VESA is BIOS extension - and yes, BIOS is API.

Originally Posted by hammer

The US didn't give a shit about the Amiga gaming i.e. it's Nintendo game console and gaming PC land.

True (so no legacy compatibility seem to be not so big issue)

[Locutus]

Originally Posted by pandy71

There was no 68k clones

It might be extremely obscure, but Apollo shipped machines with their own bitslice 68020 implementation.


In the early 90's there was also a press release of a third party BiCMOS 680x0, It likely never taped out before their funding evaporated.

[GWhizz]

No

[pandy71]

Originally Posted by Locutus

It might be extremely obscure, but Apollo shipped machines with their own bitslice 68020 implementation.

It is not clear to me if this bit-slice approach was an implementation of the 68k ISA or pre-PRISM approach... Anyway it was not clone of the 68k as it was not available as IC and/or independent product.

Originally Posted by Locutus

In the early 90's there was also a press release of a third party BiCMOS 680x0, It likely never taped out before their funding evaporated.

So based on some arguments early 90's is time where 68k dying loosing competition with other CPU's families.

Meanwhile x86 has many officially licensed clones from various silicone vendors but also clones being white room re-implementation of the x86 with with some improvements over original Intel design, some of them are pin to pin compatible so they can be used as direct Intel x86 replacement (mentioned v20/v30 from NEC) some with different pin topology require dedicated motherboards (mostly some 386/486 clones).
This shows something fundamental to understand that it was never war between 68k and x86 (this doesn't eliminate competition or "war" between users of machines using x86 and 68k).

[Thomas Richter]

Originally Posted by pandy71

This shows something fundamental to understand that it was never war between 68k and x86 (this doesn't eliminate competition or "war" between users of machines using x86 and 68k).

I'm not sure what this is supposed to prove. It only shows that there wasn't a sufficiently large market on the 68K side to make such clones worthwhile.

[pandy71]

Originally Posted by Thomas Richter

I'm not sure what this is supposed to prove. It only shows that there wasn't a sufficiently large market on the 68K side to make such clones worthwhile.

Nothing - i've just refereed to hammer claim about war between 68k (68k lost this war) and x86:

Originally Posted by hammer

1. Hardware 680x0 lost the war.

i never recalled war between 68k and x86, i've recall something that can be called war but between machines using CPU's - and it was "war" between PC and Amiga but also between Amiga and for example Atari ST.
And i agree - x86 was so popular that it was worth to be cloned - 68k was in different position (but it was very popular anyway).

[qz3fwd]

Originally Posted by pandy71

Nothing - i've just refereed to hammer claim about war between 68k (68k lost this war) and x86:


i never recalled war between 68k and x86, i've recall something that can be called war but between machines using CPU's - and it was "war" between PC and Amiga but also between Amiga and for example Atari ST.
And i agree - x86 was so popular that it was worth to be cloned - 68k was in different position (but it was very popular anyway).

I would say that m68k was not very popular in the consumer space, at least in the USA, where the units shipped/sold of 68k based architectures is really dwarfed by the volume x86 sold. I remember Mac’s back in the day at places like Circuit City had a Mac specific area with a Mac guy who only sold Mac’s and the foot traffic was always almost NILL compared to the rest of the store, and they had to have a dedicated Mac guy to try to convince customers to consider a Mac. Never saw Amigas, nor many Atari’s. Especially odd considering both Atari and Amiga were created here.

[Bruce Abbott]

Originally Posted by Thomas Richter

I'm not sure what this is supposed to prove. It only shows that there wasn't a sufficiently large market on the 68K side to make such clones worthwhile.

The other reason was that there no need for other manufacturers to clone the 68000 when they could just get a license from Motorola. That didn't happen so much on the x86 side because Intel soon realized they would lose market share if they let others copy their design.

Originally Posted by pandy71

Meanwhile x86 has many officially licensed clones from various silicone vendors but also clones being white room re-implementation of the x86 with with some improvements over original Intel design, some of them are pin to pin compatible so they can be used as direct Intel x86 replacement (mentioned v20/v30 from NEC)

Er,

NEC V20

In 1982, Intel sued NEC over the latter's ?PD8086 and ?PD8088. This suit was settled out of court, with NEC agreeing to license the designs from Intel.

In late 1984, Intel again filed suit against NEC, claiming that the microcode in the V20 and V30 infringed its patents for the 8088 and 8086 processors. NEC software engineer Hiroaki Kaneko had studied both the hardware design of the Intel CPUs and the original Intel microcode.

NEC never attempted to get official licenses, they just blatantly ripped off Intel.

And they weren't the only ones. AMD's 386DX chip was delayed for several years due to a dispute with Intel over whether its existing agreements to produce 8086 and 80286 CPUs included it.

Intel also sued Cyrix even before they released their Cx486SLC.

When Intel developed the 586 they changed the name to Pentium because they couldn't copyright a number. They were not at all happy with clones being produced, an attitude that continues today.

[pandy71]

Originally Posted by Bruce Abbott

The other reason was that there no need for other manufacturers to clone the 68000 when they could just get a license from Motorola. That didn't happen so much on the x86 side because Intel soon realized they would lose market share if they let others copy their design.

68k was on opposite extremely popular in automation https://en.wikipedia.org/wiki/VMEbus?useskin=vector

Originally Posted by Bruce Abbott

Er,

NEC V20

NEC never attempted to get official licenses, they just blatantly ripped off Intel.

And they weren't the only ones. AMD's 386DX chip was delayed for several years due to a dispute with Intel over whether its existing agreements to produce 8086 and 80286 CPUs included it.

Intel also sued Cyrix even before they released their Cx486SLC.

When Intel developed the 586 they changed the name to Pentium because they couldn't copyright a number. They were not at all happy with clones being produced, an attitude that continues today.

Most of those lawsuits was settled out of court - Intel right was possibility to raise case against other vendor and other vendors rights was to not settle case out of court - NEC case was judged on NEC favor so i've wrote nothing wrong. It was common practice to settle two independent teams, one using reverse engineering and second one using white room to create product. NEC made CPU sufficiently different than Intel so.

The court also determined that NEC did not simply copy Intel's microcode, and that the microcode in the V20 and V30 was sufficiently different from Intel's to not infringe Intel's patents.

[Locutus]

Originally Posted by pandy71

It is not clear to me if this bit-slice approach was an implementation of the 68k ISA or pre-PRISM approach... Anyway it was not clone of the 68k as it was not available as IC and/or independent product.

The OS media for it is just 68k binaries, nothing to do with PRISM. It almost surely had a custom MMU not compatible with the 68451, but if it could run AEGIS userland it was Good Enough (C).


But, this is terribly OOT (and lets be honest, the whole thread is just silly derailment anyway)

[pandy71]

Originally Posted by Locutus

The OS media for it is just 68k binaries, nothing to do with PRISM. It almost surely had a custom MMU not compatible with the 68451, but if it could run AEGIS userland it was Good Enough (C).

Sun didn't use 68451 - it was too slow for them, they created own MMU for 68000.
Had no experience with Apollo products, also can't find anything above provided leaflet - bitsavers also don't provide too many documents so from my perspective it is very interesting but also very unknown line of products.

Originally Posted by Locutus

But, this is terribly OOT (and lets be honest, the whole thread is just silly derailment anyway)

Well... as you pointed whole thread was derailed many pages earlier - now we have discussion about PC CPU's and graphics chipsets...

[chb]

Originally Posted by Locutus

It might be extremely obscure, but Apollo shipped machines with their own bitslice 68020 implementation.

While it is certainly off topic, this is one of the more interesting information in this thread. Are you sure they were implementations of the 68020 architecture/ISA, and not the 68000? All the information I can find on the web shows that the models with a custom bitslice CPU (DSP160/DN160/460/660) were released in the beginning of 1984 or earlier, before the 68020 was even announced.


It also would make more sense for them to do their own (improved) implementation of the 68000 instead of the 68020, as the 68000 did not fulfill the Popek/Goldberg virtualization requirements; AFAIR Apollo even had two 68000 operating in lockstep on their early machines as a workaround.