@Hypex

Quote:

It was however competitive with 68K which was stagnant. Even when 68K was still around ARM was said to compete and beat 68K in performance at the same clock rate. So I read online in places. But, ARM wasn't big endian, support was only added later. Had it survived, the 88K would have made a logical replacement for the 68K. Thanks Motorola!

Note why I stated "other RISC and X86 competitions".

ARM supports bi-endianness from ARM versions 3 and above.

ARM architecture supports two big-endian modes, called BE-8 and BE-32.

ARMv3 up to ARMv5 only supports BE-32 e.g. ARM6 with 12 Mhz and 10 MIPS. ARMv3 was introduced in 1991.

ARMv6 introduces BE-8 e.g. ARM11 with 532–665 MHz. Introduced in October 2001.

Last edited by Hammer on 05-Jan-2024 at 08:32 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, KS 3.2, PiStorm/RPi 3a/Emu68)

@Kronos

Quote:

Kronos wrote: @OneTimer1 The Walker was still using AGA which was "meh" in 92, so not a "seller" in 96 (or later depending on how long it was supposed to stay available) The Walker was based on a 68030 only slightly less underwhelming as the 020 in the barely selling A1200. The Walker would have made compromises in compatibility beyond the slightly off floppy in the A1200. Nothing in the Walker suggest that it could have been made cheaper than an A1200 Now if C= had introduced something like that in 92 instead of A4000/1200 it might have made some sense and even sold in o.k. numbers, but in 95/96 it was nothing more than a brainfart.

For the 1995 era context, Escom is just a "catalog engineering" company with weak internal R&D. The same for Gateway 2000. Escom's chip design R&D is weaker than the 3DO Company.

Apollo-Computer (AC68080, SAGA, and V4's Maggie3D on Intel Altera FPGA platform) has superior R&D efforts when compared to Escom.

The main R&D bulk for a major PC clone box shippers comes from X86 CPU, X86 chipset, and SVGA clone vendors e.g. Intel, AMD, ATI, NVIDIA, VIA, 3DFX, SIS, Cyrix, S3, Matrox, 3DLabs and etc.

Commodore International has a Commodore Semiconductor Group that serves a similar role as the PC market's chip component vendors.

The Amiga can't be another PC box shipper.

The modern Apple has chip R&D capability.

Last edited by Hammer on 07-Jan-2024 at 06:28 AM.
Last edited by Hammer on 05-Jan-2024 at 08:59 PM.
Last edited by Hammer on 05-Jan-2024 at 08:54 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, KS 3.2, PiStorm/RPi 3a/Emu68)

@Kronos

Quote:

Kronos wrote: @kolla An ATX Amiga would have only made sense as an A4000 like systems, which already existed as the A4000T which could be mounted in (big) off the shelf AT tower cases. So lots of engineering for solving a non problem on a niche product. Now if you were talking about something like a low end version of the Draco.. Basic backplane with ISA slots (see here ) and maybe onboard IDE. 030-060 CPU card with some RAM and minimal boot ROM. Basic RTG SW compatible with the supplied ISA-VGA (based on the same chips as the early Z2 RTG cards). Sure, might have made sense.

ATX form factor Amiga with A4000's chip components has a large chip count and uncompetitive chip integration.

For my HTPC box, my low-cost ($139 AUD) MSI B650M mATX motherboard has a very low chip count when compared to CD32. AMD APU is an SoC (System on Chip) with AMD/ASmedia's Southbridge for extra I/O. PCB quality influences PCI version support e.g. PCIe 5.0 to PCIe 3.0. PCIe 5.0 with higher cost X670E or B560E-based motherboards.

Low-cost Raspberry Pi 4B also uses SoC integration.

For the 1995 era context, PC motherboards are using higher chip integration with North and Southbridge chipsets. My 1st Pentium motherboard in 1996 was a baby ATX form factor with an Intel 430VX chipset i.e. https://theretroweb.com/motherboards/s/pcpartner-mb520n-35-8258-xx

Amiga 4000TX https://amitopia.com/the-amiga-4000tx-rev-1-1-motherboard-is-finally-ready/
A4000's native graphics subsystem can not be upgraded i.e. AAA upgrades will cause the whole board to be thrown away while the PC equivalent has upgradable native SVGA PCI cards. Amiga's game console chipset origins are not compatible with the "expandable desktop computers" concept.


https://videocardz.com/newz/amd-4800s-desktop-kit-a-pc-repurposed-apu-from-xbox-series-x-has-been-tested
This AMD 4800S recycled Xbox Series X's game console SoC with PCIe 4.0 4 lanes for video card upgrades. AMD 4800S's IGP is disabled due to chip yields or non-compliant desktop PC GPU. AMD 4800S is not cost-optimal for desktop PCs since the large semi-custom 56 CU RDNA 2 IGP is thrown away.

Last edited by Hammer on 05-Jan-2024 at 09:46 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, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hypex

Quote:

Hypex wrote: @Karlos Quote: The whole who owns what shell game is the absolute death knell of the platform as it stands and if it's come to this, as much as it pains me to say it, it's really time to move on. I can see your point. There is likely too much effort to port it in any fashion to new hardware. Let's face it, all it would do is move a buggy code base to another platform, and drag all the old bugs across!

I don’t think a complete, holus-bolus port is the prize.

Having the ownership question taken out of the equation frees up any reticent and reluctant developer to port and/or reimplement any functions without fear of potential litigation.

The same loss of enthusiasm can be seen in score-topping arcade game players when the machine is periodically and infrequently reset, wiping the leaderboard.

Last edited by agami on 05-Jan-2024 at 11:01 PM.

_________________
All the way, with 68k

@Hammer

Sometimes I think you are just a SLM (small language model) solely trained on discarded cardboard boxes found behind a PC parts shop.....

The context is that in 1995 ESCOM could have spend an 8 or 9 figure sum to develop a competitive chipset. They did not have that money, nor did they have the time so they choose wisely not to try.

At this point they had the choice to embrace what 3rd party companies had already done and create a "NG" Amiga (68k, PPC,MIPS,ARM,x86, whatever) or to hunt down a limited supply of obsolete NOS parts to build an outdated Amiga.

They did not choose wisely on this one.


As for Apollo/Vampire, they use todays HW to simulate a fan-art mid 90s chipset and took about a decade to get that point. O.k. as a hobby project, but not comparable with doing such a chip not only with the tools available at that time and within a short timeframe. No in that case they also would have made sure that it could be turned into a cost effective and reliable ASIC on whatever process node would have been available.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Hammer

Quote:

Hammer wrote: For the 1995 era context, Escom is just a "catalog engineering" company with weak internal R&D. The same for Gateway 2000. Escom's chip design R&D is weaker than the 3DO Company.

Escom's chip design was equal to that of Hyperion (they had no chip design).

I had a job interview with Petro and Dr. Kittel at Amiga Tech. in Bensheim, they told me they wanted to outsource the development (hard- software) to other companies.

Petro who worked for Escom before, never had any experience with R&D. He gave the development orders for the Walker to a company outside the Amiga market.

@thread

Quote:

Kronos wrote: The context is that in 1995 ESCOM could have spend an 8 or 9 figure sum to develop a competitive chipset. They did not have that money, nor did they have the time so they choose wisely not to try.

The most modern Amiga chip set was the CD32 design, but sadly it was still AGA. Some IO Parts where integrated into one chip, I don't know if Amiga Tech. had access to this designs, it could have lead to a cheaper and smaller A1200 board.

Quote:

Kronos wrote: At this point they had the choice to embrace what 3rd party companies had already done and create a "NG" Amiga (68k, PPC,MIPS,ARM,x86, whatever) or to hunt down a limited supply of obsolete NOS parts to build an outdated Amiga.

They produced a huge pile of A1200 they couldn't sell any more.

Quote:

Kronos wrote: As for Apollo/Vampire, they use todays HW to simulate a fan-art mid 90s chipset and took about a decade to get that point. O.k. as a hobby project, but not comparable with doing such a chip not only with the tools available at that time and within a short timeframe.

It is great what can be done with FPGAs today, Dennis van Weeren did it and the Apollo group did it again, it would have been great if Amiga Tech. would had such designs for an ASIC back in the 90ies ...

But a real future would have meant:
- leaving the dead 68k technology
- dumping ECS/AGA compatibility

Mac users today don't care for 68k compatibility or lost B/W bitplane modes of their Macintosh.

@OneTimer1

Quote:

OneTimer1 wrote: @thread The most modern Amiga chip set was the CD32 design, but sadly it was still AGA. Some IO Parts where integrated into one chip, I don't know if Amiga Tech. had access to this designs, it could have lead to a cheaper and smaller A1200 board.

A lot of that integration had beed done with the OG FAT Agnus (A500/2000B vs A1000/2000A) and Gary. While there was still some smaller parts left on the A1200 I doubt the cost savings would have have been able to offset the costs of doing a new A1200 mobo. Even less with with the limited number of NOS chips limiting the possible sales.
Which of course raises the question if these CD32 specific chips were even "found" in sufficient numbers since even reproducing them existing designs was clearly out of the scope what could/wanted to do.
Quote:

it would have been great if Amiga Tech. would had such designs for an ASIC back in the 90ies ...

Back into w##nking of to obscure alt history...
- they did not have scale to develop such chips
- they did not even have the scale to have these chips made if such designs had been given to them for free
- noone knows if these designs would have even worked in 1996 ASIC
- one can only guess at what the price and performance of those chips would have been as a 1996 ASIC (I'd guess it would run at much lower clock then it does on the FPGA)

Last edited by Kronos on 06-Jan-2024 at 09:42 AM.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Hammer

Quote:

A4000's native graphics subsystem can not be upgraded

That’s the case for any motherboard with onboard graphics, and the trick is to upgrade with a graphics card. As for "native", that’s just a matter of software, amiga is no different than any other hardware platform, the difference is AmigaOS relying on third party so called RTG solutions to achieve what other OSes have built in.

Btw, my ATX board suggestion was in respons to the Walker, which made zero sense.

Last edited by kolla on 06-Jan-2024 at 10:30 AM.
Last edited by kolla on 06-Jan-2024 at 10:23 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

The nice thing about being in 2024 is that we can implement 68K CPU expansions that are more performant than you could have imagined the day you got your first Amiga (well, assuming you aren't a late coming retro enthusiast) and FPGA are sophisticated and performant enough that you can implement your own compatible chipset components that can easily exceed the limitations of the old hardware, if you have the skills and means.

Someone out there, somewhere is probably doing that right now. A completely custom Amiga hardware compatible machine that happily supports OCS/ECS/AGA (with bandwidth enhancements) and RTG, etc.

Man, that would be my perfect birthday gift.

_________________
Doing stupid things for fun...

@Karlos

ocs for games
other graphics for rest
68k and ppc core
that will be ideal amiga in fpga
nothing more is needed

@ppcamiga1

Forget the PPC, it's an unnecessary complication, especially if you're just planning to run an Emu68 type solution for everything,or a native ARM 32-bit/Big Endian mode with a 68K emulation in it.

PPC emulation at an acceptable speed and compatibility is a challenge, and you have the added dilemma that you probably need to emulate the MMU properly as well if you want to run an NG OS on it. So you're looking at QEMU on a modern PC to get an acceptable level of performance out of it, otherwise you're setting yourself up for disappointment.

A standalone machine ideally should be a relatively low power, quiet thing and for bonus points, installable into an Amiga case.

_________________
Doing stupid things for fun...

@ppcamiga1

OS4 is garbage anyways, so what’s the purpose of PPC? Better have fast 68k.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@ppcamiga1

I couldn't disagree with you more. I think you think all these PPC machines will be sustained by your personal CPU fetish! Here's the truth, THIS IS NOT SUSTAINABLE FOR THE INVESTMENT REQUIRED FOR TREVOR TO KEEP PEDDLING WEAK PPC SOC based AmigaOnes! It's over and AmigaOS/MorphOS both have to transition to another CPU now.

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

@Karlos

You are aware that the user @ppcamiga1 in this thread, is the same @ppcamiga1 user who often plays a troll in other threads, right?

Just thought I'd check, because your logical and dignified response indicates that you are either unaware of the above fact, or for some unfathomable reason you've decided to give him yet another chance at civil discourse.

_________________
All the way, with 68k

@kolla

Quote:

kolla wrote: @Hammer BTW, my ATX board suggestion was in response to the Walker, which made zero sense.

I agree, in 1996 we'd have been better off with a BoXeR instead of a Walker.

_________________
All the way, with 68k

@agami

Quote:

agami wrote: I agree, in 1996 we'd have been better off with a BoXeR instead of a Walker.

A slightly less crappy design which may or may not have been vaporware....

Once the Draco had dropped anybody and anything that still gave AGA 5 second of thought was and is pure retro (or more likely fake retro).

In that timeframe a forward looking Amiga was still a viable concept and the fact that nothing ever came of this project is a blessing. Off course that fact that it was started is also a curse.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Quote:

Kronos wrote: @Hammer Sometimes I think you are just a SLM (small language model) solely trained on discarded cardboard boxes found behind a PC parts shop..... .

What's the matter?

Quote:

The context is that in 1995 ESCOM could have spend an 8 or 9 figure sum to develop a competitive chipset. They did not have that money, nor did they have the time so they choose wisely not to try.

Corporate's skill set and engineering talent matters, not just the money.

For example
https://en.wikipedia.org/wiki/Athlon

AMD founder and then-CEO Jerry Sanders aggressively pursued strategic partnerships and engineering talent in the late 1990s, working to build on earlier successes in the PC market with the AMD K6 processor line. One major partnership announced in 1998 paired AMD with semiconductor giant Motorola to co-develop copper-based semiconductor technology, resulting in the K7 project being the first commercial processor to utilize copper fabrication technology. In the announcement, Sanders referred to the partnership as creating a "virtual gorilla" that would enable AMD to compete with Intel on fabrication capacity while limiting AMD's financial outlay for new facilities. The K7 design team was led by Dirk Meyer, who had previously worked as a lead engineer at DEC on multiple Alpha microprocessors. When DEC was sold to Compaq in 1998 and discontinued Alpha processor development, Sanders brought most of the Alpha design team to the K7 project

Athlon's EV6 bus, high clock speed, and design were influenced by ex-DEC's engineering talent.

AMD used Motorola's fabrication tech with different clock speed results.

For competitive R&D chipset business, let's use ASmedia example;

ASRock's parent is Pegatron Corp. The major shareholder of Pegatron Corp. is Asustek Computer Inc. "ASUS" (http://www.pegatroncorp.com/investorRelation/majorShareholders) along with private and other institutional investors.

ASMedia reports "Asustek Computer Co., Ltd." as 40.74% owner (http://www.asmedia.com.tw/eng/e_shareholder_list.php)

Within ASUS 2016 audited Annual Report, ASUS owns ASMEDIA TECHNOLOGY INC. 41.08% directly and 53.37% due to directors of ASUS having ownership in ASMEDIA (https://asus.todayir.com.tw/attachment/20170614180848496941490_en.pdf pg.80);
it's also included in the corporate organizational chart as a subsidiary.

ASMedia produces designs for USB, PCI Express, and SATA controllers. Excluding the AM4's X570 PCIe 4.0 chipset, all of the AM4 chipsets for AMD's Zen micro-architecture were designed by ASMedia.

https://www.anandtech.com/show/17585/amd-zen-4-ryzen-9-7950x-and-ryzen-5-7600x-review-retaking-the-high-end/4

ASmedia designed Southbridge chipsets for AM5's B650 and X670. ASmedia's southbridge chipset is at PCIe 4.0 level.

AM5's PCIe 5.0 from APU is provided by AMD.

Nether Escom or Gateway is at the same caliber as ASMedia's chip R&D resources with leadership direction and startup funds from ASUS.

Escom has hired former Commodore Germany's personnel.

Quote:

As for Apollo/Vampire, they use todays HW to simulate a fan-art mid 90s chipset and took about a decade to get that point. O.k. as a hobby project, but not comparable with doing such a chip not only with the tools available at that time and within a short timeframe. No in that case they also would have made sure that it could be turned into a cost effective and reliable ASIC on whatever process node would have been available.

FYI, Phase 5 has used FPGA chips in its accelerator products, but it's not part of Escom.

Phase 5 has A\Box around 1997.
http://www.bambi-amiga.co.uk/amigahistory/1stabox.html
http://www.bambi-amiga.co.uk/amigahistory/aboxspec.html

AMD was pushing DDR for the original K7 Athlon while Intel was pushing RamBus for Pentium 3.

NVidia's nForce 1 chipset has a 128-bit DDR memory bus for K7 and Pentium 3-based original Xbox.

Last edited by Hammer on 07-Jan-2024 at 07:46 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, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Again 2kb off of topic PC parts babble intervened with some bad takes.

The programable logic of various kinds used in CPU cards is just a fraction of would have been needed to emulate AGA or better GFX.

A\Box was vaporware and is as such irrelevant. There is a reason why their next (and still vaporware) design the "Pre\Box" was much more conventual with a PCI based (of the shelf) GFX card.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@kolla

Quote:

kolla wrote: @Hammer Quote: A4000's native graphics subsystem can not be upgraded That’s the case for any motherboard with onboard graphics, and the trick is to upgrade with a graphics card. As for "native", that’s just a matter of software, amiga is no different than any other hardware platform, the difference is AmigaOS relying on third party so called RTG solutions to achieve what other OSes have built in. Btw, my ATX board suggestion was in respons to the Walker, which made zero sense.

Barebones DOS VGA (VBIOS) and VBE can be IGP (e.g. Ryzen 7000/8000's RDNA 2/RDNA 3 IGP) or discrete AMD/NVIDIA GPUs. Intel Xe and ARC IGP/GPUs have abandoned VGA.

Amiga's graphics architecture wasn't continued into evolved hardware with backward compatibility until Apollo Core's SAGA.

Amiga's graphics architecture topology didn't lead into discrete graphics topology e.g. you add AGA card on Amiga 3000's Zorro III bus and run AGA games?

_________________
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, KS 3.2, PiStorm/RPi 3a/Emu68)