Poster | Thread |
vox
 |  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 10:21:52
| | [ #141 ] |
|
|
 |
Elite Member  |
Joined: 12-Jun-2005 Posts: 3957
From: Belgrade, Serbia | | |
|
| @bhabbott
Falcon was fairly priced as compared to a1200 with 030, 16bit audio, midi and no dsp. More expensive versions ns had hdd and more ram, but then again cheaper then a1200 buffed alike or a4000 030 as more comparable sys. Dsp wasnt reason to its fail, it was Ataris loss of interest in home computing and not investing in its software development
Dsp was key to its limited success on demo scene, home studio use and added software, which indeed surpassed what Amigas can do, with exceptions of 060 equipped amigas. Unofficially 060 expansion exists for abait rare
Why being anti dsp? At the time it was fairly priced moto chip that adds a lot functionalities, and dsp appears on better Amiga, PC and Mac sound cards too.
Datatype example is fair idea how its use could be AmigaOS adopted, as well as e.g. AHi driver
_________________ OS 3.x AROS and MOS supporter, fi di good, nothing fi di unprofessionalism. Learn it harder way! SinclairQL and WII U lover :D YT http://www.youtube.com/user/rasvoja |
|
Status: Offline |
|
|
kolla
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 11:22:18
| | [ #142 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3475
From: Trondheim, Norway | | |
|
| @cdimauro
Quote:
Commodore's primarily (SIC!) market |
If Commodore had solely focused on what you consider its primary market, then we would not even be here._________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|
pixie
 |  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 11:27:47
| | [ #143 ] |
|
|
 |
Elite Member  |
Joined: 10-Mar-2003 Posts: 3475
From: Figueira da Foz - Portugal | | |
|
| |
Status: Offline |
|
|
vox
 |  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 15:37:06
| | [ #144 ] |
|
|
 |
Elite Member  |
Joined: 12-Jun-2005 Posts: 3957
From: Belgrade, Serbia | | |
|
| @pixie
Ah A1200 Mini is coming! Another ARM box, well I hope at least keyboard will be usable this time :D _________________ OS 3.x AROS and MOS supporter, fi di good, nothing fi di unprofessionalism. Learn it harder way! SinclairQL and WII U lover :D YT http://www.youtube.com/user/rasvoja |
|
Status: Offline |
|
|
matthey
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 16:48:42
| | [ #145 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2007 Posts: 2752
From: Kansas | | |
|
| vox Quote:
Falcon was fairly priced as compared to a1200 with 030, 16bit audio, midi and no dsp. More expensive versions ns had hdd and more ram, but then again cheaper then a1200 buffed alike or a4000 030 as more comparable sys. Dsp wasnt reason to its fail, it was Ataris loss of interest in home computing and not investing in its software development
|
Atari 68k hardware initially targeted the Mac more than Amiga but the DSP allowed them to target the Amiga. The DSP was easier to add than major chipset enhancements and cheaper than high end CPU upgrades which were preferable upgrades due to more utilization.
vox Quote:
Dsp was key to its limited success on demo scene, home studio use and added software, which indeed surpassed what Amigas can do, with exceptions of 060 equipped amigas. Unofficially 060 expansion exists for abait rare
|
Are 68060 accelerators really that rare?
Old Amiga 3000(T)/4000(T) 68060 accelerators Apollo 3060 https://bigbookofamigahardware.com/bboah/product.aspx?id=245 Apollo 4060 https://bigbookofamigahardware.com/bboah/product.aspx?id=247 AT4060 https://bigbookofamigahardware.com/bboah/product.aspx?id=248 GVP A4060 https://bigbookofamigahardware.com/bboah/product.aspx?id=1416 GVP T-Rex-II https://bigbookofamigahardware.com/bboah/product.aspx?id=255 P5 Cyberstorm Mk-I https://bigbookofamigahardware.com/bboah/product.aspx?id=1415 P5 Cyberstorm Mk-II https://bigbookofamigahardware.com/bboah/product.aspx?id=263 P5 Cyberstorm Mk-III https://bigbookofamigahardware.com/bboah/product.aspx?id=264 P5 Cyberstorm PPC604e https://bigbookofamigahardware.com/bboah/product.aspx?id=251 QuikPak 4060 https://bigbookofamigahardware.com/bboah/product.aspx?id=268 QuikPak 4000-060 XP https://bigbookofamigahardware.com/bboah/product.aspx?id=1892
There are newer cards too like the following (which both show a game menu working fine with 68060@100+MHz.
Warp 1260 105MHz AGS Amiga Game Selector Alien Breed 3D 2 MOD gameplay https://www.youtube.com/watch?v=N5T2gv_nfNs
BFG9060 68060 Amiga 4000 Accelerator review part 2/2 THIS CARD IS AWESOME!!! (101MHz) https://www.youtube.com/watch?v=7bK_Jmx6Ndk
There are 68060 accelerators for other Amiga models like Terrible Fire accelerators. There are 68040 accelerators that have been upgraded to 68060 accelerators. Individual Computers claimed to be designing 68060 accelerators. There are likely 25-50 unique accelerator designs using a 68060. There is a reason 68060 supplies have disappeared which was likely tens of thousands of CPUs. I expect there are more 68060 accelerators than PPC AmigaNOne computers even though it is unlikely to be the largest 68k Amiga market due to cost.
1. ARM emulation (THEA500 Mini, RPi hardware, A600GS) 2. WinUAE 3. real 68k Amiga hardware 4. FPGA hardware (AC/Vamp, MiSTer, MiST, FPGA Arcade, MiniMig, FleaFGPA Ohm, etc.) 5. PPC AmigaNOne
WinUAE may still be ahead of ARM emulation depending on how Amiga users are counted and whether THEA500 Mini users should count as Amiga users at all. FPGA hardware may be ahead of real Amiga hardware depending on how it is counted as some of the FPGA hardware is universal. The 68k Amiga users may be numbered in the hundreds of thousands and growing by tens of thousands per year depending on if THEA500 Mini users count due to limited hardware. One thing is for sure though. The PPC AmigaNOne is in last place with a few thousands users at best and likely falling despite new A1222 hardware with users leaving due to failure, lawsuits and broken hardware. Maybe we will get some less limited and more competitive 68k Amiga hardware when the Amiga IP squatter road blocks are finally eliminated.
Last edited by matthey on 21-Sep-2024 at 04:56 PM. Last edited by matthey on 21-Sep-2024 at 04:51 PM.
|
|
Status: Offline |
|
|
vox
 |  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 17:41:21
| | [ #146 ] |
|
|
 |
Elite Member  |
Joined: 12-Jun-2005 Posts: 3957
From: Belgrade, Serbia | | |
|
| @vox
Falcon has significantly improved graphics and OS to STE, bigger jump then ECS to AGA.
060 accelerators on Falcon are rare, that I ment.
On AMiga there was even a 4000/060 by ESCOM https://www.kevinsimon.co.uk/amiga-a4000-60/ and DraCo Clone with 060 as standard https://bigbookofamigahardware.com/bboah/product.aspx?id=43
In general, you are right number of 060 equipped AMigas likely surpass number of AmigaOne G3 and G4, SAM440, SAm460, x1000, x5000, Tabor and Pegasos boards with OS4, combined..
On Amiga, new Polish Warp060 by CS Lab is indeed most advaned and avail new. https://amigawarp.eu/product/warp-1260-board-for-amiga-1200/
Its quite expensive without CPU but were capable. In some aspects its more modern then V4 Vampires
Warp 1260 Features Warp 1260 is a composite accelerator board for the Amiga 1200 computer. It joins several features to create a great experience using your good, old Amiga:
Powerful MC68060 Turbo (some rev.6 CPU’s working even up to 105MHz)
Frequency can be easily switched by software. 64kB external L2 cache for low latency access of DDR memory (MC68060 CPU has to be provided by the customer) 256MB memory expansion
224MB for AmigaOS 32MB for RTG graphics Picasso96 compatible RTG graphics card
Fast 128bit Blitter, accelerating common 2D gfx operations. Fast CPU access to graphics memory due to unified memory architecture Connector for optional WarpedVision FlickerFixer Automatic switching between Amiga and RTG modes (when working in tandem with WarpedVision) Sana2 compatible WiFi network card
Allowing internet and local network access TCP/IP software stack (like Roadshow) required. 16Bit Audio card with mixer
High quality 16bit / 44.1kHz audio 24bit onboard sampler for original Amiga audio Mixer output pass through to Digital Video output AHI Driver is not ready yet; it will be available with future updates. USB IO card
HID (mouse/keyboard) support. Mass storage support (Bootable) Work even in the early boot menu, not requiring any drivers or additional USB stack. Fast microSD card slot with DMA support
Fast transfers (up to ~20MB/s) CPU offload thanks to Direct-Memory-Access Bootable Additional onboard Fast IDE (dedicated for CF cards)
Faster access and transfers compared to internal Amiga IDE Bootable FastROM
4 slots for different Kickstarts, selectable by software Last edited by vox on 21-Sep-2024 at 05:50 PM. Last edited by vox on 21-Sep-2024 at 05:48 PM. Last edited by vox on 21-Sep-2024 at 05:44 PM.
_________________ OS 3.x AROS and MOS supporter, fi di good, nothing fi di unprofessionalism. Learn it harder way! SinclairQL and WII U lover :D YT http://www.youtube.com/user/rasvoja |
|
Status: Offline |
|
|
matthey
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 21:13:41
| | [ #147 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2007 Posts: 2752
From: Kansas | | |
|
| @vox An ARM SoC can provide a lot of I/O modernization but it can't provide a 68060@1GHz or higher like an Amiga SoC. If Amiga users are impressed by 68060@100MHz, what would they think of a 68060@1GHz? Does anyone doubt that the games in the videos would still work with 68060@1GHz? How funny is it that rev 6 68060s sell for more than the 68060 accelerators while an Amiga ASIC SoC chip with 68060@1GHz could be produced for less than the cost of lunch? How long will the Amiga IP squatting road blocks hold out with PPC AmigaNOne failing?
|
|
Status: Offline |
|
|
OneTimer1
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 21-Sep-2024 21:57:41
| | [ #148 ] |
|
|
 |
Super Member  |
Joined: 3-Aug-2015 Posts: 1258
From: Germany | | |
|
| @vox
Quote:
On Amiga, new Polish Warp060 by CS Lab is indeed most advaned and avail new.
|
I don't criticize the design of your Warp60, but it wouldn't had been a solution for a cheap A500/A1200 update back in the early 90ies.
I can't even blame the designers of this accelerator, the 68060 CPU is much to expensive for the performance it provides, it passes the limits of sanity where you should ask yourself if expanding a 14MHz 16/32 bit design with a 50Mhz 32 Bit CPU, tuned for 32 burst access that can't be provided by the architecture of the host system, normally you would say "No let's update to a more performant system"
A 28MHz 68EC/LC30 by C= might have been a vital option but most things beyond that, needed a new design for custom chips and I/O, something that wasn't possible for C= back in the early 1990ies and it wouldn't had made much sense without improvements in GFX resolution and quality.
Back in 1990 you should have asked: What do I need to run Doom (or Wolfenstein) and speed improvements on simple GFX Write operations where the key elements.Last edited by OneTimer1 on 21-Sep-2024 at 10:00 PM.
|
|
Status: Offline |
|
|
matthey
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 22-Sep-2024 0:21:09
| | [ #149 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2007 Posts: 2752
From: Kansas | | |
|
| OneTimer1 Quote:
I don't criticize the design of your Warp60, but it wouldn't had been a solution for a cheap A500/A1200 update back in the early 90ies.
I can't even blame the designers of this accelerator, the 68060 CPU is much to expensive for the performance it provides, it passes the limits of sanity where you should ask yourself if expanding a 14MHz 16/32 bit design with a 50Mhz 32 Bit CPU, tuned for 32 burst access that can't be provided by the architecture of the host system, normally you would say "No let's update to a more performant system"
|
The 68060 was cheap and had low cost features like 32-bit data bus (fewer pins and cheaper 32-bit memory), market leading code density (improved instruction cache efficiency, reduced memory traffic & reduced memory requirements) and low power features. The 68060 offered good value in 1994.
CPU | pipeline | transistors | DMIPS/MHz | power@66MHz | price (1000s) 68040 6-stage 1,200,000 1.1 N/A $218 68060 8-stage 2,500,000 1.8 4.9W $308 PPC601 4-stage 2,800,000 ? 9W $370 P54C 5-stage 3,300,000 1.38 14W $995
Motorola Introduces Heir to 68000 Line (Microprocessor Report) https://websrv.cecs.uci.edu/~papers/mpr/MPR/ARTICLES/080502.pdf
Power watts@66MHz taken from the 68060 User's manual and the following article which has the Pentium at $950 and PPC601 more expensive at $450.
https://preserve.mactech.com/articles/mactech/Vol.10/10.02/ThePowerPC/index.html
The following is a 3rd reference with PPC601 power and pricing info.
Motorola Begins Sampling PowerPC 601 (Microprocessor Report) https://www.cecs.uci.edu/~papers/mpr/MPR/ARTICLES/070602.pdf
With all the low cost and low power 68060 features, the integer performance efficiency (performance/MHz) of the 68060 was the best of the above "desktop" CPUs. The 68060 would have the highest integer DMIPS performance at the same clock speed as the competition. The 68060 is the lowest power and would have the best power efficiency (performance/W) at the same clock speed as the competition. The 68060 would have the best price efficiency (performance/$) at the same clock speed as the competition. The 68060 has the deepest pipeline for clock speed increases of the competition. By limiting the 68060 clock speed to 50MHz, it only wins in performance efficiency instead of everything integer performance, power and price related. The Pentium and PPC601 likely have better floating point performance at the same clock speed but the deep 68060 pipeline should have allowed it to out clock the competition and remain competitive. Do you see why the 68060 could not be clocked up when the 68060@66MHz was already in testing and production announced for 4Q94 in the first article above?
The PPC601 design was award winning after all.

https://en.wikipedia.org/wiki/PowerPC_600#Design Quote:
IBM was the sole manufacturer of the 601 and 601+ microprocessors in its Burlington, Vermont and East Fishkill, New York production facilities. The 601 used the IBM CMOS-4s process and the 601+ used the IBM CMOS-5x process. An extremely small number of these 601 and 601+ processors were relabeled with Motorola logos and part numbers and distributed through Motorola. These facts are somewhat obscured given there are various pictures of the "Motorola MPC601", particularly one specific case of masterful Motorola marketing where the 601 was named one of Time Magazine's 1994 "Products of the Year" with a Motorola marking.
|
Motorola, Trevor, Ben and ppcamiga1 have shown us the vast superiority of PPC. An in-order 68k CPU could never outperform an OoO PPC CPU with double the caches. The Ministry of Truth has spoken!
OneTimer1 Quote:
A 28MHz 68EC/LC30 by C= might have been a vital option but most things beyond that, needed a new design for custom chips and I/O, something that wasn't possible for C= back in the early 1990ies and it wouldn't had made much sense without improvements in GFX resolution and quality.
Back in 1990 you should have asked: What do I need to run Doom (or Wolfenstein) and speed improvements on simple GFX Write operations where the key elements.
|
The 68060@100MHz seems to work ok with AGA in the videos above. Sure, there are lots of CBM bottlenecks and maybe even some 68k bottlenecks so I defer to the Ministry of Truth. My eyes must be deceiving me. The 68060 could have never been put in a SoC by the late 1990s or a single chip system as CBM called it complete with diagram and pinouts. That doesn't really exist there on page 7 does it?
https://ia801903.us.archive.org/22/items/amiga-chip-set-strategy/Amiga_Chip_Set_Strategy.pdf
AA+ never existed and never could have had documentation early enough to be implemented.
https://ia801803.us.archive.org/31/items/specifications-for-upgraded-aa-aa-based-systems/Specifications_for_Upgraded_AA_AA+_based_Systems.pdf
https://archive.org/download/amiga-low-end-chipset-aa-roadmap/Amiga_Low_End_Chipset_AA+_Roadmap.pdf
Of course the 68k Amiga never stood a chance and was a complete failure. CBM never could have realized the Amiga weaknesses like lack of chunky soon enough to enhance the chipset. The Ranger chipset never existed either.
https://archive.org/download/commodorestrategicplan19851987/Commodore_Strategic_Plan_1985-1987.pdf
Only the 68k Amiga makes it impossible. Only Trevor, Ben and ppcamiga1 make it possible.
Last edited by matthey on 22-Sep-2024 at 03:09 PM. Last edited by matthey on 22-Sep-2024 at 12:36 AM. Last edited by matthey on 22-Sep-2024 at 12:30 AM. Last edited by matthey on 22-Sep-2024 at 12:23 AM.
|
|
Status: Offline |
|
|
cdimauro
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 22-Sep-2024 6:21:22
| | [ #150 ] |
|
|
 |
Elite Member  |
Joined: 29-Oct-2012 Posts: 4438
From: Germany | | |
|
| @kolla
Quote:
kolla wrote: @cdimauro
Quote:
Commodore's primarily (SIC!) market |
If Commodore had solely focused on what you consider its primary market, then we would not even be here. |
That's the reason why it survived at least 'til 94: with your proposals it would have broken well before... |
|
Status: Offline |
|
|
ppcamiga1
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 22-Sep-2024 15:46:58
| | [ #151 ] |
|
|
 |
Super Member  |
Joined: 23-Aug-2015 Posts: 1017
From: Unknown | | |
|
| @matthey
nobody stop you from making 68k or any other cpu Amiga clone just stop trolling and start working on mui clone on aros
|
|
Status: Offline |
|
|
cdimauro
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 22-Sep-2024 16:04:24
| | [ #152 ] |
|
|
 |
Elite Member  |
Joined: 29-Oct-2012 Posts: 4438
From: Germany | | |
|
| |
Status: Offline |
|
|
WolfToTheMoon
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 22-Sep-2024 17:55:49
| | [ #153 ] |
|
|
 |
Super Member  |
Joined: 2-Sep-2010 Posts: 1410
From: CRO | | |
|
| @matthey
Quote:
It wasn't, really. Especially not in amiga market post commodore, where you couldn't hope to sell many units. _________________
|
|
Status: Offline |
|
|
matthey
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 0:04:23
| | [ #154 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2007 Posts: 2752
From: Kansas | | |
|
| matthey Quote:
The 68060 was cheap and had low cost features like 32-bit data bus (fewer pins and cheaper 32-bit memory), market leading code density (improved instruction cache efficiency, reduced memory traffic & reduced memory requirements) and low power features. The 68060 offered good value in 1994.
CPU | pipeline | transistors | DMIPS/MHz | power@66MHz | price (1000s) 68040 6-stage 1,200,000 1.1 N/A $218 68060 8-stage 2,500,000 1.8 4.9W $308 PPC601 4-stage 2,800,000 ? 9W $370 P54C 5-stage 3,300,000 1.38 14W $995
|
WolfToTheMoon Quote:
It wasn't, really. Especially not in amiga market post commodore, where you couldn't hope to sell many units.
|
Integration and economies of scale are required to produce cheap hardware of course. Jay Miner understood the importance of integration which made the Amiga possible and so did CBM even though their failure to further integrate the Amiga as planned led to their demise. See the last page of the following document for CBM integration benefits.
https://archive.org/download/new-amiga-systems-and-chipset-requirements/New_Amiga_Systems_And_Chipset_Requirements.pdf Quote:
PRODUCT COST EFFECTS OF INTEGRATION
- Piece parts cost of chips - Reduced Glue chips - Reduced Size PC - Reduced complexity of PC o fewer traces o less vias o less device holes - Less PC assembly time - Less PC test time - Higher PC test yields o Less repair time o Less scrap - Less assembly equipment & support - No System level chip screening o test costs less o less handling o less scrap - No sockets - Higher System level yields o less System test time o less scrap - Higher System yields - Less System Repair times - Smaller Power Supply - Less Cooling - Less Shielding - Less EMI Filtering components - Higher Reliability o Less Warranty repair/replace costs . faster repair time . Less spare parts inventory required - fewer parts to control, track, ship & inventory
|
The previous documents I posted above that were ignored as well as the Post Commodore Bankruptcy Documents have different timelines but show Amiga chipset integration before the year 2000 including 68k integration into a SoC. The last Amiga chipset we received was AGA.
AGA Lisa (ECS Denise replacement) CMOS 1500nm 84 pin, 124,500 transistors
AGA Alice (ECS Agnus replacement) NMOS 5000nm? 84 pin, 39,500 transistors
AGA Paula (ECS Paula not upgraded) NMOS 5000nm 48 pin, 21,857 transistors
AGA Akiko (integration of CIA chips, I/O and glue) ? 160 pin, ~13,000 gates (not transistors?)
The next chipset would have been AA+ and should have been available in working systems before CBM went bankrupt according to the time schedule of CBM documents. The Alice and Paula chip were integrated together.
AA+ Belle (AGA Lisa replacement) CMOS 1000nm 84 pin, 124,500+ transistors
AA+ Ariel (AGA Alice+Paula replacement) CMOS 1000nm 84 pin, ~100,000 transistors (Alice+Paula=61,357 transistors but CMOS uses more than NMOS)
AA+ ? (AGA Akiko replacement with integration of CIA chips, I/O and glue) CMOS 1000nm? 160 pin, ? transistors
Enhancements include "4/8 bit chunky pixels, a 16/24 bit true-color mode" a HAM replacement, 57.2MHz pixel clock for 800x600@72Hz non-interlaced, blitter enhancements, 16-bit audio, high density floppy support, buffered serial, etc. AA+ Basically should have the good stuff that the rushed AGA lacked. The conversion to a more modular less proprietary, low power CMOS chipset was a huge advantage and should have clocked to 57MHz. The full CMOS chipset is sometimes referred to as the "Laptop Chipset" in CBM docs for "Laptop/Palmtop" applications. Further chipsets would include integration of Belle and Ariel and then the CIA/I/O/glue chip as a "Single Chip System".
Single Chip System CMOS ?nm 208 pin, ? transistors
AA+ was scheduled to arrive in 1994 according to the Commodore Post Bankruptcy docs but an external commodity 68030 couldn't be used to clock to 57MHz severely limiting both the CPU and chipset to 28MHz. The integrated 68k+chipset SoC was scheduled for early 1995 and likely would have been a 68EC030 which Motorola had at 1000nm already. I'm not sure 57MHz yields would have been good enough though or Motorola would have been selling higher clocked 68030s. CBM could have moved to a smaller fab size but the 3-stage pipeline limits how much the CPU could be clocked up without expensive die shrinks. It may be necessary to use a 500nm chip fab size to get acceptable yields.
68EC030 Amiga SoC CMOS 500nm ? pin, 68EC030: 251,000 transistors chipset: ~500,000? transistors
The 68060 was already prepared for 500nm (original 68060 fab size), it has low power features, it has much better performance and it has an 8-stage pipeline that not only should allow better yields at 57MHz operation but may allow a CPU and chipset upgrade to 114MHz without a die shrink.
68060 Amiga SoC CMOS 500nm ? pin, 68060: 2,530,000 transistors chipset: ~500,000? transistors
Retaining the FPU would allow a 68060@57MHz which could play Quake even on a reasonably priced laptop in the late 1990s. The total transistors for the 68060 Amiga SoC would be less than a P54C Pentium. The Pentium had dropped in price so that it wasn't over three times as expensive as the 68060 and the higher clock speed added value. A 68060@57MHz Amiga SoC may have been cheap enough for mid-priced systems but maybe not for low end systems or consoles, at least without an integrated 3D chipset (memory is the other major expense). A 68060@114MHz would have increased the value and should have been possible with an 8-stage pipeline at 500nm. The Pentium had moved to smaller fab sizes and OoO by the late 1990s but the CPUs were really fat limiting the caches, expensive and power hungry. The tactic to use cheaper and older fab sizes with a deeper pipeline was a competitive advantage used by Cyrix and Exponential Technologies. Intel moved from the P5 Pentium 5-stage pipeline to a much deeper 13-stage pipeline with the 1995 OoO P6 Pentium Pro but the too deep pipeline uses more transistors with now 5.5 million transistors. They retained their smaller in-order pipeline for the masses. Motorola and IBM chose expensive die shrinks down to 250nm in the late 1990s to fend off Exponential Technology's 6-stage x704 CPU that was targeting 400MHz to 533MHz using a cheaper 500nm process. Cost advantages don't always win over deep pockets. A 68060 Amiga SoC wouldn't necessarily have been competing on the desktop or more upscale markets though. It is likely that CBM would have continued to pick up more embedded market business and turned into what the RPi Foundation is today. The 68060 was a success even at only 50MHz in the embedded market as more than a decade long life and several die shrinks shows. It outlived many of the x86 and PPC designs it was not allowed to compete with.
Well, I found that nobody reads my posts making me a troll that wastes my time. Amiga Neverland dystopia means this is not the place for intellectuals so I will do everyone a favor and limit my time here. Good luck with your 20+ year old expensive PPC niche hardware, 68k emulation and AmigaOS on top of Linux distros. My prediction is that it moves the Amiga nowhere. Integration and economies of scale hardware continue to be the way to make hardware competitive but nobody is listening to me anymore than they did to Jay Miner in 1986. RIP Amiga.
|
|
Status: Offline |
|
|
cdimauro
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 4:23:13
| | [ #155 ] |
|
|
 |
Elite Member  |
Joined: 29-Oct-2012 Posts: 4438
From: Germany | | |
|
| @matthey: I greatly appreciate your posts, Matt.
The problem is that we're really too late to think about reviving the platform. Even the A500 Mini hasn't sold so much (I think it was 150k units?) considered that 5.5 million Amigas were sold.
Maybe I'm blind, but I don't see any way. Now.
If someone has a concrete & realistic business plan, I'd like to see it. |
|
Status: Offline |
|
|
ppcamiga1
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 6:20:29
| | [ #156 ] |
|
|
 |
Super Member  |
Joined: 23-Aug-2015 Posts: 1017
From: Unknown | | |
|
| @matthey
all you wrote is worth nothing as long as there is no worth of use amiga like os on x86 and arm stop writing useless posts on ppc and 68k start working on aros especially on zune
|
|
Status: Offline |
|
|
ppcamiga1
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 6:21:08
| | [ #157 ] |
|
|
 |
Super Member  |
Joined: 23-Aug-2015 Posts: 1017
From: Unknown | | |
|
| di mauro stop trolling start working on aros on zune |
|
Status: Offline |
|
|
kolla
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 12:30:19
| | [ #158 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3475
From: Trondheim, Norway | | |
|
| @cdimauro
Quote:
cdimauro wrote: @kolla
That's the reason why it survived at least 'til 94: with your proposals it would have broken well before... |
What have I proposed? I’ve merely pointed out that looking at adding DSPs did make sense at the time, and that it could have been useful for those of us who aren’t solely games focused. This whole alternative reality thing of what they should and shouldn’t have done is all on you.Last edited by kolla on 25-Sep-2024 at 12:49 PM. Last edited by kolla on 25-Sep-2024 at 12:36 PM.
_________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|
OneTimer1
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 22:05:48
| | [ #159 ] |
|
|
 |
Super Member  |
Joined: 3-Aug-2015 Posts: 1258
From: Germany | | |
|
| @matthey
Quote:
matthey wrote:
Quote:
OneTimer1 wrote:
I don't criticize the design of your Warp60, but it wouldn't had been a solution for a cheap A500/A1200 update back in the early 90ies.
|
The 68060 was cheap and had low cost features ...
|
The 68060 wan't available in the early 90ies, according to Wikipedia is wasn't in production before 1994. According to you it cost $308 that would have increased the price of an A1200 about $600, It would have needed FastRAM adding easily $100 again.
Your ideas would have nearly doubled the price of an A1200, that is way off topic. A cheap improved Amiga would have been needed to prevent C= bankruptcy, that happened 1995. AGA would have been needed around 1990 and a compatible improvement around 1992 and not with a CPU that wasn't available.
|
|
Status: Offline |
|
|
kolla
|  |
Re: / Missed opportunitty to improve the Amiga and keeping it cheap Posted on 25-Sep-2024 23:45:43
| | [ #160 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3475
From: Trondheim, Norway | | |
|
| Phase5 CyberStorm cards with 060 showed up in early 1995. (if the manual is to be trusted) Last edited by kolla on 25-Sep-2024 at 11:46 PM.
_________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|