@Rob

Quote:

Rob wrote: @TRIPOS Quote: It has already crashed and burned. AmigaKit has removed the “Hyperion” category in their web shop since long Same with Cloanto. What do Hyperion and Cloanto have in common? What conclusions should we draw from this?

The subject was OS4, not Cloanto. OS4’s situation is problematic, not Cloantos. OS4 is dependent of AeonKit, Cloanto is not.

And really, who cares about AeonKit and their shady dealings? They are not Amiga (though they seems to think so). Cloanto is.

@cdimauro

Quote:

es, cheap means something like that. The lowest cost PowerPC processors that NXP offers can be found on that ballpark. It's quite expensive considered what else (not PowerPC) can be found (more powerful and at much less cost), but here we're talking about a processor to be used on a hobby machine just to run OS4 (as I've said times ago, PowerPCs can be considered as the "dongle key" used by Lightwave).

$100 US before OS purchase is quite a bit for a hobby. I very much doubt you will get new users at that price point. No one would care about a single tasking old OS other than a die-hard.

Quote:

Correct and, as I've said before, it should be considered that this is just a hobby market. The Pi can do much better and costs much less. That's why PiStorm with Michal's Emu68 can be (and I believe that it'll be) a game changer here. See below why.

PiStorm is a marvellous piece of kit.
Should have happened many years ago but it is giving new life to those that have working hardware.
Porting the OS native to that hardware would be a much better idea rather than trying to shoe-horn it into +20-year-old, hard to source, dying hardware bases.

Quote:

Impossible: SMP is certainly out of question. ExecSG leader already regretted from it several years ago (see also the above YT video, which is from 2015).

Now that, I don't agree on. It has to happen. We just need to move away from legacy. AROS have test versions already on 3.x. Sandboxing older requirments is always an option, but yeah, it needs to happen. This is the future, CPUS are not going to go back to single threading.
There once was a time when AMIGA had Transputer compute over may CPUs before it was mainstream.
http://www.bambi-amiga.co.uk/amigahistory/prototypes/transputer.html
https://en.wikipedia.org/wiki/Atari_Transputer_Workstation

However, as we have both recognised, NONE of this is relevant when you can't sell a device to pay for development due to buy-in COST.

Quote:

That's why the userbase is moving towards the "classic" market: it's much more affordable, there are plenty of products (even new ones), and... it's the real (more or less) Amiga. At the very end Amigans like to continue to use the old software which was developed for the Commodore Amiga. Games need a very good compatibility, and here there's no other choice than WinUAE (especially with the upcoming version). The computational power that you need is almost all about the needed emulation accuracy. Only a few 3D and/or RTG games can benefit from additional performances, but WinUAE already gives some ways (immediate blits, higher-end 68K processors, and JIT of course) to have them, just using the remaining power offered by the host processor(s). Applications needs computational power, and this comes using a JIT. Here, as I've said, the game changer will be EMU68, which allows to get very good performances at a ridiculous price

What user base!
There is very little left and if you want to garner interest in the AMIGA OS, it's going to need to harness the hardware, not leave it idle as a nice-to-have. Even Haiku is breaking away from BeOS compatibility in favour of x64, more memory for browsers and better compute.
SMP is a thing, so is higher memory access. These things need to happen. I think you will find less and less of people interested in buying old crap that can't do it when you have WinUAE for games working just fine. No need for ANY hardware to just have a look. It's pointless.

Last edited by SHADES on 05-Oct-2021 at 09:25 PM.

_________________
It's not the question that's the problem, it's the problem that's the question.

@TRIPOS

Quote:

But breaking Amiga compatibility raises the question of why on earth you would want to continue using PPC at all. Of course AeonKit has its main focus on making their OS make the most out of their PPC products, including the Tabor. Their whole purpose seems to be about building HW dongles (PPC motherboards) that locks the community in a dead-end platform. But once you have sacrificed both binary and source compatibility you could as well go all-in on SMP, 64-bit, real memory protection and resource tracking, etc, on any other platform such as X64/AMD64 or AArch64 or whatever. Sticking to PPC at that point would just be… weird!

Totally agree.

_________________
It's not the question that's the problem, it's the problem that's the question.

Yes, it looks like A-Eon was sold a bill of goods in the early days, most likely that it will be easy to port AmigaOS 4.x to any new Power ISA (PowerPC) board.

Given such assurances about the OS, the remaining piece of the puzzle was modern, um... up-to-date, hold on... better than Apple's G4 hardware for the premium offering of the A1 X1000. The launch of which would coincide with the auspicious 25th anniversary of the launch of the original Amiga 1000 in the summer (northern hemisphere) of 2010.

It's also likely that the other bill of goods Hyperion would have sold them was that SMP will also be easy to implement.

Given such assurances about the roadmap for AmigaOS 4.x, it makes perfect sense to equip the flagship A1 X1000 with the 64-bit dual-core PWRficient PA6T-1682M CPU. And also to follow it up with a planned lineup of X3500, X5000/20 and X5000/40 to be fitted with a quad-core P3041 (e500mc) CPU, dual-core P5020 (e5500) CPU, and quad-core P5040 (e5500) CPU, respectively.

Now here we are, almost 10 years after the much delayed launch of the X1000, there is no sign of AmigaOS 4.2 (a license of which was to be included with every purchase of an X1000), there is glacial progress on the A1222, none of the big-ticket ports of open source software have made it over, graphics (one of the Amiga strengths) is still one of the weakest elements in AmigaOS 4.x, and we may get an early glimpse at the possible implementation of some sort of multi-core support in the the kernel.

You know what? If AmigaOS 4 had died (like it should have), things would have been better. Instead we have this zombie OS.

_________________
All the way, with 68k

@TRIPOS

Quote:

But breaking Amiga compatibility raises the question of why on earth you would want to continue using PPC at all.

Because ppc is technology used in Amiga for last 25 years.

Quote:

Their whole purpose seems to be about building HW dongles (PPC motherboards) that locks the community in a dead-end platform. ... Sticking to PPC at that point would just be… weird!

history, memories.
Something that people like you never understand.

@ppcamiga1

Memories... Yes, of failure that PPC Amiga has been.

@Rose

Agreed!

All nostalgia and PPC have given us is, higher expense and an ever smaller trending user base with, practically zero development.
It's a dying shame too.

_________________
It's not the question that's the problem, it's the problem that's the question.

PPC success was limited.
But it is many times faster than classic and still binary and source compatible.
Something that Commodore should made by late 90.
No other made this.

@ppcamiga1

Commodore chose HP PA-RISC for their next gen devices.

_________________

@SHADES

The discussion here is too hardware related in my view. We are dinosaurs from a time when hardware defined what you could do. At that time computer were exotic, to use one at home was something special. The PC were faster typewriters... The situation today is different. Computers are in almost anything. People use devices for special purposes, they mostly not even know what is inside (and BTW they do not care about it). Expecially important are smartphones. People know what the brand is, they know if it is Android or iPhone but I do not think many people know the hardware or what the OS is based on. People today (expecially the younger people) use the devices task orientated. The brand is important but not the processor... And you have plenty of different devices so you need solutions for different devices people could use. Our classical desktops are not suited for that.

In short to be successful today you need new concepts with a new GUI hiding the hardware that makes it easier for people to use it intuitively on different devices. And your solution must be able to integrate different cloud solutions. But that (if done) would not look like amiga anymore.

@WolfToTheMoon

AMIGA Technologies chose PPC in 1995

@ppcamiga1

Which is exactly the point. It was a wrong choice and brought the platform nowhere. Commodore in 6 years sold 7 million units, and was a terrible company.

AT & associates sold ... 70 boards in 20 years ? PPC had its chance, I don't want to criminalize it. Yet, it failed miserably. Chances are, if they opted for Itanium, they would still have sold more

Last edited by IridiumFX on 07-Oct-2021 at 06:55 PM.

@IridiumFX

No one knew that Intel was able to Die shrink Intel chip to extent they did, it speculated that in order to get higher clock speed the transistor count had to go down, this was point of risk, as well as C compilers where unable to take advantage of special instructions. Intel Pentium 6, actually introduce risk.

https://en.wikipedia.org/wiki/Intel_Microcode
https://en.wikipedia.org/wiki/Micro-operation

So at least for people bet on Power had right, in that Risk was good idea, to some extent, what messed up PowerPC, as that it did not go into normal PC, with exception of Apple, instead went into embedded swamp. They where willing to cut arms, and legs off, to save power consumption.

Too many variation of PowerPC, hard to optimize if you can’t trust that instructions are available. Isel, fsel instructions are good for branchless code, but they are not always included, it all depends on powerpc isa version.

The PowerPC instruction’s does not really allow, you mix data inside instruction with exception of offsets, but move data out of instruction, should make code smaller, Instruction cache and data cache, fixed 32bit length good because how memory is organized. PowerPC lives on as Power, while giving up big endines from market pressure, in latest Power chips.

The PowerPC is actually where totterent to misalignment, something other CPU brands are not always, this makes it good when you’re working with legacy 16bit addressing and code.

Last edited by NutsAboutAmiga on 08-Oct-2021 at 07:04 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 09:19 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 09:12 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:25 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:20 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:19 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:14 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:11 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:06 PM.
Last edited by NutsAboutAmiga on 07-Oct-2021 at 07:06 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@NutsAboutAmiga

2006 was the absolute last year when you could still have defended the PPC. Because once Apple called it off, there was simply no bandwidth left for any further development. You can say POWER is the heir to the empire, but it's headed towards little endian and instruction prefixes, just like AMD 64. So why care at all ?

In hindsight, even 2006 was too late. The G4 had been stagnating for ages, the G5 was a bad joke (I had 2 PowerMac G5, including a 970MP one. Sucky till the last iteration. I know what I am talking about). PPC was probably game over as soon as PReP and CHRP failed to materialize. It failed because of many reasons, including never having a market to begin with.

If you think of all the money wasted on PPCs and NGs, you could have obtained a license to the 68K isa and even funded an fpga-to-asic conversion (and still not revived the platform, because it's a retro curiosity. Let's deal with it)

@IridiumFX

most of that large 32bit bitmaps, at high resolutions, sure binaries are fatter, but its not what makes up most of memory usage, binaries are fat on x86/x64 as well.

Also AmigaOS4.1 contains more drivers, contains USB and lot stuff do not exist in memory on AmigaOS3.x, as well JIT compiler, large buffers etc.

Last edited by NutsAboutAmiga on 07-Oct-2021 at 09:37 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@OlafS25

Quote:

by OlafS25 on 7-Oct-2021 23:58:35 @SHADES The discussion here is too hardware related in my view. We are dinosaurs from a time when hardware defined what you could do. At that time computer were exotic, to use one at home was something special. The PC were faster typewriters... The situation today is different. Computers are in almost anything. People use devices for special purposes, they mostly not even know what is inside (and BTW they do not care about it). Expecially important are smartphones. People know what the brand is, they know if it is Android or iPhone but I do not think many people know the hardware or what the OS is based on. People today (expecially the younger people) use the devices task orientated. The brand is important but not the processor... And you have plenty of different devices so you need solutions for different devices people could use. Our classical desktops are not suited for that. In short to be successful today you need new concepts with a new GUI hiding the hardware that makes it easier for people to use it intuitively on different devices. And your solution must be able to integrate different cloud solutions. But that (if done) would not look like amiga anymore.

You seem to have misunderstood my entire point. Here's why.

PPC has done nothing but slow down an already crippled platform from the get go. What I am talking about is getting away from the past, not to re-define it from scratch or build off the previous mistakes like PPC which is if nothing else, cost-prohibitive to develop products or sell them.
As you stated further down, people couldn't give a tinkers-cuss on what the hardware is. If it's expensive, many won't buy it or it stays niche with little support.

Developers, however, they like "standards" hardware platforms that are current, well documented, and remain largely compatible through each iteration. Like OCS->ECS->AGA or Pi2->Pi3->Pi4 core2, i7, etc
If there is a common GPU that can be supported, even better. Sure beats having to develop drivers for a ton of different cards.
Same with BUS architectures, PCIe remains compatible right back to version 1. Version 6 has just been finalised.
What else? Not having to re-design boards every single time there is a change. Pi (GPIO) / AMD on the same footprint up till Zen-3
That's all irrelevant if COST is prohibitive, so there we are again with BUY-IN

Cheap buy-in is the only way to get interest.
It is the only way you will make sales for an OS that is different to what's on offer.

The OS needs to develop on a platform that isn't prohibitive to the curios so that anyone who asks what is Amiga, can purchase it without second guessing their wallet.

iPhones don't run on Android devices or the other way around, regardless if people know what the hardware base is. That's not even remotely true.

I'm sure there are aspects of the OS that can keep the underlying feel and use of AMIGA OS, no matter the platform or, there would have never been OS4 to begin with. Having a cheap platform to develop on, sell and that remains compatible with each revision is a fantastic start.

Sure, brand is important however, AMIGA is hardly a brand anymore so, get the "vision" in order, look at COST and ease for development on a cheap platform and build the Brand. Apple came back from nowhere, AMD from the verge of bankruptcy and poor performance and Amiga can too.

The OS won't improve with high cost of buy-in. It just won't and it needs to.
Hardware is very relevant, to this day. I can't say it enough, buy-in costs, standards.

Last edited by SHADES on 08-Oct-2021 at 12:13 AM.
Last edited by SHADES on 07-Oct-2021 at 11:40 PM.

_________________
It's not the question that's the problem, it's the problem that's the question.

NutsAboutAmiga Quote:

No one knew that Intel was able to Die shrink Intel chip to extent they did, it speculated that in order to get higher clock speed the transistor count had to go down, this was point of risk, as well as C compilers where unable to take advantage of special instructions. Intel Pentium 6, actually introduce risk. https://en.wikipedia.org/wiki/Intel_Microcode https://en.wikipedia.org/wiki/Micro-operation

Higher clock speeds did not come from reduced transistor counts but from pipelining processors (dividing work into stages like an assembly line) where simpler instructions were helpful for this goal. Most existing 68k and x86 instructions were simple enough for 1 cycle throughput which is needed for efficient pipelining. Early RISC designs often went further to simplify instructions and eliminate microcode saving gates in the core (execution pipeline) but then higher clocked cores needed deeper pipelines, more GP registers, more ICaches for larger code and increased number of instructions and OoO to execute the increased number of instructions which all required more gates. Weak RISC cores have a transistor advantage over simple CISC cores but it mostly disappears when performance is desired and where CISC has a performance advantage.

Some RISC design principals were already utilized in the 68040 and 486. The 68060 and Pentium leveraged most of the advantages of RISC designs.

M68060 User's Manual Quote:

Leveraging many of the same performance enhancements used by RISC designs as well as providing innovative architectural techniques, the MC68060 harnesses new levels of performance for the M68000 family. Incorporating 2.5 million transistors on a single piece of silicon, the MC68060 employs a deep pipeline, dual issue superscalar execution, a branch cache, a high-performance floating-point unit (MC68060 only), eight Kbytes each of on-chip instruction and data caches, and dual on-chip demand paging MMUs (MC68060 and MC68LC060 only). The MC68060 allows simultaneous execution of two integer instructions (or an integer and a float instruction) and one branch instruction during each clock.

While micro-oped cores use RISC pipelines and simple instructions, breaking down instructions into micro ops, micro op caching, OoO execution and microcode use violates traditional RISC principles. Transistors and power are traded for higher performance. The x86 was one of the first ISAs to use micro-oped core designs because of the high decoding cost which could be commonly avoided with a micro-op cache or trace cache reusing already decoded instructions. RISC instructions should have lower decoding costs if following RISC principles but even ARM finally switched to some high performance micro-oped RISC designs to try to close the performance gap with CISC. ARM is also using OoO even on mid-performance RISC core designs today which used to be in order. ARM had tried to leverage the advantage of many smaller weaker RISC cores which was closer to the RISC philosophy but that was only good for part of the embedded market where fewer active transistors is an advantage (low power draw cores that sleep most of the time).

NutsAboutAmiga Quote:

So at least for people bet on Power had right, in that Risk was good idea, to some extent, what messed up PowerPC, as that it did not go into normal PC, with exception of Apple, instead went into embedded swamp. Where willing to cut arms, and legs off, to save power consumption.

Part of the RISC philosophy was a good idea. Some traditional RISC ideals were taken too far to be practical for high performance cores and were later abandoned. Some CISC performance advantages were realized and added back in. Traditional RISC and CISC are both mostly dead and high performance cores today are hybrid designs.

PPC had some performance advantages compared to other RISC designs like early OoO use, more powerful instructions and more registers. Most of the embedded use was for higher performance embedded applications like communications and automotive. These advantages were also disadvantages when trying to scale down to compete with ARM, along with the poor code density. ARM eventually improved their ISAs for better code density and more powerful instructions, added OoO and made their pipelines deeper so they could be clocked up more. PPC didn't have enough advantages left and castrating already loose PPC standards only made standardization problems worse.

NutsAboutAmiga Quote:

Too many variation of PowerPC, hard to optimize if you can’t trust that instructions are available. Isel, fsel instructions are good for branchless code, but they are not always included, it all depends on powerpc isa version.

Lack of standardization was a problem especially after AArch64 gave ARM the standardization they previously lacked. PPC programmer unfriendly assembly code added to the problem, especially for compiler support and embedded use.

NutsAboutAmiga Quote:

The PowerPC instruction’s does not really allow, you mix data inside instruction with exception of offsets, but move data out of instruction, should make code smaller, Instruction cache and data cache, fixed 32bit length good because how memory is organized. PowerPC lives on as Power, while giving up big endines from market pressure, in latest Power chips.

There are 3 choices of how to handle larger immediates and displacements.

1. Many instructions used to create larger immediates and displacements from smaller ones
+ small instructions
+ ICache is highly predictable
- traditional RISC fetch and instruction execution bottleneck as more instructions need to be executed
- additional instructions are mostly dependent (OoO helps at a cost)
- more registers required
- larger code

2. Move larger instructions and displacements to data sections
+ few instructions to fetch and execute
+ smaller code
+ some constant data can be reused by other code
- more stress placed on less predictable DCache (commonly more of a bottleneck than ICache)
- offsets and registers used for accessing data partially offset some of the advantage

3. Use variable length instructions to contain larger immediates and displacements
+ few instructions to fetch and execute
+ ICache is predictable
+ smaller code
+ compression of data can sometimes be used to further reduce code size
- requires a variable length instruction encoding with increased decoding overhead
- requires handling of large instructions and average instruction size increased

Most RISC cores have to use a combination of 1 and 2 above. Even ARM Thumb2 and RISC-V which use a variable length instruction encoding don't take advantage of 3 above even though the short 16 bit encodings at least reduce the code size. It is possible for a RISC ISA to take advantage which Mitch Alsup was working on. The 68k and x86(-64) take advantage.

NutsAboutAmiga Quote:

The PowerPC is actually where totterent to misalignment, something other CPU brands are not always, this makes it good when you’re working with legacy 16bit addressing and code.

I believe misaligned memory handling hardware is optional although most PPC processors had it for big endian mode. Many newer RISC ISAs support hardware misalignment. That would have been an advantage for PPC in the time of PA-RISC, MIPS and even early ARM but it is a mostly standard feature today.

Last edited by matthey on 08-Oct-2021 at 04:04 PM.

I know Im going on about Emulation alot, but there is one big positive to AmigaOS4 not progress as much as it has, getting SMP support etc, means its going to be a lot easier to emulate!

interesting little read and thread I found on a Mac site,
ARM faster for PPC Emulation
https://www.emaculation.com/forum/viewtopic.php?t=10702

Who knows in a few years time we might see Pistorm PPC Edition.

_________________
AmigaNG, YouTube, LeaveReality Studio

@ppcamiga1

if you ask people about amiga (if they remember at all) they will answer f.e. A500 not PPC. PPC was topic in mid 90s and MorphOS/AmigaOS became real about 2002 and later

@amigang

Interesting, but the FPU score is missing, this thing is lacking something,

the M1 score of 400%, is like 4 x 300mhz = 1.2Ghz that is like the score of Sam460 1.2 Ghz, it’s pretty good. But not close to X1000 or X5000 computers.

Last edited by NutsAboutAmiga on 08-Oct-2021 at 05:13 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS