@amigang

Quote:

Then why are you on Amiga forum? there must a be a reason you still use Amiga, it could be just nostalgia, but that's a powerful market. The biggest Amiga commercial venture in a while, the A500 mini, oh look at that its a Arm based computer running emulation. I guess a lot of people are going to waste there time on it, plus on all them other mini computers that had success, mini nes, mini ps1, min c64, all time wasted.

You guys still seem to not understand the purpose of @ppcamiga1. He is here to bring nonsense arguments, to drag discussion into nowhere or to initiate always one of his two favourite topics: number one being "amiga gui on unix" and number two "aros is shit".

@ppcamiga1 is known on Polish amiga forums as "swinkamorska". He is ignored by large amount of people there and gets regularly banned on at least one forum. He is pure troll and that is his only attitude here - trolling.

Therefore, please again, do not feed the troll... If you like pointless discussion, comparing apples to oranges and running in circle over the same topics go on, but it will not lead you to anything.

Much more effective would be to put up "It's time to join the forces" again ;)

@ppcamiga1

Quote:

Amiga gui and graphics on top of unix, Something like Mac Os X is the only way forward.

@deadwood's AxRuntime project seems like a good start on this.

_________________
"Unix is supposed to fix that." -- Jay Miner

@bison

It is still not compatible with mui from 1997.

@michalsc

you still fool himself and think that people will be wasting time on crap like aros x86 and/or emu68k.
it is not goig to happend.
if something can be done many times faster on commodity os on exactly the same computer
it will be done on commodity os.

amigang Quote:

Like I pointed out the mini retro console market has seen some success and created new fans of retro consoles, mini c64 was successful enough to get the full size one made, new fans, renewed interest and new games have come out as a result.

The mini consoles are cheap enough and look good enough for impulse buying including as presents. Potential customers often read reviews, know the flaws and buy the recreations anyway. It's not that they don't want something better but many of the recreations are cheap quick and dirty hardware using emulation and with cheapness limitations. Ask Trevor how the THEC64 emulation compares to the MiSTer simulation. Which do we want for an Amiga remake?

amigang Quote:

How it’s box’s, runs and presented is a big part because one big factor you forget for maybe 90% of the people who buy these type of computers they don’t care it’s emulated, as long as they can play the classic games it’s good enough. With out watching a teardown video you have no idea what the specs are, people don’t really care in this market. Mean do you know the specs difference of mini nes and mini ps1 no, just care if it ran the games. By they way they are all powered by Arm in emulation.

Some of the 8 bit hardware emulation has been adequate using 1000 times the original performance. In contrast, higher end hardware like the the PS1 mini received consistently poor reviews.

https://www.engadget.com/2018-12-12-sony-playstation-classic-review.html Quote:

But the emulation isn't always consistent. Ridge Racer Type 4 has significant frame-rate dips and jitteriness (or judder, in video-geek-speak) that makes playing the game a huge headache. That's a particular shame, since the original was a graphical milestone for the 32-bit era. There's no longer any sense of speed, and I had a hard time steering without analog controls. I'm surprised Namco's developers haven't complained that Sony's slipshod development turned one of the highlights of the PlayStation into an unplayable mess.

Look at the comments of articles and reviews too. Potential customers often know about the emulation and deficiencies but can't do much about it besides not buying it and some do choose alternatives. Better hardware with fewer limitations is possible for the price as the Raspberry Pi shows. Many people would be better off sticking a Raspberry Pi in a retro case of their choosing. The Raspberry Pi is not retro gaming oriented and does not have FPGA capabilities though. There is an opportunity to create 68k based hardware which can do better and appeal to 68k retro enthusiasts for the Amiga, Atari ST, Sega Genesis/Mega Drive, NeoGeo and x68000 which have a combined huge game library but are old enough that the chipsets do not require large and expensive FPGAs. Emulation requires more expensive hardware as resources are not used as efficiently but emulation based hardware is cheaper and quicker to develop and bring to market. I believe there is a market for higher quality mass produced "retro meets modern" hardware with fewer limitations and which should be possible to produce for a competitive price.

amigang Quote:

Again like I pointed out the tasks I think Amiga should promote are not the things that get really benchmarked, I 100% could make a pixel art work in Dpaint on my pi400 as fast as I could make it in Gpaint or what ever linux program you gave me, actually I’d be quicker with Dpaint most likely as I know the program more. That’s kinda my point.

Let's say someone tries the emulated Amiga environment and likes it. What are they going to ask?

A. How can I switch more of my computing to this emulated environment?
B. Where can I buy affordable Amiga hardware?

With emulation, the host hardware and software is more likely to be seen as superior and not the emulated environment. A good example was the 68k Amiga emulating the 68k Macintosh. The 68k Mac had some quality software and being able to use this software was viewed as an advantage of the Amiga and not the Mac. I don't expect many people switched to the Macintosh as a result, at least until the demise of the Amiga. This was even best case emulation as the CPU was the same so performance was often as good and sometimes better than a real Mac. Realistically, my opinion is that an emulation only Amiga looking for a host is a dead Amiga. I can't see any software only solution which would make the Amiga relevant again.

@matthey

I willing to admit emulation is not the perfect solution, but it is a market and an area I feel Amiga or Cloanto has massively under-utilised.

1. Mini console market
-finally we are seeing a mini Amiga being made, I personally think it’s not ambitious enough, for very low extra costs they could of maybe built the machine with wifi, added a online store where Cloanto could of sold extra Amiga games, they could of opened this up to allow third party developers to sell there stuff, that way it might of encourage new Amiga game development and open up a new money steam for both Cloanto and the community. It should also be promoting the serious side of Amiga and show of that be used to make stuff on, thats it more than just a gaming system.

2. Steam / console game packs
Sega has re-release its mega drive classic pack and sonic the hedgehog on I think every god dam platform. Sonic it self on android has 10 million downloads, some I guess paid £2 to own it with out adverts or some put up with adverts, anyway it’s another money stream. Why hasn’t the Amiga got the same? Again it’s a money stream for Cloanto to collect from and then re-invest in improving its own software or getting more classic games rights they could sell.

3. Games / software store added to Amiga forever
Same as before having a built in store say like steam where it be easy for the end user to buy new games for the emulation again might bring in new money source.

4. Port to more platforms
Add to this Amiga forever the complete pack should be on a lot more systems, not just windows and a poor mans linux / Mac version. It should be on Android, steam dex, iOS (if Apple allowed) web based version etc.

These are all markets that could grow some small new interest for Amiga, i not saying it be a huge market and the masses will come over to use Amiga but we may pick up a few new fans. It’s just one direction Amiga could go into all I feel at very low cost and little development needed.

Last edited by amigang on 22-Nov-2021 at 09:44 AM.

_________________
AmigaNG, YouTube, LeaveReality Studio

A curious thing I find with this poll is that the first two options are about hardware, without any context regarding the software. The the third option is focused on an OS without much context to the overall operating environment and the hardware.

When the context is omitted, it's no wonder we get responses which are narrowly focused on parts of a whole. It's counter productive.

There are many ways to bring back the fun that's absent in today's personal computing landscape. Yes, most of them will require a lot of funding if they are to reach even a user base of 1M, and No, an Amiga experience layer on top of unix/linux (kernel) is not the "only" way.

The "back to our roots" approach is a legitimate one. I think there is merit in updating the prematurely abandoned 68k architecture, and building a niche around a hobby board platform is a workable approach, but the full package will require an operating environment that has an OS that is better than AROS 68k and some new development tools, and a marketplace for hardware and software modules. The 1GHz 68k ASIC board is not the end goal, it's just the starting piece.

A complete break from the the PPC legacy Amiga in a new system which lives atop a unix/linux kernel a la Mac OS X is also a workable avenue. But that's just the beginning. To move more people than just us nostalgic few, it is going to require the other things Apple brought to OS X to round out the operating environment, such as modern APIs and an SDK, and some killer apps that showcase why using them should be seriously considered, and also what can be achieved in a short space of time.

The main point worth driving is that we shouldn't get hung up on a single technology or approach.
Emulation on its own will not do anything, but included into a complete packaged offering it can certainly play a key role. At the very least it can be an essential part of a "Dev Kit" while the real thing is being worked on.
A new x86 or ARM board on its own will not grow a user base, but with an intuitive operating environment and some use-case context, it could deliver the goods.

Productizing is not an exact science, and there are no guarantees, but that doesn't mean that it should not be attempted.

_________________
All the way, with 68k

@agami

Agreed.

People seem to forget who own the Amiga brand now, Cloanto, a company that let’s face it has made nearly all its money from the emulation market.

I also remember when winuae first got ppc support, os4 classic sold out everywhere.

Emulation is not a perfect market but it’s an market never the less.

Last edited by amigang on 22-Nov-2021 at 07:34 AM.

_________________
AmigaNG, YouTube, LeaveReality Studio

@agami

68k was abandoned because it was too slow.
After so many years it is still too slow.
1 GHz 68k ASIC will be nice but probably never happend.
Amiga community have to accept that "classic" Amiga may run on any big endian cpu.
Which means no x86 no arm no risc-v.
x86 or arm boards are waste of time.
They will never be fast and modern.
Amiga gui and graphics on top of unix are only way forward.
Time wasted for arm boards should be used to made common graphics for ppc and 68k.
It may be fpga and have ps1 level performance but it should be cheap.

@ppcamiga1

Quote:

Time wasted for arm boards should be used to made common graphics for ppc and 68k.

Noone is wasting time on ARM boards, and common graphics for ppc and 68k is already what IComp and ThoR are doing - your precious chunky modes.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@ppcamiga1

What do you expect Amiga to be nowadays if anything but a retro machine? 68k is perfectly suited fo Amiga, emu68 gives it speeds way beyond what's needed, what do you want Amiga to run, 3D stuff? Get real

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@pixie

Please, do not feed the troll ;) It brings nothing ;)

@kolla

people waste time on something like buffy or pistorm.
there is no cheap 3d graphics for 68k

@pixie

this crap emu68k is emulator not real thing
it is not retro

To answer the original question, what I would accept in the sense of a dream Amiga, is now very difficult. At one time I was a fan of PPC NG Amigas. Yet, the biggest developments of late for both hardware and software are 68k (actual 68k, FPGU 68k, ARM emulated 68k, and good ol' rabbit hole UAE implementations). To see specs of a 68k Amiga get over 1000 MIPS is very impressive. As the technology improves, we may even see 68k Amiga technology surpass the SAM460/A1222+ in speed. AmigaOS3.2 keeps adding features from 4.x. Who knows, 68k may even get a GL library.

With all of that said, I do have some PPC specific software that I would hate to fall into disuse. There is also a great many PC software packages I use that do not run on the Amiga (no ports). So, my current dream machine would be a type of UAE that has in addition to good classic emulation, a very fast PPC emulation capable of SAM 460 speeds, running on cheap PC hardware, allowing rabbit hole for running a decent browser and other software packages.

My other dream is an actual 68k based Amiga fast enough and graphically capable enough to do all of the above.

_________________
I sold my SAM460ex lite... waiting for money to buy a Raspberry Pi... or a Classic A1000 with Buffee... or an A1222... and OS4.3 FE update 11

@michalsc

Message understood, ppcamiga1 does seem really locked in on his vision and no other can succeed, i give him some credit for his passion on his vision, it may be the right way to go, this is all fun dreaming and speculation anyway.

Last edited by amigang on 22-Nov-2021 at 09:22 PM.

_________________
AmigaNG, YouTube, LeaveReality Studio

PhantomInterrogative Quote:

To answer the original question, what I would accept in the sense of a dream Amiga, is now very difficult. At one time I was a fan of PPC NG Amigas. Yet, the biggest developments of late for both hardware and software are 68k (actual 68k, FPGU 68k, ARM emulated 68k, and good ol' rabbit hole UAE implementations). To see specs of a 68k Amiga get over 1000 MIPS is very impressive. As the technology improves, we may even see 68k Amiga technology surpass the SAM460/A1222+ in speed. AmigaOS3.2 keeps adding features from 4.x. Who knows, 68k may even get a GL library.

I don't consider 1000 DMIPS for a modern 68k CPU to be impressive. Motorola claimed "over 100 MIPS at 66 MHz" for the 1994 released 68060 which is 1.52 DMIPS/MHz. A 1 GHz 68060 would have 1520 DMIPS and a 68060 would reach 1000 DMIPS at 658 MHz. The 68060 was strong in DMIPS/MHz for the day but Motorola chose not to clock it up even though the 8 stage pipeline made it a better candidate to clock up than the early Pentium which had 5 stages up until the Pentium MMX with 6 stages in 1997. Motorola could have exploited the advantage by clocking up the 68060 but the 68060@50MHz was already embarrassing the PPC 603 in DMIPS/MHz and the shallow pipeline PPC processor designs were difficult to clocked up. Apple was not happy that the PPC CPUs did not clock up as fast as the Pentium but. ironically, they abandoned the 68k which ended up with the 68060 which was the best candidate to clock up.

The 68k Amiga has full OpenGL support (not OpenGL ES) through StormMesa including hardware accelerated rendering with Warp3D and software rendering. It is a mostly outdated version of OpenGL not using modern GPU shaders but works ok with old OpenGL code. There are severe optimization issues in the 68k Warp3D libraries and StormMesa was a huge project which is difficult to fully debug and optimize.

PhantomInterrogative Quote:

My other dream is an actual 68k based Amiga fast enough and graphically capable enough to do all of the above.

Many people on here still think RISC has better performance potential than the 68k. The 68k runs circles around RISC cores when using memory.

addq.l #1,(a0) ; 1 cycle, 2 bytes

The 68060 is 1 cycle with 2 bytes of code while the classic 5 stage RISC pipeline with result bypass/forwarding takes 4 cycle and 12 bytes of code.

load r3,(r4) ; 1 cycle, 4 bytes
bubble ; 1 cycle
add r3,r3,1 ; 1 cycle, 4 bytes
store r3,(r4) ; 1 cycle, 4 bytes

The 68060 pipelines not only register accesses like the classic RISC pipeline but DCache accesses which execute with 1 cycle throughput on L1 DCache hits (modern x86-64 cores have 90-95% hit rates on L1 DCache). The 68060 could only have one DCache access per cycle while modern x86-64 CPUs dual port the L1 DCache to further exploit the advantage allowing multiple instructions to execute together more often (existing 68k code without superscalar scheduling would benefit the most). The x86-64 is a memory monster and the 68k can be too.

The 68060 and x86-64 cores can use wider fixed length instructions with larger immediates and displacements after the instruction buffer and for the execution pipelines than is practical for RISC.

add.l #$76543210,(a0) ; 1 cycle, 6 bytes

RISC code would add an immediate load of the upper bits followed by a shift and or instruction taking 6 cycles and using 24 bytes (instruction scheduling can fill the bubble at least). The 68k and x86-64 have more complex addressing modes than RISC which can often execute in 1 cycle while RISC tacks on more instructions to support them. RISC with 32 GP registers and 3 op instructions may execute fewer instructions when all the data is already loaded in registers of deep functions but the 68k still has an advantage with large immediates used on GP registers and generally has smaller code. RISC has to clock up the core, use aggressive OoO execution and/or use instruction fusion to try to execute the many additional instructions with dependencies and bubbles while x86-64 remains the memory munching monster and king of performance despite its flaws. The 68k has flaws too but it has some great traits and a cleaner ISA than x86(-64). Too bad Motorola jumped the fence castrating itself in the process and finding the grass wasn't greener on the other side anyway.

Last edited by matthey on 23-Nov-2021 at 04:04 PM.

@matthey

After all 68k is still as slow as it was in 1994.
Even gvb not change this.

@amigang

Our community has better emulator than emu68k for more than twenty one years.
JIT for UAE is avaible since august year 2000. It results in nothing.
Twenty one years of nothing.
Repeating old mistakes is simple stupid.
I expect change to x86/arm to something fast and modern.

This is kind of an interesting question, as I have multiple "Versions" of Amiga. I have the classic A1200, still the best version in my opinion. I have a NG Amiga with AmigaOne X5000 which seems familiar, but not as cool as the 68K versions. Then I also have a great fast Amiga clone/emulator with Raspberry PI 400 and Chris's awesome work on Pimiga 1.5. I'm not sure I would need a "Next" Amiga, but more of a refinement of what's already out there and much more 3rd party support for hardware, new games and more productivity software. Unfortunately, I'm mostly using my Amigas for past games and software tinkering. I'm not very good with the NG X5000, so I use it periodically. And I know, I can hear some of you now - Use it more! I'm trying to find stuff on it, via the CDs from Amiga Future to enjoy. My biggest issue is I don't really know how to use it fully. Anyhow, this is all just my 2 cents. I'm always happy to see more come out for the Amiga and try to support any Kickstarters that help our hobby become more influential with the masses and us happier. Keep the dream alive all.


Ed

_________________
Amiga 1200 - ACA 1233 68030 128MB Ram
8GB CF With tons of Classics

AmigaOne X5000

Raspberry PI 400 - PiMiga 1.5
"That which doesn't kill you, only makes you stronger" - Someone important, but I forgot who