Click Here
home features news forums classifieds faqs links search
6071 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
» Home
» Features
» News
» Forums
» Classifieds
» Links
» Downloads
Extras
» OS4 Zone
» IRC Network
» AmigaWorld Radio
» Newsfeed
» Top Members
» Amiga Dealers
Information
» About Us
» FAQs
» Advertise
» Polls
» Terms of Service
» Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
9 crawler(s) on-line.
 103 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 vox:  5 mins ago
 zipper:  8 mins ago
 kolla:  15 mins ago
 OneTimer1:  29 mins ago
 BigD:  1 hr 12 mins ago
 OlafS25:  1 hr 14 mins ago
 amigakit:  1 hr 42 mins ago
 NutsAboutAmiga:  1 hr 45 mins ago
 Gunnar:  2 hrs 23 mins ago
 Frank:  2 hrs 30 mins ago

/  Forum Index
   /  Amiga General Chat
      /  Commodore Amiga Global Alliance
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 Next Page )
Poll : Commodore Amiga Global Alliance
Yes, I would Join! £30
Yes, for less
Maybe
No
Bad idea, I have a better one....
Pancakes!
 
PosterThread
OlafS25 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 8:34:02
#101 ]
Elite Member
Joined: 12-May-2010
Posts: 6321
From: Unknown

@BigD

PPC was the official route after the bankruptsy of Commodore. At that time many users already left. Most of them can only remember 68k based hardware (mostly A500). From emotions and memory PPC is as good or bad as X86/AMD64 or ARM for most people., Just another hardware.

Last edited by OlafS25 on 05-Sep-2022 at 08:34 AM.

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 11:04:42
#102 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3621
From: Germany

@BigD

Quote:

BigD wrote:
@cdimauro

Quote:
So, there cannot be a PPC Amiga failure simply because there was/is no PowerPC Amiga.


What are Blizzard PPC/Cyberstorm PPC equipped 'Classic' Amigas then?

They are Amigas with ADDED some coprocessors or replacements.

So, they are still Amigas, since they keep the same chipset, processor, and even o.s., even if they are not using them or part of them.

But the same happened with the Amigas where there was ADDED the InMOS's Transputers, to give the most notable example. Or DSPs, etc.
Quote:
They ARE Commodore/Escom machines that run PPC programs!

Sure, and? See above.
Quote:
Heck PPC even became the official roadmap for upgrading the platform (after some 180 degree turns etc).

PowerPCs were ALSO (read: there were OTHERS as well) a processors family for the NEW Amiga platform.

BTW, Windows NT was ALSO a future base for the NEW Amiga platform. Did you know it?
Quote:
So unless you're hung up about the AmigaOne branding I can't really follow what you mean!

I've absolutely NO problems with AmigaOne branding since they are DIFFERENT machines. So, NOT Amiga.
Quote:
The fact that 90% of the user base does not want to run PowerPC software is another thing entirely!

And that's another thing, right?

I'm talking about Amigas and what I've said applies. As well as AmigaOnes are different machines and this applies as well.

It's "just" a matter to understand it and do NOT confuse the two different things.

 Status: Offline
Profile     Report this post  
BigD 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 14:06:34
#103 ]
Elite Member
Joined: 11-Aug-2005
Posts: 7307
From: UK

@cdimauro

I think it's far easier (and accurate) to view it as support for 68k winding down and people pushing for PPC instead with some limited custom chip emulation (The BoXer would have had custom chip hardware too). This did not bring the majority of the user base along for the ride but that's ok because there's been a 68k resurgence through FPGA AND emulation. However, both are technically branches of 'Amiga' development.

It is true that IP disputes have muddied the waters IMHO. But, the NeXTSTEP/OSX Macs are STILL Macs for example! Same difference!

Last edited by BigD on 05-Sep-2022 at 02:46 PM.
Last edited by BigD on 05-Sep-2022 at 02:07 PM.

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

 Status: Offline
Profile     Report this post  
Karlos 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 14:45:10
#104 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4394
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@BigD

What does it matter? Picking a subset of user mode instructions that work well on 68020 and tuning JIT emulations for those and tying it all to a standardized ABI and set of interfaces, everyone wins. Compile once, run on everything, whether it's a 68060, FPGA, 603, G4, PA Semi, ARM or x64. Performant and mature emulation of 68K exists on all of the above.

Save all the architecture specific features and optimisation for libraries, drivers, datatypes and such.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 15:27:16
#105 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3621
From: Germany

@Karlos: exactly. And at least 68k has the vast majority of software and supported in several ways outside of that ecosystem.

 Status: Offline
Profile     Report this post  
AmigaMac 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 16:00:56
#106 ]
Super Member
Joined: 26-Oct-2002
Posts: 1094
From: 3rd Rock from the Sun!

@amigang

It’s a very good idea and needs to be embraced by all of the Amiga communities across the globe.

_________________

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 16:24:14
#107 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12795
From: Norway

@cdimauro

That was idea behind Java, yet no fully working Java on AmigaOS, then we had AmigaDE/AA/AA2 again only worked on mobile devices, now inventing a new x68 instruction set, fine, we optimize for API calls, and create stack engine pretty easy, that can convert to native code easy. But doing so won’t automatically give you all the 680x0 software, it be a new thing. So not so easy, complicated.

On other hand you can use the 68020 instruction set++, but that includes a lot of flags, Z,N,C,V maybe you don’t wont emulate things on so low level, because its inefficient. But at least its fast on real ASIC (as its native).

Something in between might ideal, that can translate into native code, LLVM?

or maybe open source is best option, compile to native binary for OS of your choice. with few changes here and there.

Last edited by NutsAboutAmiga on 05-Sep-2022 at 05:43 PM.
Last edited by NutsAboutAmiga on 05-Sep-2022 at 05:07 PM.
Last edited by NutsAboutAmiga on 05-Sep-2022 at 04:25 PM.

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

 Status: Offline
Profile     Report this post  
redfox 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 17:01:02
#108 ]
Elite Member
Joined: 7-Mar-2003
Posts: 2064
From: Canada

I think we need some more words of wisdom from MEGA_RJ_MICAL



 Status: Offline
Profile     Report this post  
matthey 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 17:09:01
#109 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

agami Quote:

Ah, the sunk cost fallacy. While it is a shame that ultimately all the work you and others have done related to AmigaOS 4.x is approaching a dead end and objectively amounts to very little, the number of years worked on it are irrelevant.


From investing, I call this "throwing good money after bad". Reality doesn't have to be so harsh. There is little doubt that the PPC AmigaOne venture has been a financial failure. Some of that failed investment may be recovered and reused in other projects. I believe many Amiga users like AmigaOS 4 so it is perhaps not a complete marketing failure and AmigaOS 4 has provided technology advances for Amiga like computers. Ben Hermans recognized that AmigaOS 4 was unprofitable and that he was throwing good money after bad so he stopped development and found a way to reuse the technology and investment in a bigger market with the 68k AmigaOS. Ben is intelligent even though his moral compass and logic of thinking it is better to take advantage of business partners instead of honestly working together with them is flawed. In conclusion, the effort should be recycleable with recompiles on either the 68k or ARM hardware.

agami Quote:

And go forward with 68k.


Yes, the majority of the Amiga market likes and prefers the 68k so it has a marketing tailwind. ARM would be easier and cheaper to develop but further developing the 68k is more compatible and interesting. ARM further divides the Amiga community while the 68k unites it. Real 68k hardware is needed to bring Amiga users and the 68k forward. Emulated 68k CPUs and 68k FPGA CPUs are not good enough. Competitive mass produced hardware is needed "to push that price-performance ratio harder than ever before".

agami Quote:

As is the number of those who are "enjoying" AmigaOS 4.x, It is highly likely that there are an equal or greater amount of individuals that are enjoying:
- Playing a movie on either HD-DVD disc, LaserDisc, SuperVHS, BetMax
- Playing music on either SACD, DCC, MiniDisc, 8-track, reel-to-reel
- Playing UMD-based games on a Sony PSP
- Typing a letter/essay/novel/script on an electric or mechanical typewriter
- Shooting a movie or taking photos using film stock
- Riding a horse to work

And like your enjoyment of AmigaOS 4.x, none of the above listed are credible nor practical technologies for growing adoption in a future context. Some of them are failures, as in they never really reached their potential; and others were just left behind by the ebbs and flows of progress. AmigaOS 4.x and anything on PPC are in both categories.


Putin recently cut off petro exports to much of the EU so some Europeans may go back to riding horses to work. Germany has restarted old power plants all the way back to coal plants and that is unlikely enough to get them through the winter. Green energy was supposed to replace all this. Sometimes older technologies are too good and/or cheap to be easily replaced.

agami Quote:

Perhaps your involvement in the AmigaOS 4.x ecosystem is your swansong: If so, that is perfectly fine, for you. That doesn't mean that AmigaOS4.x/PPC should be the last chapter in the Amiga saga. Certainly not because a small group of developers spent close to two decades in the ecosystem, and you and a handful of others can still find some enjoyment in the platform today.


Retaining valuable developers is important whichever direction is pursued after PPC AmigaOS 4 hardware. I would hope that PPC replacement hardware is at least as powerful as the average AmigaOne hardware.

kolla Quote:

Memory is a much more limiting factor to what can be accomplished than speed is, you should know that.


It is not a memory limitation holding back the 68k Amiga. It is a performance limitation. It is not difficult to add 2GiB of memory to a 68k Amiga. It is not even that expensive as can be seen by the cost of a 2GiB Raspberry Pi 4. There are likely some cases where that much memory would be useful but performance would do more to modernize the 68k Amiga. What would you choose?

A) 125MHz 68060 with 2GiB of memory
B) 250MHz 68060 with 1GiB of memory
C) 500MHz 68060 with 0.50GiB of memory
D) 1GHz 68060 with 0.25GiB of memory

I believe most 68k Amiga users would choose D although some may choose C. I expect A and B would not be popular choices. D would likely open up more old game ports although C or B may be better for productivity software like a modern browser. There is another factor which is the performance of the memory where it may be possible for less memory to be higher performance and more memory benefits more from multilevel caches. This is a hypothetical analysis and both performance and memory are important and have synergies when increased together.

Last edited by matthey on 05-Sep-2022 at 05:12 PM.

 Status: Offline
Profile     Report this post  
kolla 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 17:36:57
#110 ]
Elite Member
Joined: 20-Aug-2003
Posts: 2859
From: Trondheim, Norway

@matthey

I would naturally choose alternative A.

I hate it when build jobs break because of “out of memory”, then I rather go for slower CPU, I don’t mind waiting.

2GB of ram would also mean holding entire animations and audio tracks in ram instead of relying on storage I/O.

Last edited by kolla on 05-Sep-2022 at 05:43 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 17:39:21
#111 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12795
From: Norway

@matthey

Why do you need choice between 4 options, why cant you simply buy the ram you need and plug it in?

Last edited by NutsAboutAmiga on 05-Sep-2022 at 05:40 PM.
Last edited by NutsAboutAmiga on 05-Sep-2022 at 05:39 PM.

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

 Status: Offline
Profile     Report this post  
ferrels 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 17:43:56
#112 ]
Cult Member
Joined: 20-Oct-2005
Posts: 922
From: Arizona

@AmigaMac

Quote:
It’s a very good idea and needs to be embraced by all of the Amiga communities across the globe.


Yeah, all 5 of you will hold hands around the campfire and sing "Michael Rowed the Boat Ashore".

FFS, this is just another delusional thread patronized by the same bunch of crazies who think that the Amiga can be revived. The Amiga is dead, both classic and NG. It's been a retro-hobby even before the term retro-hobby existed....which started when Commodore closed it doors back in the early 1990s. The Amiga offers absolutely nothing that modern consumers and users expect or want from a computer. Believing that the Amiga will make a comeback is as ludicrous as believing that landline phones will make a comeback.

 Status: Offline
Profile     Report this post  
QuikSanz 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 18:57:28
#113 ]
Super Member
Joined: 28-Mar-2003
Posts: 1236
From: Harbor Gateway, Gardena, Ca.

The name Commodore is what puts me off. My A2000 Is Commodore and has parts from all over. My A4000T is an AT, no commodore in there, guess there is no place for me. AGA is good, covers all..

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 19:19:22
#114 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3621
From: Germany

@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote:
@cdimauro

That was idea behind Java, yet no fully working Java on AmigaOS, then we had AmigaDE/AA/AA2 again only worked on mobile devices, now inventing a new x68 instruction set, fine, we optimize for API calls, and create stack engine pretty easy, that can convert to native code easy. But doing so won’t automatically give you all the 680x0 software, it be a new thing. So not so easy, complicated.

On other hand you can use the 68020 instruction set++, but that includes a lot of flags, Z,N,C,V maybe you don’t wont emulate things on so low level, because its inefficient. But at least its fast on real ASIC (as its native).

Something in between might ideal, that can translate into native code, LLVM?

or maybe open source is best option, compile to native binary for OS of your choice. with few changes here and there.

Yes, something like Java and LLVM, but much simpler to implement, because... most of it is already there.

68k is the common ground on Amiga and post-Amiga systems: even PowerPC Amiga o.s.-/like, because they implement a 68k -> PowerPC JIT compiler and 68k applications there are "first class citizens".

So, it's "just" a matter to standardize it to have a common ground, which is then fully implemented on all platforms.

 Status: Offline
Profile     Report this post  
pixie 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 19:49:17
#115 ]
Elite Member
Joined: 10-Mar-2003
Posts: 3115
From: Figueira da Foz - Portugal

@cdimauro

PPC is long dead, why not embrace 68k as the defacto standard?

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

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 20:00:32
#116 ]
Elite Member
Joined: 29-Oct-2012
Posts: 3621
From: Germany

@pixie

Quote:

pixie wrote:
@cdimauro

PPC is long dead, why not embrace 68k as the defacto standard?

That's the point...

 Status: Offline
Profile     Report this post  
matthey 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 21:26:54
#117 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

cdimauro Quote:

Actually it's better to spend the transistor budget to more L1 cache. It would be good to have 32kB x 2 for instructions and data caches. And even better would be to have an L0 cache for some uops, to get rid of the mess/effort of decoding 68k instructions all the time.


There is little doubt that increasing a 68060 like CPU L1 caches to at least 16kiB I+D would be good. Larger is not always better as access latency is increased. There is a balance between L1 access latency and the performance boost from having more data available in L1. Many modern high performance CPUs have found 32kiB I+D L1 caches to be the best compromise but that doesn't mean it is always the case. The 68k extreme code density could allow the I cache size to be decreased while retaining performance compared to other CPUs but it is usually better to benefit from the L1I cache having significantly more code than some RISC like CPUs with poor code density (the 68060 8kiB L1I can contain as much code as a PPC CPU with 32kiB L1I). A 68060 like CPU would most resemble the in order Intel Atom Bonnell architecture although there are significant differences between x86 and 68k.

https://en.wikichip.org/wiki/intel/microarchitectures/bonnell

They chose to use an 8 way 32kiB L1I which is a high capacity L1I cache even with x86 code density but oddly reduced the L1D to 6 way 24kiB, likely for timing purposes. More set associative ways also increases latency so maybe they found 6 way 24kiB L1D was better than 4 way 32kiB L1D.

Bonnell does not break x86 instructions down as far in order to reduce power. This is more like 68060 decoding but not as efficient.

Quote:

Bonnell is a departure from all modern x86 architectures with respect to decoding (including those developed by AMD and VIA and every Intel architecture since P6). Whereas modern architectures transform complex x86 instructions into a more easily digestible µop form, Bonnell does almost no such transformations. The pipeline is tailored to execute regular x86 instructions as single atomic operations consisting of a single destination register and up to three source-registers (typical load-operate-store format). Most instructions actually correspond very closely to the original x86 instructions. This design choice results in lower complexity but at the cost of performance reduction. Bonnell has two identical decoders capable of decoding complex x86 instructions. Being variable length instruction architecture introduces an additional layer of complexity. To assist the decoders, Bonnell implements predecoders that determine instruction boundaries and mark them using a single-bit marker. Two cycles are allocated for predecoding as well as L1 storage. Boundary marks are also stored in the L1 eliminating the need to preform needlessly redundant predecoding. Repeated operations are retrieved pre-marked eliminating two cycles. Bonnel has a 36 KiB L1 instruction cache consisting of 32 KiB instruction cache and 4 KiB instruction boundary mark cache. All instructions (coming from both cache or predecode) must undergo full decode. It's worthwhile noting that Intel states Bonnell is a 16-stage pipeline because for the most part, after a cache hit you'll have 16 stages. This is also true in some cases where the processor can simultaneously decode the next instruction. However, in the cases where you get a miss, it will cost 3 additional stages to catch up and locate the boundary for that instruction for a total of 19 stages.


Rather than a micro-op cache, Bonnell is using more like a mini trace cache by storing decoding data in the L1I. Now we see why such a high capacity L1I is needed and this wastes some of the advantage of good code density which was further wasted when Atom moved to significantly fatter x86-64, as you know. The 68060 has less decoding overhead (16 bit encodings are inherently lower overhead than 8 bit encodings) and maintained competitive high performance without breaking down instructions into uops and requiring extra resources for a muop cache. On top of that, the 68k has better code density and I believe it could maintain much of the code density while moving to 64 bit unlike x86-64.

cdimauro Quote:

eDRAM is too much expensive for a cheap SoC. Better to get rid of it.


That's entirely possible but eDRAM has been used for low cost console SoCs where cost is more important that higher margin desktop SoCs. Moving memory onto the SoC has a performance and power advantage, reduces SoC pins which are far more expensive than transistors and allows to reduce the size of boards. Do you have knowledge of actual licensing costs?

cdimauro Quote:

See above on this: L0+L1 should be enough for this kind of SoC.


As above, I doubt a uop L0I cache is necessary. The write buffer can act like a mini L0 cache and the 68060 has room for improvement here. If using cheap commodity memory, then a L2 cache would certainly be beneficial and practically required for running nearly modern programs.

cdimauro Quote:

No, please. Who's using dual ported VRAM nowadays? Just use the regular DDR memory, which has already enough bandwidth.


Dual porting SRAM is simple and cheap for professionals. The Amiga chipset certainly works with interleaving of memory bandwidth but there still may be some loss of efficiency that could be regained with dual porting of SRAM. Certainly 2MiB of on SoC chip memory SRAM is no problem and would give blazing performance but modernizing also means significantly increasing chip memory if not making all memory chip memory with minimal performance loss so there is a good chance this is not practical too.

cdimauro Quote:

Caches and/or large bandwidth is and should be enough to address those small data requests.


I believe modern DDR memory is designed to service caches at the end of long multi-level caches, much of the accesses being large prefetches. It makes sense to use modern memory with large multi-level caches to hold the large prefetches. As much memory bandwidth as possible for a SoC is certainly beneficial especially with a high performance GPU.

cdimauro Quote:

Please get rid of this Stone Age technology. Modern memory has already more than needed to address the small requests of legacy/retro system.

Using SDRAM only for this reason will only cripple the whole platform. OR complicate it, if you want to use both SDRAM and DDR at the same time.

So, DDR should be the only memory system. Unless there's a clear proof that DDR causes problems.


Modern DDR memory is also SDRAM but DDR (Double Data Rate) instead of SDR (Single Data Rate). The commands and access sizes changed from generation to generation of SDRAM making them incompatible. A hardware translation layer could likely make different memory look like older memory for simulation purposes kind of like eDRAM can be made to behave like SRAM on an ASIC.

cdimauro Quote:

Nintendo, Sony, SEGA with their respective "mini" versions of the their old consoles.

It should be enough, but if it's not you can take a look also at the C64 Mini and Amiga Mini.


Most of these were quick and cheap eye candy money grabs that have inferior game performance compared to the much older original hardware.

cdimauro Quote:

Do you understand that the current hardware-based systems use FPGAs (so, not ASICs) and they are utterly expensive?

How much it costs a MISTer or a Vampire V4 (let's talk only about the standalone versions, which are "self-contained" AKA SBC)?

How much it costs a low-end (because nowadays it's enough to address most of the retro systems) MiniPC?

Actually the differences between FPGAs and PCs (or RPis) is about how you use the transistors budget. But both require A LOT of transistors.

At least PCs are much more flexible and development of emulators takes a lot LESS effort compared to HDL. Not even counting the performance that they could get if speed is important (and here using a JIT makes at least an order of magnitude of difference).


Cheap FPGAs are cheap and are still very good for retro chipset simulation as these are parallel workloads. Cheap CPUs are good at sequential workloads and do reasonably well with CPU emulation of old CPUs as long as the code and data is in their limited caches. FPGAs large enough to do retro CPU simulation aren't more expensive than large CPUs which give improved emulation. Compare the cost of your x86-64 CPU to the MiSTer FPGA cost. The whole MiSTer base board only cost $140 at one time and that not only included the FPGA but also an ARM SoC, 1GiB of memory, etc. The MiSTer FPGA can't reach anywhere near the CPU performance of your CPU but I believe it still more accurately simulates most retro CPUs than using your CPU emulation. At least on my mid generation Intel Core i5 using Windows, it has inconsistent performance and sometimes noticeable latency. I'm sure your higher performance computer is better but all it takes is for something to fall out of cache to set you back more than a decade in performance. The differential between CPU performance and memory performance has only grown which is why modern CPUs have ridiculous amounts of caches but the Achilles heel remains.

cdimauro Quote:

Because of a psychological reason: they need to "touch" the hardware.

Something which is also happening the mentioned "Mini" retro systems, even if all of them use an emulator inside!

People touch the hardware --> Oooh, it's cool: it's like the original platform!

You see? Psychology...


I consider the Modern Vintage Gamer to be not only credible but an authority on retro gaming who owns many of the original systems. He noticed a different and admitted he was wrong to think emulation was better. In my experience, people usually don't admit they were wrong unless they were wrong.

cdimauro Quote:

On both videos the author says that FPGA systems do cycle-accurate implementations. So, letting people think that this is not possible with emulators.

This is clearly false and totally misleading. As we know, most emulators provide cycle-accurate emulation as well. Since years. And actually better than FPGAs, because emulators are the most accurate.


You have a valid argument here. Even FPGA simulation is not necessarily cycle exact, just more accurate. Retro CPU emulation can be cycle exact especially if you would throw away your jitter causing OS and your caches were large enough for all retro gaming purposes, which they may be. Chipset emulation is more difficult especially worst case depending on the complexity of the chipset which is again about jitter.

cdimauro Quote:

Aside this, today latency could be greatly reduced with GSync or FreeSync displays. In the last years almost all displays (TVs, monitors) implement FreeSync and some GSync.

If you have no such display or you need something better, then you have Lagless VSync: https://blurbusters.com/blur-busters-lagless-raster-follower-algorithm-for-emulator-developers/
WinUAE already implemented it, but other emulators are adopting it.

That should be enough. But one more thing about 3D systems. Emulation could do much better the original systems even in terms of latency. For example: https://www.neogaf.com/threads/dolphin-emulating-wii-and-gamecube-games.395121/page-310#post-246464070
And, in general, 3D systems are not that easy to be implemented using FPGAs. Plus they require A LOT of resources (e.g. LEs). And they might require higher clocks. For both things FPGAs aren't usable.


Part of the problem is that modern CPU and GPU latency and jitter have grown. Turning up the clock and adding more caches, necessitating multi-level caches, actually makes this problem worse. That little Raspberry Pico beats modern CPUs in this regard. It can probably emulate a 6502 and maybe even a 68000 with similar accuracy to a FPGA. Hitting the hardware with no OS isn't much easier than HDL programming though. A system like the Amiga was the next best thing with a very thin and responsive OS on small footprint DMA driven hardware. The 68k Amiga was good at emulation before it became extinct and emulated instead, often by fatter systems. It's kind of like a big truck coming to pick up an efficient little motorcycle to take it places. All that matters is how how much power and payload the big truck has while little attention is payed to how inefficient it is to haul around the efficient little motorcycle. Yea, the motorcycle needs repairs and some modern improvements but it doesn't need a big truck to carry it around.

Last edited by matthey on 05-Sep-2022 at 09:39 PM.
Last edited by matthey on 05-Sep-2022 at 09:34 PM.

 Status: Offline
Profile     Report this post  
matthey 
Re: Commodore Amiga Global Alliance
Posted on 5-Sep-2022 22:30:17
#118 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

kolla Quote:

I would naturally choose alternative A.

I hate it when build jobs break because of “out of memory”, then I rather go for slower CPU, I don’t mind waiting.


lol. For some reason I'm not surprised by your choice. Just don't think you are normal. I have experienced out of memory which interrupted long build jobs and it is irritating but that was with ~100MiB. I can't imagine ever running out of memory on an Amiga with several times that amount of memory let alone 2GiB of memory. I expect it would happen if enough more modern programs were ported to the Amiga but that would be much less likely without improved CPU performance.

kolla Quote:

2GB of ram would also mean holding entire animations and audio tracks in ram instead of relying on storage I/O.


Your storage I/O must be slow then but that is common on the Amiga. I'm used to CSMK3 Ultra SCSI performance which at least relatively feels completely modern.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Commodore Amiga Global Alliance
Posted on 6-Sep-2022 0:19:49
#119 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12795
From: Norway

@pixie

Quote:
PPC is long dead, why not embrace 68k as the defacto standard?


Naive PPC code will always be faster. "DEAD" is relative, as long as you get CPU's that are faster then 680x0 CPU's or JITed code on ARM, the PowerPC will stay.

But if look at larger picture connectivity, is lacking on ARM, 68K upgrades, you simply can’t plug in modern SATA, SOUND and VIDEO cards.

If the future is ARM hardware similar to AmigaONE's and Sam460's then, then future should be system trendy 680x0 code. or generally more open source games and apps.

Last edited by NutsAboutAmiga on 06-Sep-2022 at 12:31 AM.
Last edited by NutsAboutAmiga on 06-Sep-2022 at 12:28 AM.
Last edited by NutsAboutAmiga on 06-Sep-2022 at 12:21 AM.

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

 Status: Offline
Profile     Report this post  
MEGA_RJ_MICAL 
Re: Commodore Amiga Global Alliance
Posted on 6-Sep-2022 1:26:57
#120 ]
Super Member
Joined: 13-Dec-2019
Posts: 1200
From: AMIGAWORLD.NET WAS ORIGINALLY FOUNDED BY DAVID DOYLE

GLOBAL PADDING

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle