Click Here
home features news forums classifieds faqs links search
6179 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
22 crawler(s) on-line.
 95 guest(s) on-line.
 2 member(s) on-line.


 DiscreetFX,  kolla

You are an anonymous user.
Register Now!
 kolla:  16 secs ago
 DiscreetFX:  4 mins ago
 carefuleileen:  6 mins ago
 matthey:  20 mins ago
 danwood:  22 mins ago
 minator:  1 hr 33 mins ago
 number6:  2 hrs 1 min ago
 agami:  2 hrs 9 mins ago
 amigasociety:  2 hrs 45 mins ago
 jingof:  3 hrs 47 mins ago

/  Forum Index
   /  AROS Software
      /  Goodbye PowerPC?
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 Next Page )
PosterThread
V8 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 8:11:58
#161 ]
Regular Member
Joined: 30-Mar-2022
Posts: 141
From: Unknown

@OlafS25

Quote:
MUI5, if you mean the downloadable version, is kind of freeware but not open source. MorphOS team see MUI as core part of the OS and do not want sources to be open Stefan Stuntz was, as I read it, more open but others from the team not


Maybe they should be approached on the subject again?

We all know the reason for not open sourcing it then but that reason is now pretty much moot and irrelevant. Times have changes so maybe the reluctance to open source it has changed too.

 Status: Offline
Profile     Report this post  
AmigaBlitter 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 8:13:04
#162 ]
Elite Member
Joined: 26-Sep-2005
Posts: 3523
From: Unknown

There was the Libre SOC program that i followed for quite some time, but it's seem derailed now for variuos reason. Don't know if someone would/could continue the work.

It was a nice project.

_________________
retired

 Status: Offline
Profile     Report this post  
agami 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 9:04:59
#163 ]
Super Member
Joined: 30-Jun-2008
Posts: 1950
From: Melbourne, Australia

@ppcamiga1

Quote:
ppcamiga1 spewed nonsensically:

to attract people outside current amiga community

The only users that can be attracted from outside of the "current Amiga community", for anything based on Amiga and related tech, are:
- former members of the Amiga community
- former users of Amiga computers that were not involved enough to consider themselves as community members
- former Amiga application users and Amiga game players who cared more for the software than the hardware upon which it ran.

Any of this stuff you list below, does nothing for the potential users listed above.
Quote:
amiga solutuions on pc have to be decent
amiga solutuions on pc have to be unix based

aros x86 should be canceled
aros arm should be canceled
emu68 should be canceled
zune should be made mui compatible
then everyting below amiga gui and graphics should be cut off
and amiga gui and graphics should be ported to unix

The only person who agrees with your list is you.

I get it, you love MorphOS on your Apple G4 Mac mini, and for you the Amiga user experience starts and ends with MUI.
But MUI is not going to get more users to buy dead PowerPC machines.
MUI is not enough to even get any returning users to install some darwin + MUI hybrid on their old PC, even if it had iBrowse 4.

The solutions for the users listed above (in reverse order):
- Cheap and easy access to updated/refreshed versions of their beloved software (A500 mini, A600GS, Amiga Forever, WinUAE, etc.)
- New hardware supporting modern peripherals that embody the Amiga retro experience (A600GS, Apollo V4 Standalone, upcoming RGL Maxi), and for those that still have an Amiga stored away (Apollo V4 accelerators, PiStorm + emu68)
- For former members of the Amiga community? I'm guessing the only thing that'd get those people back into anything Amiga would be if some rich person like Peter Thiel committed to spend $250M+ on resurrecting the Amiga, buys all the IP, brings the old engineering gang together as advisors, and commits to make a line of desktop and laptop machines running AmigaOS 5, outperforming Apple's systems at half the price.

Only that last one is going to see developers come back to anything that is legitimately an Amiga. No MUI needed.

Last edited by agami on 23-Sep-2024 at 09:11 AM.
Last edited by agami on 23-Sep-2024 at 09:10 AM.
Last edited by agami on 23-Sep-2024 at 09:06 AM.
Last edited by agami on 23-Sep-2024 at 09:05 AM.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
pixie 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 10:01:56
#164 ]
Elite Member
Joined: 10-Mar-2003
Posts: 3470
From: Figueira da Foz - Portugal

@agami

If you compare Reaction (OS4) vs Mui (MorphOS) I know what I would choose. I don't say it's needed, I say I'm quite tired of having to re invent the wheel each time.

There was Feelin, sadly it never went anywhere, not even on a fresh start OS such as AROS

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

 Status: Offline
Profile     Report this post  
amigang 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 10:10:56
#165 ]
Elite Member
Joined: 12-Jan-2005
Posts: 2135
From: Cheshire, England

@agami

Quote:
The only users that can be attracted from outside of the "current Amiga community", for anything based on Amiga and related tech, are:
- former members of the Amiga community
- former users of Amiga computers that were not involved enough to consider themselves as community members
- former Amiga application users and Amiga game players who cared more for the software than the hardware upon which it ran.


I completely disagree, they are the easiest to reach, but I met people who didn't grow up with Amiga.

They have joined because they like the community, or they have an interest in general retro tech, or they are just impressed by the tech.

The market value of retro gaming market was valued at $1billion in 2021. Now many will only be interested in one camp, like Nintendo fans, but Id gone to a few retro gaming shows where i seen all fans mix and jump on different machine and see different product, I feel the Amiga community can definitely reach more of these people if more effort was put in.

Now I'm a realist, and you're not going to get suddenly millions of fans but its one area we could grow a bit.

So here is my list of fans we could get

- people who are generally intrested in Retro Tech
- people who are generally intrested in Retro Gaming
- people who are generally intrested in Retro music production
- people who are generally intrested in Retro Graphics
- people who are generally intrested in coding (still a great platform to learn the basic of programing thanks to Amos and other program)
- people who are generally looking for close computer community, with forums and shows to vist
- people who are generally intrested in another way of doing computing
- people who are generally crazy

Last edited by amigang on 23-Sep-2024 at 10:22 AM.
Last edited by amigang on 23-Sep-2024 at 10:22 AM.
Last edited by amigang on 23-Sep-2024 at 10:19 AM.

_________________
AmigaNG, YouTube, LeaveReality Studio

 Status: Offline
Profile     Report this post  
OlafS25 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 11:52:47
#166 ]
Elite Member
Joined: 12-May-2010
Posts: 6491
From: Unknown

@amigang

retro is a wide field. On youtube I searched for "retro" and "games". There was a younger man there, officially interested in retro gaming. But for him game consoles from 2010er years were already retro. So retro depends on every person. And even if you take the people from 80s and early 90s, you have a lot of different camps including people interested in 8bit systems, game consoles or amiga/atari, perhaps even PCs with MSDOS games.

But you are right, at least I see some potential if you would get a linux distribution with preconfigured desktop and optimized for retro, that means games, emulators, easy use of media like adf and so on. It would be at least a try. I would like to do something in that directlion when the base is available. How successful it would be, who knows

 Status: Offline
Profile     Report this post  
amigang 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 14:51:40
#167 ]
Elite Member
Joined: 12-Jan-2005
Posts: 2135
From: Cheshire, England

@OlafS25

yes, plus I was thinking after my post, let just say Amiga could only ever reach 5% of that $1billion market, that still $50 million market that I think Amiga could reach.

_________________
AmigaNG, YouTube, LeaveReality Studio

 Status: Offline
Profile     Report this post  
OneTimer1 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 20:27:36
#168 ]
Super Member
Joined: 3-Aug-2015
Posts: 1219
From: Germany

Quote:

OlafS25 wrote:

retro is a wide field. On youtube I searched for "retro" and "games" ...


Ack!

I once saw a posting in a forum, where someone called "MineCraft a retro game."

If someone would redesign a old Spectrum/C64 game he would make it in 3D, 2k resolution (or better) 24Bit colors and maybe with a multiplayer option over internet. Most "Jump 'n' Run" games today are looking like 3rd person shooters.

Nintendo is a good example for a company making retro games, they are recycling their old game characters over and over again and customers love them.

Last edited by OneTimer1 on 23-Sep-2024 at 09:26 PM.

 Status: Offline
Profile     Report this post  
OneTimer1 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 21:35:23
#169 ]
Super Member
Joined: 3-Aug-2015
Posts: 1219
From: Germany

Quote:

amigang wrote:


... let just say Amiga could only ever reach 5% of that $1billion market, that still $50 million market that I think Amiga could reach.


In it's best time the Amiga sold less computers than Apple and Apple had a market share of 12%, most of the Amigas where cheap A500, so when you are asking for a market share in price it was much less than Apple.

So even in its best time it was less than 5% of the money in the worldwide PC market, I see no way how Amiga could get back to this dream times again.

Last edited by OneTimer1 on 23-Sep-2024 at 10:09 PM.

 Status: Offline
Profile     Report this post  
kolla 
Re: Goodbye PowerPC?
Posted on 23-Sep-2024 22:47:35
#170 ]
Elite Member
Joined: 20-Aug-2003
Posts: 3461
From: Trondheim, Norway

@agami

Quote:
amiga solutuions on pc have to be unix based


You’d be surprised… I’d buy that, easily.

It’s not without irony that the only vendor in end-user unix systems is… Apple.
Not Sun, not SGI, not HP, not Fujitsu, not Microsoft, not DEC, not IBM, not Compaq, not Oracle, not SCO etc etc… but Apple.

Last edited by kolla on 23-Sep-2024 at 10:49 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Online!
Profile     Report this post  
agami 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 3:06:18
#171 ]
Super Member
Joined: 30-Jun-2008
Posts: 1950
From: Melbourne, Australia

@kolla

Quote:
kolla wrote:
@agami

Quote:
amiga solutuions on pc have to be unix based

You’d be surprised… I’d buy that, easily.

It’s not without irony that the only vendor in end-user unix systems is… Apple.

I'm not surprised, because you are in the "current Amiga community". I would likely buy into the same.

Clearly @ppcamiga1 continually argues for the people on his Troll List to repeat what Apple did with OS X, and graft an Amiga UX atop a mature Unix base.
Never mind that Apple had a much larger team, plus a $427M ($837M today) shortcut in purchasing NeXT (NeXTSTEP).

If someone could do that for Amiga, at the quality level of OS X 10.1, then you bet we'd see people from outside the "current Amiga community" return.
I argue that such an accomplishment would draw in users even if it wasn't marketed as an Amiga.

Last edited by agami on 24-Sep-2024 at 07:47 AM.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
agami 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 4:02:28
#172 ]
Super Member
Joined: 30-Jun-2008
Posts: 1950
From: Melbourne, Australia

@amigang

Quote:
amigang wrote:
@agami

Quote:
The only users that can be attracted from outside of the "current Amiga community", for anything based on Amiga and related tech, are:

I completely disagree, they are the easiest to reach, but I met people who didn't grow up with Amiga.

Sounds more like you partially disagree, and I would say mostly because you've overlooked the key qualifier highlighted above.

Quote:
They have joined because they like the community, or they have an interest in general retro tech, or they are just impressed by the tech.

Edge cases do not a target market make.

Quote:
So here is my list of fans we could get

- people who are generally intrested in Retro Tech
- people who are generally intrested in Retro Gaming
- people who are generally intrested in Retro music production
- people who are generally intrested in Retro Graphics
- people who are generally intrested in coding (still a great platform to learn the basic of programing thanks to Amos and other program)
- people who are generally looking for close computer community, with forums and shows to vist
- people who are generally intrested in another way of doing computing
- people who are generally crazy

I on the other hand completely agree with you, but what you are doing is expanding the market beyond the scope of the original scope inferred by @ppcamiga1's Amiga MUI GUI wet dream.

To most of the consumers in the expanded Retro Tech market (above), having things branded as Amiga, or having MUI on a unix base, means very little.

My response intended to highlight the gap between Amiga retro and Amiga NG (actual).
- We have plenty of solutions for the Amiga retro market.
- The kind of Amiga NG solution of which @ppcamiga1 is talking about, and the kind that would be deserving of the label "NG", we have none of those. Furthermore, it's no small task.

Last edited by agami on 24-Sep-2024 at 07:46 AM.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 4:54:40
#173 ]
Elite Member
Joined: 29-Oct-2012
Posts: 4396
From: Germany

@kolla

Quote:

kolla wrote:
@agami

Quote:
amiga solutuions on pc have to be unix based


You’d be surprised… I’d buy that, easily.

It’s not without irony that the only vendor in end-user unix systems is… Apple.
Not Sun, not SGI, not HP, not Fujitsu, not Microsoft, not DEC, not IBM, not Compaq, not Oracle, not SCO etc etc… but Apple.

Then the switch to the dark side will be complete...

 Status: Offline
Profile     Report this post  
Karlos 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 10:06:58
#174 ]
Elite Member
Joined: 24-Aug-2003
Posts: 4944
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

All ppcamiga needs, apart from therapy, is to run AxRT.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
pixie 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 10:24:55
#175 ]
Elite Member
Joined: 10-Mar-2003
Posts: 3470
From: Figueira da Foz - Portugal

@Karlos

But without knowing

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

 Status: Offline
Profile     Report this post  
jonssonj 
Re: Goodbye PowerPC?
Posted on 24-Sep-2024 14:21:09
#176 ]
Regular Member
Joined: 1-Mar-2004
Posts: 299
From: Sweden, Bjärred

My 2 cents is that we need more people interested in developing for the Amiga. I would like more of this content:

https://www.youtube.com/@KSD-AmigaCorner/videos

more learning content, :), and less arguing content, :)

_________________
A1 X1000 is here !!!

 Status: Offline
Profile     Report this post  
Hypex 
Re: Goodbye PowerPC?
Posted on 25-Sep-2024 6:19:02
#177 ]
Elite Member
Joined: 6-May-2007
Posts: 11351
From: Greensborough, Australia

@cdimauro

Quote:
Sorry, I don't know how to help you, because it doesn't happen to me


That looks like a terrible line from a software developer.

Well, I got it working in FireFox, for one article. Loaded another and it didn't ask to translate again. Maybe the quick brown fox jumps over the lady code.

Quote:
You don't need to register to DeepL: it works even without it. However, you can translate more words (up to 5000 characters) if you register.


I had another go and decided to put it in the too hard basket. Nothing obvious to translate site. Pressed a button that said try for free. Wants a sign in and opened another page. Plug ins that open pages so I lose my place in the tabs are grounds for instant dismissal and uninstalling.

Quote:
You're good for doing QA then.


It's good for starters.

Quote:
In fact, I've a completely different experience (always was good with Windows, and had problems with Linux).


I have a friend that thinks I have too much open at once. But then installs things that open on boot. I do make use of multitasking and running a few apps at once if I go from one to another while working no stuff.

I find Linux can be faulty with mobiles and AFT MTP driver. Limitations in copying files. Hangs. Immediate error even plugging it in. Where as in Windows it tends to work fine.

Quote:
In fact, I've a completely different experience (always was good with Windows, and had problems with Linux).


Quote:
It worked with big endian systems as well.


Is there any info online about it? I can't find any except for discussion about it. It makes me wonder, since the PPC was primarily a big endian ISA, why MS would bother running the PPC as LE if they could just use native BE mode? It was early then for PPC and LE had limitations. They must have been one of, if not the first, to get a kernel booting itself in LE mode. A few decades before there were any common LE bootloaders and other kernels running PPC in LE.

Quote:
As I've said, NT was ported to very different architectures and with different endians. It has no problem with that, because it had a good design.


The design would have kept the internals private so that does help. I do wonder if they needed to invent another exe format. I used to refer to MZ exe files as Microsoft Z, because it looked like it and I didn't know what MZ was nor what the Z would be.

Quote:
There are several things which aren't correct.


In response to your points...

1. This is really an old example. RGB15PC and similar have been used in Amiga screen modes since early RTG days. Correlating to RGB15. So really RGB15 LE or RGB15 BE in 16 bit word. My point here was that storing the bytes backward in LE makes no sense of the bit packed data inside as it breaks it up. But I can see it's about storage byte order and not bitfields which are just internal format. I could extend this, as AMD used to support BE in driver objects passed to hardware, then dropped support and thus broke support for R series cards up which can only use framebuffer and no hardware 3d.

2. I don't know the exact details but those who work on OS4 browser ports would know what it is. I keep reading about endian issues when porting and IIRC relating to JavaScript. I'm not aware of any JIT support being in OS4 browsers. There was a known issue with FireFox rendering images in LE RGB in the past as well.

3. Strings really are endian agnostic since they are bytes in the case of ASCII. I've seen more code that cheats by looking up 32-bit 'words" in reverse for LE, than BE code cheating by looking up the same "word" in order. I couldn't imagine the mental gymnastics in reversing numbers in code because it assumed CPU looked at it that way, but reading backwards doesn't come naturally to me either. For such tricks my example was referring to using a MAKEID('T','E','S','T') type macro as a portable solution, which is both readable and compatible across endians. However, for ELF parsers the code I have seen checks the ID by the byte, and doesn't attempt any cheats.

4. I realise it's not clear and after reposting the idea and finding other people couldn't figure that out either lol. It depends on how code is written and the example is slightly low level as it involves using CPU registers for optimum efficiency. It's easier to understand if going above the operation of packing variable sized bit-fields to loading or storing the data. To keep it simple I'll only deal with 32 and 64 bit scalars and leave larger scalars found in vectors or floats out of it. So, assuming there is a bit packed stream of data of size 64 bits, containing this:
00 01 02 03 04 05 06 07

For x86/32, if the data is processed and stored in one 32-bit register, then storing it sequentially in memory will end up as:
03 02 01 00 07 06 05 04

This is fine but if we want to optimise our code for x86/64, processing in 64 bit, then this happens on store:
07 06 05 04 03 02 01 00

Now, this data would then be saved to disk for retrieval later, so at this point is fine. The problem arises if we want to load the data on a CPU with with a different register size.

If x86/32 reads in data from x86/64 the stream will load as:
04 05 06 07 00 01 02 03

If x86/64 reads in data from x86/32 the stream will load as:
04 05 06 07 00 01 02 03

In either case, every pair of 32 bit words, will read in backwards! The data may still be correct in each 32-bit word, but words are in incorrect order. Thus, the data becomes corrupt.

If the same operation is performed on BE CPU, such as PPC32 and PPC64, then the data is always exactly the same.

This could be seen as an esoteric example but demonstrates where LE byte order of different scalars is interoperable and incompatible. Older Amiga code works with variable bit streams in whole registers. And I've seen data formats such as StarCraft archives where the data is stored in blocks of 32 bit width.


5. By using the term endian to describe bit order has slightly confused the meaning. But I did this to keep things simple. According to GCC docs endian can affect positions and alignment of bitfields as well.

6. I could add the PPC32 bit order of 0/MSB to 31/LSB.

In my experience the bits were always numbered downwards, from MSB to LSB, greater to lower. Then IBM or Motorola messed this up on PPC and decided to flip the numbers. It's an ISA, not a FFT, what on earth? But, a point I was making is that for LE or BE, at least in the modern age, that the MS to LS bits in bytes are in the same order, represented left to right. Like so:

ASCII "A" or $41:
LE: 0100 0001 / 41
BE: 0100 0001 / 41

But you knew that.

Quote:
That should be more or less everything, but probably I forgot something (I recall some jokes against the little endian system, but I'm not able to find it now).


That article became a monster. It started off as a small article. Then I ended up researching where x86 came from and I was unaware of anything before 8088 nor the history before Intel. I learned a lot doing it and must have spent a month on the whole thing. Unfortunately I didn't keep a list of citations I'm aware of. A lot were in the text itself but they kept turning text blue which looked too distracting to me so I removed them. Blogspot already made it hard to neatly format text and I gave up fighting with it.

I imagine these days most of the jokes would be against big endian.

Quote:
That's part of the C++ standard: it has nothing to do with Windows.


I would have been focussing on this remark:
All native scalar types are little-endian for the platforms that Microsoft Visual C++ targets (x86, x64, ARM, ARM64).

Quote:
Google is very well known of unfinished products.


They've had enough time. I'm suspicious that Google can delete pages without asking but AmigaWorld can't redirect the page without asking me. Google is your friend, your stalker and has a backdoor to your browser.

 Status: Offline
Profile     Report this post  
Jose 
Re: Goodbye PowerPC?
Posted on 25-Sep-2024 15:12:10
#178 ]
Super Member
Joined: 10-Mar-2003
Posts: 1002
From: Unknown

Regarding potential new users I see a huge amount of people dissatisfied with big tech these days, lots and lots of people still use Win7 for a reason, starting from Win10 on Windows has become a spyware machine... But a lot of them hate Linux. The prospect of a system that allows one to got deep in it if one wants to yet is presented in a digestible manner like AmigaOS would be appealing IMHO.. But for that we need a system that is a bit more mature in terms for SMP and memory protection...

Last edited by Jose on 25-Sep-2024 at 03:12 PM.

_________________

José

 Status: Offline
Profile     Report this post  
cdimauro 
Re: Goodbye PowerPC?
Posted on 26-Sep-2024 5:38:34
#179 ]
Elite Member
Joined: 29-Oct-2012
Posts: 4396
From: Germany

@Hypex

Quote:

Hypex wrote:
@cdimauro

Quote:
Sorry, I don't know how to help you, because it doesn't happen to me


That looks like a terrible line from a software developer.

Indeed. "It works on my computer."
Quote:
Quote:
It worked with big endian systems as well.


Is there any info online about it? I can't find any except for discussion about it.

https://en.wikipedia.org/wiki/Windows_NT#32-bit_platforms

NT supported MIPS, PowerPC and SPARC, which at the time were only big endian (no bi-endianess).
Quote:
It makes me wonder, since the PPC was primarily a big endian ISA, why MS would bother running the PPC as LE if they could just use native BE mode? It was early then for PPC and LE had limitations. They must have been one of, if not the first, to get a kernel booting itself in LE mode. A few decades before there were any common LE bootloaders and other kernels running PPC in LE.

In fact, it didn't happened: there were no LE PowerPCs at the time, but only BE ones.
Quote:
Quote:
As I've said, NT was ported to very different architectures and with different endians. It has no problem with that, because it had a good design.


The design would have kept the internals private so that does help. I do wonder if they needed to invent another exe format. I used to refer to MZ exe files as Microsoft Z, because it looked like it and I didn't know what MZ was nor what the Z would be.

MZ files were always used, with very limited extensions to support different architectures: https://en.wikipedia.org/wiki/Portable_Executable
Quote:
Quote:
There are several things which aren't correct.


In response to your points...

1. This is really an old example. RGB15PC and similar have been used in Amiga screen modes since early RTG days. Correlating to RGB15. So really RGB15 LE or RGB15 BE in 16 bit word. My point here was that storing the bytes backward in LE makes no sense of the bit packed data inside as it breaks it up. But I can see it's about storage byte order and not bitfields which are just internal format.

Well, bit ordering is also important and would have further created other incompatibilities.
Quote:
I could extend this, as AMD used to support BE in driver objects passed to hardware, then dropped support and thus broke support for R series cards up which can only use framebuffer and no hardware 3d.

That's exactly my point: it's up to the vendor decide which formats to support. And that's nothing which the OS can control: its drivers have to support what it is found.
Quote:
2. I don't know the exact details but those who work on OS4 browser ports would know what it is. I keep reading about endian issues when porting and IIRC relating to JavaScript. I'm not aware of any JIT support being in OS4 browsers.

Well, it can happen with a Javascript engine (non JITted) as well, but it should be really weird to have it: developers should have done a very bad and dirty job to cause endianess issues in this case.
Quote:
There was a known issue with FireFox rendering images in LE RGB in the past as well.

That's a bug. And bug needs to be fixed.
Quote:
3. Strings really are endian agnostic since they are bytes in the case of ASCII. I've seen more code that cheats by looking up 32-bit 'words" in reverse for LE, than BE code cheating by looking up the same "word" in order.

But that's not cheating: it's like it should be.
Quote:
I couldn't imagine the mental gymnastics in reversing numbers in code because it assumed CPU looked at it that way, but reading backwards doesn't come naturally to me either.

That happens because you're looking at the them as a human being.
Quote:
For such tricks my example was referring to using a MAKEID('T','E','S','T') type macro as a portable solution, which is both readable and compatible across endians.

Exactly! That's the point.
Quote:
However, for ELF parsers the code I have seen checks the ID by the byte, and doesn't attempt any cheats.

So, they are not optimized.
Quote:
4. I realise it's not clear and after reposting the idea and finding other people couldn't figure that out either lol. It depends on how code is written and the example is slightly low level as it involves using CPU registers for optimum efficiency. It's easier to understand if going above the operation of packing variable sized bit-fields to loading or storing the data. To keep it simple I'll only deal with 32 and 64 bit scalars and leave larger scalars found in vectors or floats out of it. So, assuming there is a bit packed stream of data of size 64 bits, containing this:
00 01 02 03 04 05 06 07

I assume the first byte in memory is the LSB.
Quote:
For x86/32, if the data is processed and stored in one 32-bit register, then storing it sequentially in memory will end up as:
03 02 01 00 07 06 05 04

It depends entirely on how you have imagined the above data. If, as I've said before, first byte for you is the LSB, then on a little endian CPU it's:

00 01 02 03 04 05 06 07

Otherwise (first by is the MSB) it's:

07 06 05 04 03 02 01 00

But certainly not what you've reported.
Quote:
This is fine but if we want to optimise our code for x86/64, processing in 64 bit, then this happens on store:
07 06 05 04 03 02 01 00

Now, this data would then be saved to disk for retrieval later, so at this point is fine. The problem arises if we want to load the data on a CPU with with a different register size.

If x86/32 reads in data from x86/64 the stream will load as:
04 05 06 07 00 01 02 03

If x86/64 reads in data from x86/32 the stream will load as:
04 05 06 07 00 01 02 03

In either case, every pair of 32 bit words, will read in backwards! The data may still be correct in each 32-bit word, but words are in incorrect order. Thus, the data becomes corrupt.

If the same operation is performed on BE CPU, such as PPC32 and PPC64, then the data is always exactly the same.

This could be seen as an esoteric example but demonstrates where LE byte order of different scalars is interoperable and incompatible. Older Amiga code works with variable bit streams in whole registers. And I've seen data formats such as StarCraft archives where the data is stored in blocks of 32 bit width.

I think that the example is still not correct, because you're thinking about the data as a human being.

In fact, storying data with LE CPU it's perfectly coherent with both 32 and 64 bit architectures. As an example, let's take -2, which is:
FE FF FF FF in 32 bits and:

FE FF FF FF FF FF FF FF in 64 bits.

As you can see, the number is naturally extended moving from 32 to 64 bits.

I leave to you the example with a BE system.
Quote:
5. By using the term endian to describe bit order has slightly confused the meaning. But I did this to keep things simple. According to GCC docs endian can affect positions and alignment of bitfields as well.

Exactly, but in this case / point was your idea to mix your attitude as human being on how to see the numbers, with how a processor internally works.

As I've told you, I represent LE data always starting with the MSB/MSb from left to right, because... it's me like a human being.
Quote:
6. I could add the PPC32 bit order of 0/MSB to 31/LSB.

Oh. Coherent.
Quote:
In my experience the bits were always numbered downwards, from MSB to LSB, greater to lower. Then IBM or Motorola messed this up on PPC and decided to flip the numbers. It's an ISA, not a FFT, what on earth?

But that's OK: it's a legit decision, and makes the ISA coherent: both byte and bit ordering are "big".
Quote:
But, a point I was making is that for LE or BE, at least in the modern age, that the MS to LS bits in bytes are in the same order, represented left to right. Like so:

ASCII "A" or $41:
LE: 0100 0001 / 41
BE: 0100 0001 / 41

But you knew that.

Yes. But if you try to build such numbers one by one with a big -endian bit ordering... you don't get the same result.
Quote:
Quote:
That should be more or less everything, but probably I forgot something (I recall some jokes against the little endian system, but I'm not able to find it now).


That article became a monster. It started off as a small article. Then I ended up researching where x86 came from and I was unaware of anything before 8088 nor the history before Intel. I learned a lot doing it and must have spent a month on the whole thing. Unfortunately I didn't keep a list of citations I'm aware of. A lot were in the text itself but they kept turning text blue which looked too distracting to me so I removed them. Blogspot already made it hard to neatly format text and I gave up fighting with it.

I imagine these days most of the jokes would be against big endian.

It depends on the audience, but yes: it might be.
Quote:
Quote:
That's part of the C++ standard: it has nothing to do with Windows.


I would have been focussing on this remark:
All native scalar types are little-endian for the platforms that Microsoft Visual C++ targets (x86, x64, ARM, ARM64).

Ehm... no. There's no little-endian enforcing there. Not needed: LE is the natural byte order, and you don't need to for it.

 Status: Offline
Profile     Report this post  
agami 
Re: Goodbye PowerPC?
Posted on 27-Sep-2024 2:39:08
#180 ]
Super Member
Joined: 30-Jun-2008
Posts: 1950
From: Melbourne, Australia

@Karlos

Quote:
Karlos wrote:

All ppcamiga needs, apart from therapy, is to run AxRT.

Yep. It covers all his fetishes:
- running native linux Quake binaries, no emulation
- modern GPU, no AGA

I wonder if deadw00d is considering an ARM port for AxRT. Sure, app developers would need to compile their apps for AROS ARM, but then one could have a nice and cheap Raspberry Pi 5 + AxRT mini computer. Or perhaps even run AxRT on A600GS.

_________________
All the way, with 68k

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 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