@michalsc

For Emu68, is there a pathway to bypass the 68K translation layer for native ARM instructions?

IEEE library port would probably benefit from it.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@michalsc

Quote:

michalsc wrote: @ZXDunny But matthey is right. 1GHz 68060 is much better than JIT running on ARM using the same clock speed. Much better code density and for sure its L1, L2 and L3 caches are running at the same over 1GHz clock speed as the 68060 itself. That must be, otherwise we wouldn't have such low latencies for loads and stores. I just wonder why we all have not yet switched to MC68060 running at 1 or 2GHz. ... oh, wait... ;)

Money doesn't grow on trees.

Freescale didn't spin off 68K into a separate company e.g. 68K Inc, like MIPS Inc.

Freescale used 68K's embedded revenue stream to fund PPC related road maps. PPC gains compressed 16bit ISA via PPC e200 cores.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@Karlos

Quote:

Karlos wrote: @cdimauro It's unrealistic because no matter how good the potential of the 68K ISA for embedded applications, there are already many competitors, including ARM and x86. Not to mention all the microcontrollers for less demanding applications. The relative cost for ARM now is low enough that any hardware you get for such purposes is likely overkill already and maybe you care more about power efficiency than you do about overall performance for your embedded application. I don't see a 68K hardware revival happening other than a vanity project from an eccentric with the resources to burn.

ARM is changing its policy and also entering the market with its own chips. So, it's ready to become a treat even to its current customers.

Besides that, there are already architectures other than x86 and ARM. RISC-V is another fresh architecture which gained consensus from nowhere, and it's a treat to ARM and other architectures which operate on the embedded market.

In short: there's room for competition, IF there's added value to the customers. It's "just" a matter of finding resources.

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Right. Because a system with an MMU was EXPENSIVE (NOT only for the hardware implementation). The same reason why not only the Amiga OS, but other systems / OSes have NOT used an MMU since very long time. Something which you still don't get, because as a bot you don't understand and you're not able to contextualize. Motorola made MMU to be a premium, and 68551 were late, which caused Commodore to commit R&D resources for 68000's MMU and 68020's MMU. The 1st C= custom MMU works, but it was slow and needed a TLB cache. Commodore's memory protected/multi-user OS development was focused on aiding AT&T's expensive license Unix platform instead of AmigaOS. MS's 1988 focused on in-house OS/2 3.0 project aka Windows NT to replace MS Xenix. Steve Job's 1986 focused on freebe *nix kernel with value-add higher layers for NextStep, that later served as the foundation for MacOS X. Commodore's 2nd custom MMU was abandoned for 68551 which caused delays for A2620's release. In 1988, Commodore's MMU engineer, Bob Welland later moved to Apple and focused on RISC ARM-based devices. The first true ARM MMU design was Bob Welland, for Apple. https://www.linkedin.com/in/bob-welland-1349193/details/experience/ At Apple, Bob Welland designed "ARM MMU Architecture". Bob Welland argued for MMU for the masses and RISC during his time at Commodore.

Still this MMU padding, which is a complete non-sense, since MAINSTREAM OSES HAVE NOT USED MMUS!

I've capitalized it for YOUR benefit. But I doubt that a bot could get it.
Quote:

Both 16-bit Windows/32-bit Win32S and AmigaOS have shared memory address space. Win16 was boxed into a virtual machine on memory protected Windows NT/Windows 95. MS boxed their legacy software environment e.g. Windows NT's NTVDM (NT Virtual DOS Machine). In 1997, Microsoft purchased Connectix, a Virtual PC specialist, and led into Hyper V. When PC and Mac switched to C3 rated memory protected/multi-user 32bit OS, both platforms were ready. Windows 3.1 enhanced mode was a 32-bit protected mode virtual machine manager that ran WIndows 3.1 standard mode, Win32S, and DOS boxes virtual machines. Windows 3.1 and Win32 have shared memory address space with no memory protection. Windows NT 3.1 boxed Win16 environment. My point, MS mastered virtual machine software tech.

And here again mixing apples and oranges: Windows NT was NOT a mainstream OS!

The mainstream Windows version suffered of the same problems of ALL other mainstreams OSes, Amiga OS included.

When Windows '95 was out I've created a simple 3 (THREE) bytes program which completely froze it:
- open DEBUG.COM;
- type "a 100";
- type "CLI";
- type "JMP 101";
- type "G";
- enjoy!

Guess what: even with some good memory protections added to the system, Windows (read: NOT NT!) was forced to share part of the memory and I/O resources (which included interrupts handling), because it had to retain DOS-compatibility.
And this continued until Windows XP. Release date for XP: 2001. Read again: 2001.

Compatibility was so much important the Apple was forced on a similar situation (but worse than Microsoft with Windows), which prevented to get a robust and modern OS 'til MacOS X, as I've already written. Release date for OSX: 2001. . Read again: 2001.

That was the common ground for all mainstream OSes, 'til the new millennium.

So, the MMU was NOT a requirement neither a concern.

Yes, Windows 9x demanded it for the virtual memory, but just because it was already built-in on the 386, and it became a requirement. But for the software of the time, it wasn't a requirement.
Quote:

Go read Commodore - The Final Years book, you ignoramus.

The problem is not reading, rather UNDERSTANDING, and this is something which bots can't do.

You can read whatever you want, but if you're not able to understand and contextualize, then reading was completely meaningless. Which is what happens with you, in fact.

Thanks to what you and Bruce have shared from such book, I was able to grab the relevant information, draw the lines between the points, and get an overall picture of the situation.

BTW, some information was coming from the first book of Bagnall: "Commodore : a company on the edge", which I've read some years ago (it's a bit old now).
This book already reported some useful information about your beloved engineers. In fact, he reported that after the original Amiga engineers left the company, the ones which remained were not able to advance the chipset because they lacked experience on designing chips, and it required years to acquire this ability.
I haven't it anymore, because it was from a friend which borrowed it, so I can't report the precise quotes (but I remember very well this part which I've reported).

Anyway, and I've said, reading is NOT enough to understand. Which is you biggest problem.
Quote:

PiStorm-Emu68 software is designed as a hypervisor and has a high resistance against Amiga software crash.

Irrelevant. Hammer's padding.
Quote:

32bit ARM's birth was mostly due to Commodore (Jack Tramiel)'s slow 8bit 65xx R&D road maps.

Irrelevant. Hammer's padding.
Quote:

ARM MMU's were due to Commodore (Henri Rubin)'s anti-MMU for the masses. Commodore management didn't plan for the next generation C3/POSIX rated AmigaOS.

Irrelevant. Hammer's padding + MMUs were NOT important for the mainstream market, as I've already reported.
Quote:

Commodore didn't keep up with the competition, ex-Commodore engineers aided the competition.

See above: FALSE for the OS, because the situation was the same for all consumer platforms, and only at the beginning of the new millennium the situation radically changed.
Quote:

Quote: I reveal you a secret: MacOS X is VERY DIFFERENT from MacOS (which... did NOT... require an MMU! Now take some salts to recover from the shock). The same thing could have happened to the Amiga. There was simply (!) not enough time to correctly evolve the platform. Sic et simpliciter. Quote: Don't assume I don't know MacOS X is different from MacOS.

That's not the point. You can know MacOS X but you don't know how the situation with the mainstream OSes of the time (before OS X) was AND you're NOT able to contextualize.
Quote:

Both MS and Apple (Steve Jobs) boxed their legacy software environments.

WHEN? That's the point. And the point is that you're NOT able to contextualize, as I've said before.
Quote:

PowerPC versions of Mac OS X up to and including Mac OS X 10.4 Tiger include a compatibility layer for running older Mac applications, the Classic Environment (aka Blue box). Read https://en.wikipedia.org/wiki/Sandbox_(computer_security) Sandboxes may be seen as a specific example of virtualization

See above: this is about the NEW age for consumers' OSes.
Quote:

Your assumption is wrong.

My assumption is that bots are NOT able to understand and contextualize. And I'm totally right here.

@cdimauro

Quote:

And here again mixing apples and oranges: Windows NT was NOT a mainstream OS!

Yeah, but let's be honest, everything Windows prior to NT was unmitigated turd. NT 4.0 was the first version of Windows that didn't completely sap your will to live after 30 minutes. Windows 2000 was actually bearable and occasionally even useful. Then there was XP, which I hated out of the box (but endured after getting rid of all the default theme), but I'm told was apparently quite popular :D

_________________
Doing stupid things for fun...

@Karlos

Quote:

Yeah, but let's be honest, everything Windows prior to NT was unmitigated turd.

Windows 3.x 386 reach to a stage that they are good enough for many users, hence the sales boom with Windows 3.0.

Windows 3.0 is the platform to run MS Excel, MS Word and MS Office. Business process document automation was the killer app with MS Office.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@Karlos

Quote:

Karlos wrote: @cdimauro Quote: And here again mixing apples and oranges: Windows NT was NOT a mainstream OS! Yeah, but let's be honest, everything Windows prior to NT was unmitigated turd. NT 4.0 was the first version of Windows that didn't completely sap your will to live after 30 minutes. Windows 2000 was actually bearable and occasionally even useful. Then there was XP, which I hated out of the box (but endured after getting rid of all the default theme), but I'm told was apparently quite popular :D

But mainstream customers haven't used Windows NT until it was presented in its more commercial form: XP.

The first time which I've used NT was when I was doing my stage at STMicroelectronics for my thesis (a JPEG2000 decoder to be implemented in hardware), and it was 2003 if I recall it correctly.

NT was very solid, but required much more resources, and was lacking multimedia/game capabilities / extensions. That's the reason why it was used only by some professionals.

Consumers, like me, "just" used the mainstream Windows versions, with their fragility (like happened with ALL consumers systems: Macs, Ataris, Amigas, II GSes, Archimedes, ...).
Only on 2000s (some) consumers systems started switching to much more robust/solid OSes versions, which required an MMU and other protections.


@Hammer

Quote:

Hammer wrote: @Karlos Quote: Yeah, but let's be honest, everything Windows prior to NT was unmitigated turd. Windows 3.x 386 reach to a stage that they are good enough for many users, hence the sales boom with Windows 3.0.

Prove it! IF you have numbers (for all hardware platforms supported by Windows 3.0: 8088/8086, 80286, 80386+) that can justify your statement.
Quote:

Windows 3.0 is the platform to run MS Excel, MS Word and MS Office. Business process document automation was the killer app with MS Office.

Office applications worked also on Windows 3.0 / Real mode (8088/8086)...

@cdimauro

Quote:

Still this MMU padding, which is a complete non-sense, since MAINSTREAM OSES HAVE NOT USED MMUS! I've capitalized it for YOUR benefit. But I doubt that a bot could get it.

MS Windows 3.1 Enhance Mode has virtual memory you ignorant fool.

Apple Computer first officially added support for virtual memory with the release of System 7 in 1991 you ignorant fool.

Windows 3.1 uses a hardware feature of the 386 processor and its built-in Memory Management Unit (MMU). Virtual 8086 (Address) Mode creates multiple instances of an 8086 processor with each having separate RAM.

MMU is needed for "next gen" OS development.

Intel 80376 doesn't have MMU and it was replaced by the much more successful 80386EX in 1994. 80376 doesn't have legacy 16bit and MMU i.e. it's Intel's 68020 counterpart based on 80386.

NEC V30's 8086 clone was pushed out of the market.

EMM386.EXE requires 386 MMU since it emulates EMM.
EMM286.EXE requires 286 MMU since it emulates EMM.

Windows 3.x has three modes of operation:
Real Mode, Windows and the applications are subjected to the limitation of the 8086. This mode was abandoned by Windows 3.1.

Standard Mode requires 286's protected mode. It's sayonara to NEC V30's 8086 clone.

Enhanced Mode requires 386's protected mode. It is built on the 32-bit “Virtual Machine Manager� (VMM) that leverages the virtual 8086 mode to create and manage several virtual machines: one for the Windows Operating Environment, the System VM, and one for each DOS session. Enhanced Mode also introduces Virtual Device Drivers (VxDs): 32-bit protected mode drivers that can share system resources to more than one process.


Quote:

And here again mixing apples and oranges: Windows NT was NOT a mainstream OS!

In the USA and for MS OS offerings, Windows NT was the mainstream OS for C3 / Orange book/TCSEC and POSIX compliance you ignorant fool.


Quote:

The mainstream Windows version suffered of the same problems of ALL other mainstreams OSes, Amiga OS included. When Windows '95 was out I've created a simple 3 (THREE) bytes program which completely froze it: - open DEBUG.COM; - type "a 100"; - type "CLI"; - type "JMP 101"; - type "G"; - enjoy!

1. That's a red herring since that sequence is not a normal operation from a normal user.

2. Your example doesn't dismiss 386's MMU requirement for Windows 95. Memory protection is applied for userland Win32 apps.

3. Windows 95 is NOT designed for C3 / Orange book/TCSEC compliance, hence Windows 95's memory protection coverage is less than NT's.

Under Windows NT, DEBUG doesn't have direct access to hardware due to the extra abstraction layer. DEBUG doesn't have sector access for HDD and it has FDD sector access.


Quote:

Guess what: even with some good memory protections added to the system, Windows (read: NOT NT!) was forced to share part of the memory and I/O resources (which included interrupts handling), because it had to retain DOS-compatibility. And this continued until Windows XP. Release date for XP: 2001. Read again: 2001.

1. Windows 2000 (NT5.0, December 15, 1999 RTM) and Windows XP (NT 5.1) have very similar Games for Windows support as Windows XP e.g. Windows 2000 supports DirectX 7 to DirectX9c. Both Windows 2000 and XP don't support Windows Vista exclusive DirectX10.

2. For Windows are based on the "NT", to run DOS (and 16-bit Windows) applications, NT-based versions of Windows use a "Virtual DOS Machine" (VDM) a full featured MS-DOS emulator and virtual machine with very limited access to real hardware.

On DOS text UI mode, Doom DOS version will run on Windows NT 3.1 until Doom switches to the full screen mode. MS didn't add sufficient VGA emulation for VDM. There's a reason for Microsoft's push for DirectX and Games for Windows.

NTVDM has DOS compatibility up to a certain level. NTVDM can run DOS-based Win16.

Windows 9x is needed for FULL DOS compatibility.

Your DEBUG example is a DOS program that can run on Windows 95's VDM and NT's NTVDM. LOL

My university's 1997 PC fleet has Windows NT 4.0 (released in 1996) and we run Windows OpenGL / DirectX3a games on them in a large after hours LAN party. WinQuake and GLQuake was released in Jan 1997. University's PC fleet resource needs to be controlled.


Quote:

Compatibility was so much important the Apple was forced on a similar situation (but worse than Microsoft with Windows), which prevented to get a robust and modern OS 'til MacOS X, as I've already written. Release date for OSX: 2001. . Read again: 2001.

MMU is needed for "next gen" OS development and building the install base with MMU.

Steve Job's NextStep was available for 68K with MMU, PPC and i386.

A desktop platform vendor must plan with a proper road map. Since Commodore didn't properly invest in AmigaOS's C3 / Orange book/TCSEC compliance R&D, Commodore has statements like running Windows NT on PA-RISC based Amiga Hombre.

Commodore management wasted OS development resource and time to benefit AT&T's Unix, not AmigaOS. Commodore didn't have "Amiga 1st" ideology. Apple has "MacOS 1st" ideology.

Quote:

That was the common ground for all mainstream OSes, 'til the new millennium.

1. That's wrong for C3 / Orange book/TCSEC workplace requirement.

2. That's wrong for MMU install base build-up. Imagine in the year 1999 when MS Windows 2000 was released to manufactuerers (RTM) with near zero MMU install base. That's a Commodore level fuckup.

Imagine MacOS X was released without an existing MMU install base... LOL. A f__kup like Commodore.


Quote:

So, the MMU was NOT a requirement neither a concern.

MMU is a concern for the "next gen" OS development and building MMU install base.

MMU is a concern for faking EMM compatibility instead of using hardware bank-switching. Faking memory expansion type is a concern for the PC.


Quote:

Yes, Windows 9x demanded it for the virtual memory, but just because it was already built-in on the 386, and it became a requirement. But for the software of the time, it wasn't a requirement.

That's a red herring. My argument is not about minimum system requirements.

NEC V30 clone died since it didn't have 386's 32-bit protected mode. For x86 cloners, AMD separates from NEC on x86 MMU mastery.

https://hackaday.com/2024/02/09/breaking-through-the-1-mb-barrier-in-dos-with-unreal-mode-and-more/

This article builds on the first one which covered the older methods and covered the basics of protected mode. Where protected mode is convenient compared to real mode is that with the former the memory accesses go via the MMU and thus allows for access to 16 MB on the 80286 and 4 GB on the 80386. The segment descriptors and resolving of these that make this possible can be (ab)used on the 80286 and up by realizing that these segment descriptors are also used in real mode. Unreal mode is thus about switching to protected mode, loading arbitrary segment descriptors and switching back to real mode. As this is outside the original processor spec, it is commonly called ‘unreal mode’.


There was an EMM286.EXE and HIMEM.SYS (version 2.06 or later) using the 286 CPU’s MMU instead of any hardware bank-switching like most of the EMS cards of the 286’s era. For legacy support, faking memory expansion type is a concern on the PC.

You are wrong.

Reminder, Linux was born on 386 PC, not on an expensive MMU equipped Amiga with an install base of a few thousands!

https://archive.computerhistory.org/resources/access/text/2013/04/102723262-05-01-acc.pdf
DataQuest 1991, page 119 of 981

For 1992,
68EC020-16, $16.06
68EC020-25, $19.99
68020-25, $35.13
68EC030-25, $35.94 (missing MMU, not Unix capable, used in A4000/030)
68030-25, $108.75 (MMU equiped)
68040-25, $418.52 (MMU equiped)
68EC040-25, $112.50 (missing MMU and FPU, Commodore management rejected glue chips for Amiga)

AM386-40, $102.50 (MMU equiped)
R3000-25, $96.31 (MMU equiped)

386DX-25, $103.00 (MMU equiped)
80486SX-20, $157.75 (MMU equiped for Xenix 386, Windows NT, and Linux)
80486DX-25, $376.75 (MMU equiped)
80486DX-33, $376.75 (MMU equiped)
80486DX-50, $553.25 (MMU equiped)
80486DX2-50, $502.75 (MMU equiped)


All 32-bit X86 CPUs have MMU. 68030-25 already lost "value for money" when compared to AM386-40 and MIPS R3000. Motorola's 68030-25 nearly copies Intel's 386DX-25 price guide.


For 1993 Q1,
68EC020-16, $15.00
68EC020-25, $18.00 ($3 difference from EC020-16)
68020-25, $33.00
68EC030-25, $34.00 (missing MMU, not Unix capable, used in A4000/030)
68030-25, $95.00 (MMU equiped)
68040-25, $350.00
68EC040-25, $95.00 (missing MMU and FPU, Commodore management rejected glue chips for Amiga)

386DX-25, $65.00 (MMU equiped)
80486SX-20, $93.00 (MMU equiped for Xenix 386, Windows NT, and Linux)
80486DX-25, $288.00 (MMU equiped)
80486DX-33, $288.00 (MMU equiped)
80486DX-50, $375.00 (MMU equiped)
80486DX2-50, $385.00 (MMU equiped)

AM386-40, $48.00 (MMU equiped)
R3000-25, $87.00 (68040 level with MMU)

A no brainer to why Motorola lost both desktops and workstation markets. Motorola positions MMU equiped 68K as a premium. Motorola was sleeping on the wheel when AMD lowered AM386-40's price to $48. Intel wasn't sleeping on the wheel.

For handhelds, Palm selects TI's ARM 925T with MMU to replace Motorola's MMU-less 68000 based Dragon Ball VZ. Motorola repeats the same market segment lost like the earlier desktops and workstation markets.

Last edited by Hammer on 11-Mar-2025 at 03:52 AM.
Last edited by Hammer on 11-Mar-2025 at 03:50 AM.
Last edited by Hammer on 10-Mar-2025 at 07:29 AM.
Last edited by Hammer on 10-Mar-2025 at 06:40 AM.
Last edited by Hammer on 10-Mar-2025 at 06:35 AM.
Last edited by Hammer on 10-Mar-2025 at 06:34 AM.
Last edited by Hammer on 10-Mar-2025 at 06:15 AM.
Last edited by Hammer on 10-Mar-2025 at 06:07 AM.
Last edited by Hammer on 10-Mar-2025 at 05:59 AM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@cdimauro

Quote:

Prove it! IF you have numbers (for all hardware platforms supported by Windows 3.0: 8088/8086, 80286, 80386+) that can justify your statement. Quote:

1. I don't need to prove it when the PC market is very large.

2. Windows 3.1 sold over three million copies during the first three months of its release.. This is not including the pirate Windows 3.1 copies. Windows 3.0, 3.1 and 3.11 wasn't copy protected.

3. Windows 3.1 drops support for real mode, which drops 8086 support.

4. Windows 3.11 drops support for standard mode, which drops 80286 support.

5. From 1992, Intel has a 486 revenue majority.

From https://www.intel.fr/content/dam/doc/report/history-1994-annual-report.pdf
Intel reported the following
i. In 1994's fourth quarter, Pentium unit sales accounted for 23 percent of Intel's desktop processor volume.
ii. Millions of Pentiums were shipped.
iii. During Q4 1993 and 1994, a typical PC purchase was a computer featuring the Intel 486 chip.
iv. Net 1994 revenue reached $11.5 billion.
v. Net 1993 revenue reached $8.7 billion.
vi. Growing demand and production for Intel 486 resulted in a sharp decline in sales for Intel 386 from 1992 to 1993.
vii. Sales of the Intel 486 family comprised the majority of Intel's revenue during 1992, 1993, and 1994.
viii. Intel reached its 6 to 7 million Pentiums shipped goal during 1994. This is only 23 percent unit volume.

By the end of 1994, Intel's Pentium PC install base crushed the entire Amiga install base of 4 to 5 million units!

By the numbers, Intel's unified X86 PC platform is a monster compared to the Amiga i.e. it mirrored the USA's superpower might against the smaller German military during WW2.

6. Windows 95 sales, first year: 40 million, Windows 95 requires at least 386 and 4 MB RAM equipped PC. This is not including the pirate Windows 95 copies. 40 million of these PCs are Windows 3.11 capable.

AmigaOS 3.1's copy protection is via physical Kickstart ROM, hence it harder to pirate AmigaOS 3.1 when compared to Windows 3.1.



Quote:

Office applications worked also on Windows 3.0 / Real mode (8088/8086)...

Real mode (8086 and clones like NEC V30) has lower memory capability.

MS Excel 5.0 (1993) requires standard mode aka 286 protected mode.

Last edited by Hammer on 10-Mar-2025 at 07:21 AM.
Last edited by Hammer on 10-Mar-2025 at 07:08 AM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@cdimauro

Quote:

ARM is changing its policy and also entering the market with its own chips. So, it's ready to become a treat even to its current customers.

https://www.theregister.com/2025/02/06/arm_qualcomm_nuvia/
Arm gives up on killing off Qualcomm's vital chip license
"The British are coming, the British are coming ... to terms with their loss".

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

Jesus... another thread completly derailed by same age old story.. Guys can you please stop iccupying every thread with the same bickering.

It started so good with a possibilty of adding W3D to a game engine.. FFS!!!!

_________________

@Yssing

It's still doable. From what I've seen of the engine it's pretty much untextured polygons. All that is needed is to convert the fully transformed fixed point coordinates to floating point at the point of rasterization. While that will add some overhead, it's more than paid back by having the hardware render through polygon itself.

_________________
Doing stupid things for fun...

@Yssing

Quote:

Yssing wrote: Jesus... another thread completly derailed by same age old story.. Guys can you please stop iccupying every thread with the same bickering. It started so good with a possibilty of adding W3D to a game engine.. FFS!!!!

It was already long that there were OT posts and/or content on this thread, but you started complaining only now that I've written. Hypocritical!

@Yssing

Quote:

Yssing wrote: Jesus... another thread completly derailed by same age old story.. Guys can you please stop iccupying every thread with the same bickering. It started so good with a possibilty of adding W3D to a game engine.. FFS!!!!

Nothing is derailed. Warp support is definitely much higher in my ToDo list than supporting 040-060 (as we established in the other thread), because it doesn't look like it should take more than about a week.

But, right now, I am focused on finishing the darn game

Once the Warp library for PiStorm's VideoCore is released, I can start working on adding the Warp support.

I'm developing and testing the game myself under 1920x1080 in WinUAE, so I want the game to be playable in that res on an actual HW.

Vampire remains playable up to 960x540/1280x720 and I think I had a build which did ~10 fps in 1920x1080. If I introduced cockpit HUD, then it could totally be playable in that res even on Vampire (and without Maggie support, which I hope to get to later this year)

But, once the scanline traversal and pixel fill gets offloaded to Pi4's VideoCore, there's no reason it wouldn't do 60 fps at FullHD under Warp if it's doing ~77 fps (in the benchmark) in 320x200 now and that includes all rasterizing on CPU.

@Heimdall

Thanks for the update :)

_________________

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Still this MMU padding, which is a complete non-sense, since MAINSTREAM OSES HAVE NOT USED MMUS! I've capitalized it for YOUR benefit. But I doubt that a bot could get it. MS Windows 3.1 Enhance Mode has virtual memory you ignorant fool.

From which of my writings you've deducted that I didn't know that?

Don't worry, you don't have to answer: it's a rhetorical question.

Bot!
Quote:

Apple Computer first officially added support for virtual memory with the release of System 7 in 1991 you ignorant fool.

Same as above. BOT!

Do you know that there were Amiga applications which enabled virtual memory as well?
Quote:

Windows 3.1 uses a hardware feature of the 386 processor and its built-in Memory Management Unit (MMU).

Really? Ohhh... what a news!!!

But having and MMU is using it is NOT the same thing.
Quote:

Virtual 8086 (Address) Mode creates multiple instances of an 8086 processor with each having separate RAM.

Irrelevant. Plus, they have a separate VIRTUAL address space, but they can share the same physical memory.
Quote:

MMU is needed for "next gen" OS development.

Oh, really?

Could you please tell me when the 80286 and 80386 were commercialized and when Microsoft released Windows XP?

When the 68451 PMMU was commercialized and when Apple released MacOS X? When the first PowerPC was commercialized, and when Apple released MacOS X?

Just to get an idea of how long it took from the commercialization of a processor with some features, and when a modern OS (using them) was finally available to the end users.
Quote:

Intel 80376 doesn't have MMU and it was replaced by the much more successful 80386EX in 1994. 80376 doesn't have legacy 16bit and MMU i.e. it's Intel's 68020 counterpart based on 80386.

Irrelevant: see above.
Quote:

NEC V30's 8086 clone was pushed out of the market.

LOL
Quote:

EMM386.EXE requires 386 MMU since it emulates EMM.

Or course, IF you've run 8086 boxes which take advantage of the more available memory.
Quote:

EMM286.EXE requires 286 MMU since it emulates EMM.

286s extended memory could only be accessed in protected more, so you've extended memory in this case and not expanded memory.

286s have no VM8086 boxes, and cannot support expanded memory.

They can only run 8086 real mode or 286 protected mode.
Quote:

Windows 3.x has three modes of operation: Real Mode, Windows and the applications are subjected to the limitation of the 8086. This mode was abandoned by Windows 3.1. Standard Mode requires 286's protected mode. It's sayonara to NEC V30's 8086 clone. Enhanced Mode requires 386's protected mode. It is built on the 32-bit “Virtual Machine Manager� (VMM) that leverages the virtual 8086 mode to create and manage several virtual machines: one for the Windows Operating Environment, the System VM, and one for each DOS session. Enhanced Mode also introduces Virtual Device Drivers (VxDs): 32-bit protected mode drivers that can share system resources to more than one process.

Irrelevant: see above above.
Quote:

Quote: And here again mixing apples and oranges: Windows NT was NOT a mainstream OS! In the USA and for MS OS offerings, Windows NT was the mainstream OS for C3 / Orange book/TCSEC and POSIX compliance you ignorant fool.

STRA-LOL

https://en.wikipedia.org/wiki/Trusted_Computer_System_Evaluation_Criteria
Trusted Computer System Evaluation Criteria (TCSEC) is a United States Government Department of Defense (DoD) standard that sets basic requirements for assessing the effectiveness of computer security controls built into a computer system.
[...]
In the early 1970s, United States Air Force requirements for the development of new computer system capabilities were addressed to the Air Force Electronic Systems Division (ESD) later known as the Electronic Systems Center at Hanscom Air Force Base in Massachusetts.
[...]

Me: Windows NT was NOT a mainstream OS!
You: Windows NT was the mainstream OS for C3 / Orange book/TCSEC

You're a bot!
Quote:

Quote: The mainstream Windows version suffered of the same problems of ALL other mainstreams OSes, Amiga OS included. When Windows '95 was out I've created a simple 3 (THREE) bytes program which completely froze it: - open DEBUG.COM; - type "a 100"; - type "CLI"; - type "JMP 101"; - type "G"; - enjoy! 1. That's a red herring since that sequence is not a normal operation from a normal user.

It's not a red herring. You don't even know its meaning, and you're applying it with your usual non-sense.

That was PERFECTLY VALID application! Prove me the opposite, IF you're capable of!

BTW, it was a very simple DOS application. ANY DOS application could do something like that, bring the system completely down! And even Windows apps can do it.

You clearly have no clue at all of how Windows worked!
Quote:

2. Your example doesn't dismiss 386's MMU requirement for Windows 95.

That's only because Microsoft dropped all other execution modes and selected 386 in protected mode with PMMU usage.
Quote:

Memory protection is applied for userland Win32 apps.

False: see above!
Quote:

3. Windows 95 is NOT designed for C3 / Orange book/TCSEC compliance, hence Windows 95's memory protection coverage is less than NT's.

STRA-LOL. Again?!?

The context was ALL ABOUT MAINSTEAM OSES here! Windows NT was NOT a mainstream OS!
Quote:

Under Windows NT, DEBUG doesn't have direct access to hardware due to the extra abstraction layer. DEBUG doesn't have sector access for HDD and it has FDD sector access.

See above: OUT OF CONTEXT! You don't know how to follow a discussion and completely derail changing the context!
Quote:

Quote: Guess what: even with some good memory protections added to the system, Windows (read: NOT NT!) was forced to share part of the memory and I/O resources (which included interrupts handling), because it had to retain DOS-compatibility. And this continued until Windows XP. Release date for XP: 2001. Read again: 2001. 1. Windows 2000 (NT5.0, December 15, 1999 RTM) and Windows XP (NT 5.1) have very similar Games for Windows support as Windows XP e.g. Windows 2000 supports DirectX 7 to DirectX9c. Both Windows 2000 and XP don't support Windows Vista exclusive DirectX10.

Irrelevant.

How long it took to have a modern, mainstream OS? Care to give an answer?
Quote:

2. For Windows are based on the "NT", to run DOS (and 16-bit Windows) applications, NT-based versions of Windows use a "Virtual DOS Machine" (VDM) a full featured MS-DOS emulator and virtual machine with very limited access to real hardware. On DOS text UI mode, Doom DOS version will run on Windows NT 3.1 until Doom switches to the full screen mode. MS didn't add sufficient VGA emulation for VDM. There's a reason for Microsoft's push for DirectX and Games for Windows. NTVDM has DOS compatibility up to a certain level. NTVDM can run DOS-based Win16.

Irrelevant / PADDING.
Quote:

Windows 9x is needed for FULL DOS compatibility.

Not only for that: it's also for the Win16 and Win32S applications.
Quote:

Your DEBUG example is a DOS program that can run on Windows 95's VDM and NT's NTVDM. LOL

Right. And the context was... rolling drum... MAINSTREAM OSES!
Quote:

My university's 1997 PC fleet has Windows NT 4.0 (released in 1996) and we run Windows OpenGL / DirectX3a games on them in a large after hours LAN party. WinQuake and GLQuake was released in Jan 1997. University's PC fleet resource needs to be controlled.

Irrelevant / PADDING.
Quote:

Quote: Compatibility was so much important the Apple was forced on a similar situation (but worse than Microsoft with Windows), which prevented to get a robust and modern OS 'til MacOS X, as I've already written. Release date for OSX: 2001. . Read again: 2001. MMU is needed for "next gen" OS development and building the install base with MMU. Steve Job's NextStep was available for 68K with MMU, PPC and i386.

See above: HOW LONG IT TOOK? Answer!
Quote:

A desktop platform vendor must plan with a proper road map. Since Commodore didn't properly invest in AmigaOS's C3 / Orange book/TCSEC compliance R&D, Commodore has statements like running Windows NT on PA-RISC based Amiga Hombre.

So, Commodore was planning for a modern OS, right?
Quote:

Commodore management wasted OS development resource and time to benefit AT&T's Unix, not AmigaOS. Commodore didn't have "Amiga 1st" ideology. Apple has "MacOS 1st" ideology.

No, neither Apple had it. In fact, when Mac OS X was released? Answer!
Quote:

Quote: That was the common ground for all mainstream OSes, 'til the new millennium. 1. That's wrong for C3 / Orange book/TCSEC workplace requirement.

See above, bot!
Quote:

2. That's wrong for MMU install base build-up. Imagine in the year 1999 when MS Windows 2000 was released to manufactuerers (RTM) with near zero MMU install base. That's a Commodore level fuckup. Imagine MacOS X was released without an existing MMU install base... LOL.

Again, HOW LONG IT TOOK?!?
Quote:

A f__kup like Commodore.

Wasn't Commodore planning for using Windows NT?

You aren't even able to contextualize your own words, BOT!
Quote:

So, the MMU was NOT a requirement neither a concern.

MMU is a concern for the "next gen" OS development and building MMU install base.[/quote]
Again, HOW LONG IT TOOK?
Quote:

MMU is a concern for faking EMM compatibility instead of using hardware bank-switching. Faking memory expansion type is a concern for the PC.

A PMMU, to be more precise. A 286 had no PMMU and wasn't able to emulate the expanded memory.
Quote:

Quote: Yes, Windows 9x demanded it for the virtual memory, but just because it was already built-in on the 386, and it became a requirement. But for the software of the time, it wasn't a requirement. That's a red herring. My argument is not about minimum system requirements.

I don't care about YOUR argument: I only care about the context of the discussion.
Quote:

NEC V30 clone died since it didn't have 386's 32-bit protected mode.

SUPER-LOL
Quote:

For x86 cloners, AMD separates from NEC on x86 MMU mastery.

Really?!?
Quote:

https://hackaday.com/2024/02/09/breaking-through-the-1-mb-barrier-in-dos-with-unreal-mode-and-more/ This article builds on the first one which covered the older methods and covered the basics of protected mode. Where protected mode is convenient compared to real mode is that with the former the memory accesses go via the MMU and thus allows for access to 16 MB on the 80286 and 4 GB on the 80386. The segment descriptors and resolving of these that make this possible can be (ab)used on the 80286 and up by realizing that these segment descriptors are also used in real mode. Unreal mode is thus about switching to protected mode, loading arbitrary segment descriptors and switching back to real mode. As this is outside the original processor spec, it is commonly called ‘unreal mode’.

First of all, it's not outside of the processor spec: the so called "unreal mode" isn't a real execution mode, rather a perfectly valid way to set the processor segment/selector registers. Again, you report things that you've no clue at all, because you've never opened an x86/x64 architecture manual in your life.

Second, this proves that you can access all physical memory WITHOUT requiring an MMU.
Quote:

There was an EMM286.EXE and HIMEM.SYS (version 2.06 or later) using the 286 CPU’s MMU

See above: the 286 has NO PMMU, so it's NOT able to map pages for implementing the expanded memory.

HIMEM is used to use the "high-memory", up to 1MB + 64kB - 16 byte. Which was perfectly feasible for ALL 286+ processors WITHOUT switching to protected mode.

Again, you don't know of what you talk about.
Quote:

instead of any hardware bank-switching like most of the EMS cards of the 286’s era.

See above: 286 can't support expanded memory!
Quote:

For legacy support, faking memory expansion type is a concern on the PC.

Of course.
Quote:

You are wrong.

WHERE?!? PROVE IT!
Quote:

Reminder, Linux was born on 386 PC,

Irrelevant: it was NOT a Windows MAINSTREAM OS!
Quote:

not on an expensive MMU equipped Amiga with an install base of a few thousands!

And? Tell me when a MODERN, MAINSTREAM Windows OS arrived! WHEN?!?
Quote:

https://archive.computerhistory.org/resources/access/text/2013/04/102723262-05-01-acc.pdf DataQuest 1991, page 119 of 981 For 1992, 68EC020-16, $16.06 68EC020-25, $19.99 68020-25, $35.13 68EC030-25, $35.94 (missing MMU, not Unix capable, used in A4000/030) 68030-25, $108.75 (MMU equiped) 68040-25, $418.52 (MMU equiped) 68EC040-25, $112.50 (missing MMU and FPU, Commodore management rejected glue chips for Amiga) AM386-40, $102.50 (MMU equiped) R3000-25, $96.31 (MMU equiped) 386DX-25, $103.00 (MMU equiped) 80486SX-20, $157.75 (MMU equiped for Xenix 386, Windows NT, and Linux) 80486DX-25, $376.75 (MMU equiped) 80486DX-33, $376.75 (MMU equiped) 80486DX-50, $553.25 (MMU equiped) 80486DX2-50, $502.75 (MMU equiped) All 32-bit X86 CPUs have MMU. 68030-25 already lost "value for money" when compared to AM386-40 and MIPS R3000. Motorola's 68030-25 nearly copies Intel's 386DX-25 price guide. For 1993 Q1, 68EC020-16, $15.00 68EC020-25, $18.00 ($3 difference from EC020-16) 68020-25, $33.00 68EC030-25, $34.00 (missing MMU, not Unix capable, used in A4000/030) 68030-25, $95.00 (MMU equiped) 68040-25, $350.00 68EC040-25, $95.00 (missing MMU and FPU, Commodore management rejected glue chips for Amiga) 386DX-25, $65.00 (MMU equiped) 80486SX-20, $93.00 (MMU equiped for Xenix 386, Windows NT, and Linux) 80486DX-25, $288.00 (MMU equiped) 80486DX-33, $288.00 (MMU equiped) 80486DX-50, $375.00 (MMU equiped) 80486DX2-50, $385.00 (MMU equiped) AM386-40, $48.00 (MMU equiped) R3000-25, $87.00 (68040 level with MMU) A no brainer to why Motorola lost both desktops and workstation markets. Motorola positions MMU equiped 68K as a premium. Motorola was sleeping on the wheel when AMD lowered AM386-40's price to $48. Intel wasn't sleeping on the wheel.

Irrelevant. HAMMER's PADDING.
Quote:

For handhelds, Palm selects TI's ARM 925T with MMU to replace Motorola's MMU-less 68000 based Dragon Ball VZ. Motorola repeats the same market segment lost like the earlier desktops and workstation markets.

And again, you don't know of what you talk about.

In fact, you've no clue of how PalmOS worked. That's why... rolling drum... it have NOT used the MMU even on ARM.

You're an hopeless IGNORANT!

And a BOT!

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Prove it! IF you have numbers (for all hardware platforms supported by Windows 3.0: 8088/8086, 80286, 80386+) that can justify your statement. Quote: 1. I don't need to prove it

YOUR statement -> YOU have to prove it!
Quote:

when the PC market is very large.

Irrelevant. You're miserably trying to change the context.

You made a PRECISE statement, that you CANNOT prove. That's the point, BOT!
Quote:

2. Windows 3.1 sold over three million copies during the first three months of its release.. This is not including the pirate Windows 3.1 copies. Windows 3.0, 3.1 and 3.11 wasn't copy protected. 3. Windows 3.1 drops support for real mode, which drops 8086 support. 4. Windows 3.11 drops support for standard mode, which drops 80286 support. 5. From 1992, Intel has a 486 revenue majority. From https://www.intel.fr/content/dam/doc/report/history-1994-annual-report.pdf Intel reported the following i. In 1994's fourth quarter, Pentium unit sales accounted for 23 percent of Intel's desktop processor volume. ii. Millions of Pentiums were shipped. iii. During Q4 1993 and 1994, a typical PC purchase was a computer featuring the Intel 486 chip. iv. Net 1994 revenue reached $11.5 billion. v. Net 1993 revenue reached $8.7 billion. vi. Growing demand and production for Intel 486 resulted in a sharp decline in sales for Intel 386 from 1992 to 1993. vii. Sales of the Intel 486 family comprised the majority of Intel's revenue during 1992, 1993, and 1994. viii. Intel reached its 6 to 7 million Pentiums shipped goal during 1994. This is only 23 percent unit volume. By the end of 1994, Intel's Pentium PC install base crushed the entire Amiga install base of 4 to 5 million units! By the numbers, Intel's unified X86 PC platform is a monster compared to the Amiga i.e. it mirrored the USA's superpower might against the smaller German military during WW2. 6. Windows 95 sales, first year: 40 million, Windows 95 requires at least 386 and 4 MB RAM equipped PC. This is not including the pirate Windows 95 copies. 40 million of these PCs are Windows 3.11 capable.

Totally irrelevant, according to YOUR previous statement.
Quote:

AmigaOS 3.1's copy protection is via physical Kickstart ROM, hence it harder to pirate AmigaOS 3.1 when compared to Windows 3.1.

?!?
Quote:

Quote: Office applications worked also on Windows 3.0 / Real mode (8088/8086)... Real mode (8086 and clones like NEC V30) has lower memory capability.

And... WHO CARES?!?
Quote:

MS Excel 5.0 (1993) requires standard mode aka 286 protected mode.

Guess what: Microsoft change the Office requirement after that it dropped the support of the real mode...

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: ARM is changing its policy and also entering the market with its own chips. So, it's ready to become a treat even to its current customers. https://www.theregister.com/2025/02/06/arm_qualcomm_nuvia/ Arm gives up on killing off Qualcomm's vital chip license "The British are coming, the British are coming ... to terms with their loss".

You didn't get it and you went completely wrong.

This:

Arm to launch its own chip in move that could upend semiconductor industry
https://www.ft.com/content/95367b2b-2aa7-4a06-bdd3-0463c9bad008

was the context!

Very latest news: SoftBank (ARM's owner) just acquired Ampere Computing.

"The Japanese are coming, the Japanese are coming..."

@cdimauro

Quote:

So, it's ready to become a treat even to its current customers.

Well, that’s funny :)

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro Quote: So, it's ready to become a treat even to its current customers. Well, that’s funny :)

Indeed, with a missing h. :-/