@Karlos

Quote:

Karlos wrote: @cdimauro HAM is suitable for some games if you know how to use it. A poster on EAB has an example of a scrolling tile engine in which every tile left edge begins with a palette colour and HAM does the rest giving lots of shade and brightness variations. The fringing that would be visible on the left edge of the display is hidden behind a sprite status bar. An early tech demo of his engine is shown here: https://youtu.be/L1s5Vxb9BBY

Yes, and that's part of this:

"pre-computed animations"

But making games using BOBs which aren't surrounded of a one pixel border using one of the 16 primary colours (like in the video) is not possible due to the artifacts which are generated.

@cdimauro

One typical artifact is long lines, because start color is not visible, and other color channels is not reset, if you always did the RGBRGBRGBRGB pattern, you avoid long line artifacts. HAM is that different from sub pixels, except Pixels are BIG.
You get little bleeding but using hires or super hires to reduce this effect.

I guess another thing you can do, is draw different versions of the tiles,
with different cut of points.

You also avoid scrolling backgrounds,
you walk into a room, and into next and so on.

Last edited by NutsAboutAmiga on 17-Jun-2023 at 08:30 PM.
Last edited by NutsAboutAmiga on 17-Jun-2023 at 08:29 PM.

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

@cdimauro

Does it matter? For the style of game, the graphics are a perfect suit and you get to have your cake and eat it.

_________________
Doing stupid things for fun...

@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @cdimauro One typical artifact is long lines, because start color is not visible, and other color channels is not reset, if you always did the RGBRGBRGBRGB pattern, you avoid long line artifacts. HAM is that different from sub pixels, except Pixels are BIG. You get little bleeding but using hires or super hires to reduce this effect.

Hires and Super Hires aren't suited for games because they hog too much memory bandwidth plus, in this case, also space in memory. AGA games were usually in Lores because of that, and several times not even able to display 256 colours.

And this available only for AGA: OCS/ECS hadn't this possibility.
Quote:

I guess another thing you can do, is draw different versions of the tiles, with different cut of points.

That's a very difficult thing to achieve because you can have different colors at the beginning of each line in a tile. Even using 8x8 tiles (which is an absolute waste with planar graphics, as I've shown in one of my articles), you have to combine 8 rows of one tile with 8 rows of the following one: too much.
Quote:

You also avoid scrolling backgrounds, you walk into a room, and into next and so on.

This, added to the above, seriously limits the design of games and the work of artists.

@Karlos

Quote:

Karlos wrote: @cdimauro Does it matter? For the style of game, the graphics are a perfect suit and you get to have your cake and eat it.

That's ok for this specific game it works, but my statement was general (as above my answer to NutsAboutAmiga).

@cdimauro

Quote:

Hires and Super Hires aren't suited for games because they hog too much memory bandwidth plus

what if you draw it into fast memory, and copied only parts that is changed to chip memory.

I seen some HAM videos, to produce a HAM image, all you need is a fast CPU.

Last edited by NutsAboutAmiga on 18-Jun-2023 at 10:41 AM.
Last edited by NutsAboutAmiga on 18-Jun-2023 at 10:37 AM.

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

@Karlos

HAM was used in some card games, like strip poker pro.

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

@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @cdimauro Quote: Hires and Super Hires aren't suited for games because they hog too much memory bandwidth plus what if you draw it into fast memory, and copied only parts that is changed to chip memory. I seen some HAM videos, to produce a HAM image, all you need is a fast CPU.

Both weren't an option at the time: Amiga 1000 = 1985, Amiga 500/2000 = 1987.
Quote:

NutsAboutAmiga wrote: @Karlos HAM was used in some card games, like strip poker pro.

Obvious. :)

@NutsAboutAmiga
Quote:

I guess another thing you can do, is draw different versions of the tiles, with different cut of points. You also avoid scrolling backgrounds, you walk into a room, and into next and so on.

-=Hamulet=- Demo

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

@pixie

looks pritty good, is this on stock hardware, or does it need some fast ram, etc.

Last edited by NutsAboutAmiga on 18-Jun-2023 at 10:56 PM.

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

@NutsAboutAmigaQuote:

@pixie looks pritty good, is this on stock hardware, or does it need some fast ram, etc.

The author said regarding the original techdemo that it would run on a stock Amiga, I guess this shouldn't be much different otherwise he would have stated on the release of the game.
Quote:

It runs on OCS, and requires only a 512KB chip ram Amiga, so normally any plain Amiga 500/2000, or Amiga 1000 with 512KB should work Techdemo of my game :: eab.abime.net

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

@Karlos

Quote:

I think "control words" might be a bit grand. Ok, the upper bitplane pair decides how the remaining bitplanes are to be interpreted. You can still do that in chunky. It's not really necessary to separate kr out.

Oh, yeah I meant control bits, since it's a few bits and not a word. But with my idea it's not the usual bitplane approach but having a control plane and colour plane. Both planes would contain packed data. With HAM6 control plane would have 2-bit values packed with colour plane containing 4-bit colour values for palette or RGB modifying. HAM8 could have been extended this way without loosing 2 bits of colour for modification. The control plane could be the same as HAM6 but colour plane would use 8 bits instead for full 256 colour palette and 8-bit RGB modification. Though, HAM6 could be called HAM4 since it would be 4 bit HAM, and the new HAM8 a full 8-bit HAM. This is similar to some ideas they had for RGB planar in AAA with three planes of packed RGB.

Quote:

For MC64K, I did make a chunky ham mode that has only 5 bits per gun. However, it has 3 control bits that allow any of the guns to be set simultaneously, e.g. 000xxxxx gives a palette entry x (from a 15 bit RGB) and 111xxxxx sets red green and blue to x. Thus you get all possible 32 grey scale shades for free, without having to use a palette entry.

That's interesting and looks efficient. So just to understand, what about 1 to 6 control data? Would something like upper 100, 010 or 001 set R, G or B, allowing to set a single gun or combination of guns to the lower 5 bits?

@Hammer

Quote:

After my Dad bought ex-corporate IBM Model 55SX (386SX-16,x 387 FPU, 5 MB RAM, onboard IBM VGA) in 1990, IBM VGA is slow for Amiga 500 era games.

VGA had no blitter. But it had hardware scrolling. So simple single playfield 2d games should have been fine.

Quote:

VGA's 256 color mode is 8-bit.

Oh I meant the colour resolution. 6 bit RGB. A bit like 262,144 colours of HAM8 being the full palette range of VGA.

Quote:

256-color VGA games ran fine on MCGA as long as they stay to the basic 320×200 256-color mode (from a palette of 262,144, Mode 13h) and didn't attempt to use VGA-specific features such as multiple screen pages.

I see that. VGA is mostly known for it. But it did exist in MCGA.

MCGA could be confusing. Looks like it could mean Monochrome Colour Graphics Adapter. Yes VGA needed to fix that confusion.

@Hypex

Quote:

VGA had no blitter. But it had hardware scrolling. So simple single playfield 2d games should have been fine.

they got that in VESA in 1996, better late, then never

See: "VBE/accelerator functions (VBE/AF) (August 1996)"

https://en.wikipedia.org/wiki/VESA_BIOS_Extensions

Last edited by NutsAboutAmiga on 19-Jun-2023 at 07:51 PM.
Last edited by NutsAboutAmiga on 19-Jun-2023 at 07:48 PM.

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

@Hypex

Regarding the chunky HAM I described, think of it as RGBxxxxx:

000 xxxxx - use palette colour xxxxx
001 xxxxx - hold previous and set blue to xxxxx
010 xxxxx - hold previous and set green to xxxxx
011 xxxxx - hold previous and set green and blue both to xxxxx
100 xxxxx - hold previous and set red to xxxxx

You get the picture. By the time you have 111 xxxxx, you are setting all 3 channels to xxxxx, e.g. setting grey.

This mode is coupled to an RGB 555 model, since the colour resolution is 5 bits let gun.

_________________
Doing stupid things for fun...

@Hypex

Quote:

Hypex wrote: @Hammer Quote: After my Dad bought ex-corporate IBM Model 55SX (386SX-16,x 387 FPU, 5 MB RAM, onboard IBM VGA) in 1990, IBM VGA is slow for Amiga 500 era games. VGA had no blitter. But it had hardware scrolling. So simple single playfield 2d games should have been fine. Quote: VGA's 256 color mode is 8-bit. Oh I meant the colour resolution. 6 bit RGB. A bit like 262,144 colours of HAM8 being the full palette range of VGA. Quote: 256-color VGA games ran fine on MCGA as long as they stay to the basic 320×200 256-color mode (from a palette of 262,144, Mode 13h) and didn't attempt to use VGA-specific features such as multiple screen pages. I see that. VGA is mostly known for it. But it did exist in MCGA. MCGA could be confusing. Looks like it could mean Monochrome Colour Graphics Adapter. Yes VGA needed to fix that confusion.

1987 era IBM 8514 has Blitter and fixed-function 2D acceleration and it was cloned e.g. ATI Mach 8 / Mach 32, Matrox MG-108, Paradise Systems' Plus-A / Renaissance Rendition II, Desktop Computing's AGA 1024 (also clone TIGA capable), and 8514ish Tseng Labs ET4000.

IBM created VGA and 8514 in 1987 and it was cloned.

Amiga OCS was price-performance competitive between 1985 to 1990.

IBM didn't kill the Amiga, lower-cost PC clones killed Amiga from the desktop computer gaming scene.

Lenovo is the world's largest PC vendor which shows IBM's incompetence and bloated business administration.

_________________
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

@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @Hypex Quote: VGA had no blitter. But it had hardware scrolling. So simple single playfield 2d games should have been fine. they got that in VESA in 1996, better late, then never See: "VBE/accelerator functions (VBE/AF) (August 1996)" https://en.wikipedia.org/wiki/VESA_BIOS_Extensions

Beyond DOS VGA, PC desktops "designed for Windows 3.x" 2D accelerators tookover for desktop PC apps i.e. Microsoft's retargetable graphics path.

DOS 1996-era Duke Nukem 3D and Quake can use VESA's VBE/AF.

Following Doom95 (released in late 1995), WinQuake was released in 1997 and run on Windows 95's DirectDraw.

Last edited by Hammer on 20-Jun-2023 at 03:59 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:

Nevertheless, the question is always the same: those comparisons as totally irrelevant and useless. Because such monster CPUs weren't available at the time.

For PlayStation 1/ PC VGA 320x200p gaming, C= AGA framebuffer is good enough. The missing factor in math compute power.

About 100 MIPS compute math power would be needed against Sony's PS1 (mostly integer-based games console). This is for the late 1994 release and 3rd party developer program in late 1993.

Dave Haynie argued for DSP3210 (FP40 / FP32 / INT16 / INT24/ INT32) math compute power with AGA bundle. DSP3210 has about 33 MFLOPS FP32 @ 66 Mhz.

Both David Pleasance and Dave Haynie are aware of Amiga's uncompetitive math compute power.

AAA alone will not solve Amiga's uncompetitive math compute power problem.

----

For 1993-era Doom-type gaming, 68030 @ 40 Mhz with AGA or 386DX-33 with fast VGA is the reasonable gameplay minimum.

Commodore UK has recognised this fact with CPU accelerated A1200 bundle proposal and Commodore International rejected it. Amiga's 3rd party CPU accelerator vendors do not have Commodore's economy of scale.
This is for the 1993 context.

Quote:

Besides that, HAM mode is NOT suitable for games due to the graphical artifacts which it generates.

Think of JPEG/MPEG artifacts and "compressed pixels" in modern times.

Without a fast CPU, HAM mode is useful for Links Golf, Strip Poker, Porn, 256 color VGA adventure games, and quick time event (QTE) Time Gal type video games.


Quote:

Plus, it's limited to 16 color shades, whereas VGA offered 64 shades since 1987 and that allowed to reproduce more realistic colors (even using just 256 colors. But it was possible to change some colors each scan line by reprogramming some RAMDAC's entries, like the Copper allowed with the Amiga chipset. However I never tried and I don't know how many of them can be changed during the horizontal blank period).

1987 IBM VGA doesn't have access to 1993 Pentium class CPUs and it's a severe bottleneck for any fast X86 CPUs.

IBM VGA is a crap retro VGA.

Quote:

Now one question to you: why you became a PiStorm evangelist, since it lacks 68K's PMMU and 80bit FPU precision... exactly like the Apollo core's 68080 which you continuously bashed for exactly the same things?

It depends on the context e.g. running Amiga pro-apps / Mac soft-VM that needs 80-bit precision FP or running games that needs FP32 e.g. 3DNow (packed FP32 from 64-bit SIMD) enhanced Quake II.

X86-64 CPUs still have FP80 support.

Furthermore, PiStorm and Raspberry Pi 3a+ is a low-cost solution for very fast soft 68040 (via Emu68) and P96 RTG accelerator.

Apollo core's 68080 V2 is below the standard FP64 threshold i.e. FP52 (Ref 1)
68080 V4 has support for FP64.

"While V2 first had softfloat, then hard fpu but limited precision (iirc 52bits), V4SA also didn't have 64bit from beginning, only 56."

ARM Cortex A53 supports FP64.

PiStorm-Emu68 setup can brute force soft FP80 above any 68K hardware implementations.

With 128-bit SIMD support, two 64-bit integers can fake FP80.

Reference
1. https://amitopia.com/gold-2-11-is-out-for-68080-amiga-accelerator-users/

Last edited by Hammer on 21-Jun-2023 at 01:55 AM.
Last edited by Hammer on 21-Jun-2023 at 01:54 AM.
Last edited by Hammer on 20-Jun-2023 at 05:36 AM.
Last edited by Hammer on 20-Jun-2023 at 05:31 AM.
Last edited by Hammer on 20-Jun-2023 at 05:25 AM.
Last edited by Hammer on 20-Jun-2023 at 05:20 AM.
Last edited by Hammer on 20-Jun-2023 at 05:16 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

@pixie

Quote:

pixie wrote: @NutsAboutAmiga Quote: I guess another thing you can do, is draw different versions of the tiles, with different cut of points. You also avoid scrolling backgrounds, you walk into a room, and into next and so on. -=Hamulet=- Demo

@pixie

Quote:

pixie wrote: @NutsAboutAmigaQuote: @pixie looks pritty good, is this on stock hardware, or does it need some fast ram, etc. The author said regarding the original techdemo that it would run on a stock Amiga, I guess this shouldn't be much different otherwise he would have stated on the release of the game. Quote: It runs on OCS, and requires only a 512KB chip ram Amiga, so normally any plain Amiga 500/2000, or Amiga 1000 with 512KB should work Techdemo of my game :: eab.abime.net

Thanks for it. It was nice to see this game and the video.

I've written an article about HAM mode and games, taking Hamulet as an example: Amiga in modalità HAM: gioia per gli occhi, ma per pochi giochi
It's in Italian, but it can be easily translated with DeepL, Google Translate, ...

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Nevertheless, the question is always the same: those comparisons as totally irrelevant and useless. Because such monster CPUs weren't available at the time. For PlayStation 1/ PC VGA 320x200p gaming, C= AGA framebuffer is good enough.

Not at the time, because on '92 PC's graphic cards were much more powerful and with more features. Plus, they had the packed more already implemented, which was important. Last but not really least, they had processors with better performances (even the 286 was much better, especially for games).

BTW, Playstation arrived later.
Quote:

The missing factor in math compute power.

Performance, in general.
Quote:

About 100 MIPS compute math power would be needed against Sony's PS1 (mostly integer-based games console). This is for the late 1994 release and 3rd party developer program in late 1993.

The PS1's CPU didn't reached 100MPS, but much less. I assume that you talked of the graphics processor.
Quote:

Dave Haynie argued for DSP3210 (FP40 / FP32 / INT16 / INT24/ INT32) math compute power with AGA bundle. DSP3210 has about 33 MFLOPS FP32 @ 66 Mhz. Both David Pleasance and Dave Haynie are aware of Amiga's uncompetitive math compute power.

Why a DSP processor which is totally aliens to the Amiga ecosystem, when you could have had better processor (68040) which was useful with all Amiga software (like x86 processors for PCs)?
Quote:

AAA alone will not solve Amiga's uncompetitive math compute power problem.

AAA would have been too late.
Quote:

---- For 1993-era Doom-type gaming, 68030 @ 40 Mhz with AGA or 386DX-33 with fast VGA is the reasonable gameplay minimum. Commodore UK has recognised this fact with CPU accelerated A1200 bundle proposal and Commodore International rejected it. Amiga's 3rd party CPU accelerator vendors do not have Commodore's economy of scale. This is for the 1993 context.

Whatever: Commodore hasn't provided machines with better performances. And hasn't supported developers.
Quote:

Quote: Besides that, HAM mode is NOT suitable for games due to the graphical artifacts which it generates. Think of JPEG/MPEG artifacts and "compressed pixels" in modern times.

Modern is different. Now we were talking about mid '80s to mid '90s.
Quote:

Without a fast CPU, HAM mode is useful for Links Golf,

Have you tried? I don't think so. I mean: with a plain Amiga 500 (1MB max).
Quote:

Strip Poker,

Easy to make it using HAM.
Quote:

Porn,

Quote:

256 color VGA adventure games,

Not suitable: too many characters displayed in some scenes, with characters also being larger than 16 pixel.

Sprites cannot cover this. And BOBs can't be used due to the fringing the HAM generates.
Quote:

and quick time event (QTE) Time Gal type video games.

Movies require a lot of storage: CDRom need to be used.
Quote:

Quote: Plus, it's limited to 16 color shades, whereas VGA offered 64 shades since 1987 and that allowed to reproduce more realistic colors (even using just 256 colors. But it was possible to change some colors each scan line by reprogramming some RAMDAC's entries, like the Copper allowed with the Amiga chipset. However I never tried and I don't know how many of them can be changed during the horizontal blank period). 1987 IBM VGA doesn't have access to 1993 Pentium class CPUs

Unbelievable: a 1987 computer that cannot access a 1993 CPU!
Quote:

and it's a severe bottleneck for any fast X86 CPUs.

Seriously?
Quote:

IBM VGA is a crap retro VGA.

Ah, again the same mantra. You're a broken record. But repeating always the same thing will NOT make them true.
Quote:

Quote: Now one question to you: why you became a PiStorm evangelist, since it lacks 68K's PMMU and 80bit FPU precision... exactly like the Apollo core's 68080 which you continuously bashed for exactly the same things? It depends on the context e.g. running Amiga pro-apps / Mac soft-VM that needs 80-bit precision FP or running games that needs FP32 e.g. 3DNow (packed FP32 from 64-bit SIMD) enhanced Quake II.

There was no "context" before: you continuously bashed and hit the Vampire because of its FPU.

Now you completely changed your mind and became a PiStorm evangelist, which suffers from exactly the same issues.

You're not trustable.
Quote:

X86-64 CPUs still have FP80 support.

Right. Which is not used by WinUAE by default, BTW.
Quote:

Furthermore, PiStorm and Raspberry Pi 3a+ is a low-cost solution for very fast soft 68040 (via Emu68) and P96 RTG accelerator.

It doesn't matter: see above. Your wasn't a cost problem when you bashed the Vampire: it was because it lacked an 80-bit FPU. Full stop.
Quote:

Apollo core's 68080 V2 is below the standard FP64 threshold i.e. FP52 (Ref 1) 68080 V4 has support for FP64. "While V2 first had softfloat, then hard fpu but limited precision (iirc 52bits), V4SA also didn't have 64bit from beginning, only 56."

Sure and?
Quote:

ARM Cortex A53 supports FP64. PiStorm-Emu68 setup can brute force soft FP80 above any 68K hardware implementations. With 128-bit SIMD support, two 64-bit integers can fake FP80.

PADDING...
Quote:

Reference 1. https://amitopia.com/gold-2-11-is-out-for-68080-amiga-accelerator-users/

So, it's since 5 (FIVE) years that the 68080 support a 64-bit FPU.

Anyway, it doesn't matter: see above. You bashed the Vampire for this and now you incensate the PiStorm which has exactly the same issues. You're not coherent. Rather, the opposite.

@cdimauro

Quote:

Not at the time, because on '92 PC's graphic cards were much more powerful and with more features. Plus, they had the packed more already implemented, which was important. Last but not really least, they had processors with better performances (even the 286 was much better, especially for games).

For 1992-1993 PC DOS gaming, it's mostly a fast VGA 320x2x0 frame buffer.

Quote:

BTW, Playstation arrived later.

PlayStation was attracting 3rd party developers around late 1993.

On 27 October 1993, Sony publicly announced that it was entering the game console market with the PlayStation.


A team from Epic Sony visited more than a hundred companies throughout Japan in May 1993 in hopes of attracting game creators with the PlayStation's technological appeal.[58] Through a series of negotiations, Sony acquired initial support from Namco, Konami, and Williams Entertainment, as well as 250 other development teams in Japan alone. Namco in particular was keen to participate in the PlayStation project as a third-party developer since Namco rivalled Sega in the arcade market.[59] Attaining these companies secured influential games such as Ridge Racer (1993) and Mortal Kombat 3 (1995), Ridge Racer being one of the most popular arcade games at the time, and it was already confirmed behind closed doors that it would be the PlayStation's first game by December 1993.

Namco's research managing director Shegeichi Nakamura met with Kutaragi in 1993 to discuss the preliminary PlayStation specifications, with Namco subsequently basing the Namco System 11 arcade board on PlayStation hardware and developing Tekken to compete with Virtua Fighter. The System 11 launched in arcades several months before the PlayStation's release, with the arcade release of Tekken in September 1994

Namco's PlayStation-based System 11 was released several months before the PlayStation's release.

PlayStation hardware was operational before the Japanese retail Q4 1994 release. Sony made sure Playstation has AAA games during the retail release.

Your argument is not based on 3rd party developer experience for PlayStation.

Quote:

The PS1's CPU didn't reached 100MPS, but much less. I assume that you talked of the graphics processor.

PS1's CPU has 33 MIPS and the graphics co-processor has 66 MIPS.

100 MIPS requirement is from a gaming PC Pentium perspective.

Quote:

Not suitable: too many characters displayed in some scenes, with characters also being larger than 16 pixel. Sprites cannot cover this. And BOBs can't be used due to the fringing the HAM generates.

Refer to Amiga's Links Golf game. Direct adventure PC VGA game ports that use Amiga HAM mode will require hard disk storage.

HAM mode has certain color artwork rules to minimize fringing.

Quote:

Unbelievable: a 1987 computer that cannot access a 1993 CPU!

Amiga's decline started around 1991. Wing Commander for PC was its "Defender of the Crown" moment and Doom's 1993 release reinforced the gaming PC's position.


Quote:

There was no "context" before: you continuously bashed and hit the Vampire because of its FPU. Now you completely changed your mind and became a PiStorm evangelist, which suffers from exactly the same issues.

Apollo Core's AGA promise for Vampire V2 FAILED.

68080 V2 is stuck in a 52-bit floating point.

68080 V4 initially didn't have a 64-bit floating point.

Quote:

You're not trustable.

PADDING.

Quote:

So, it's since 5 (FIVE) years that the 68080 support a 64-bit FPU.

For "FPU Precision On V4SA" topic
http://www.apollo-core.com/knowledge.php?b=4¬e=34161
Gunnar von Boehn: R5 release supports 56 Bit FPU precision.
Posted 14 Feb 2021 06:15

http://apollo-core.com/knowledge.php?b=1¬e=34946
V4SA R6 was released on 18th of April 2021 HAHAHAHA

For Gold 2.7 core for all Vampires 500, 500+ and 600
https://forum.amiga.org/index.php?topic=73325.0
Posted date: March 04, 2018

Other forums have already reported that the FPU implementation doesn't work properly.

For example, launch SYS:Tools/Calculator and try to calculate 5 * 1.3, and the result will be 6.499999999069.

More about issues in the FPU implementation over here:
https://blog.alb42.de/2018/03/03/vampire-2-7-fpu-part-2/

This FPU is just "Quake enabler". HAHAHAHAHA


Posted: March 05, 2018
Kolla: As I commented, it's not really an FPU as much as a Quake enabler - any hope for a proper FPU on V2, is long gone. In many cases, FEMU works better (more accurate) than the FPU of core 2.7, but sadly it has its issues and development has stopped. So, another FPU emulator is needed for V2.

HAHAHAHAHA

You're not trustable. Look in the mirror.

PiStorm32 Lite and Raspberry Pi 4 is cheap.

Fact: There is no sales contract for zero-cost open-source software like Emu68.

I did have a Natami account.

I'm aware of the FP80 issues with certain software that runs on the Amiga, MacOS and Vampire AC68080 e.g. http://apollo-core.com/knowledge.php?b=5¬e=24818&order=&x=3&z=NHLVYf

PiStorm32 Lite + RPI 4B (ARM Cortex A73) + Emu68 can brute force softFPU Shapeshifter that beats Power Macintosh 6100/60 FPU performance


https://eab.abime.net/showpost.php?p=1605865&postcount=91

A1200/Vamp v2 Coffin56/Gold2.12fw - 1526s (25m 26s)
Amiga4000D 3.2.1 WarpEngine060@96MHz - 1477s (24m 37s)
Vampire V4SA+ Coffin57/SA_8435.jic (x14) - 1405s (23m 25s)
Amiga4000T 3.2.1, BFG9060@100MHz - 1358s (22m 38s)
A1200/PiStorm32 w/Pi4b+ @2.2GHz.CaffeinOS - 166secs
X5000/20 2GHz MorphOS 3.18 - 119secs

MACPRO5,1 2010 2x3.5GHz 6core. FSUAE/Coffin on BigSur - 80secs

SGI FUEL 800MHz / IRIX 6.5.30f - 65s (62 without render preview turned on)
SGI 4x1GHz (1 THREAD) - 53s (50s no preview)
SGI 4x1GHZ (MAX THREADS) - 15s (14s no preview)

Last edited by Hammer on 23-Jun-2023 at 02:30 AM.
Last edited by Hammer on 23-Jun-2023 at 02:16 AM.
Last edited by Hammer on 23-Jun-2023 at 02:13 AM.
Last edited by Hammer on 23-Jun-2023 at 01:53 AM.
Last edited by Hammer on 23-Jun-2023 at 01:48 AM.
Last edited by Hammer on 23-Jun-2023 at 01:39 AM.
Last edited by Hammer on 23-Jun-2023 at 01:35 AM.
Last edited by Hammer on 23-Jun-2023 at 01:29 AM.
Last edited by Hammer on 23-Jun-2023 at 01:24 AM.
Last edited by Hammer on 23-Jun-2023 at 01:18 AM.
Last edited by Hammer on 23-Jun-2023 at 01:16 AM.
Last edited by Hammer on 23-Jun-2023 at 01:14 AM.
Last edited by Hammer on 23-Jun-2023 at 01:13 AM.
Last edited by Hammer on 23-Jun-2023 at 12:48 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