@kolla

Quote:

kolla wrote: A cool thing about the DSP 3210 is that it actually exists, on the AA3000 and AA3000+ motherboards, with software available on Aminet. Which is than what can be said about many other things thrown up here.

Dave Haynie's DSP 3210 with AGA config died due to internal corporate politics.

DSP 3210 (16 MIPS INT32, 33 MFLOPS FP32 at 66 Mhz) with AGA is just a stopgap that nearly rivals early 80486 / 68040 class CPUs in integer (INT32) and Pentium 60 class FPU (~30 MFLOPS for FP32 workloads).

Nintendo's Super FX2 can do 20 MIPS INT16 for geometry and pixel workloads.

DSP 3210's INT32 ALU can cover integer-based pixel workloads while FP32 FPU can cover geometry workloads.

By 1995, gaming PC's fast SVGA frame buffers (operating in VGA's typical gaming resolution) were backed by Pentium 75/90/100/120 class CPUs. In 1996, Pentium evolved into 133/150/166/200 SKUs, and Quake was released.

3DO was released in October 1993 in North America and June 1994 in the EU.
3DO has supply problems during Xmas Q4 1993 to Q2 1994. EA extensively backed the 3DO platform.

Amiga CD32's specs are a dead duck in 1993 and it has its own supply problems in 1993.

Sony's PlayStation was released in Xmas Q4 1994 in Japan and Q4 1995 in North America and the EU. The PlayStation has strong exclusive game titles and a strong hardware supply chain for 1995.

For the 1990s game market, both Atari and Commodore's senior management are out of touch.

David Pleasance (Commodore UK MD, pushed for CPU accelerated A1200 bundles) and Dave Haynie (pushed for math co-processor accelerated AGA bundles, chief engineer at Commodore International) are aware of the competition, but Commodore International's senior management was out of touch.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@cdimauro

Quote:

So? It was NOT the ONLY VGA card which was available at the time. AND, which is even worse, you're comparing it (and only this card) with the AGA chipset which arrived FIVE YEARS LATE (when SVGAs were very common). As I've said before, you mix-up two completely different contexts: a total non-sense!

There are many slow VGA / SVGA clones. Read https://thandor.net/benchmark/32

The PC market is very large with sufficient hardware specs to play Doom.
https://doom.fandom.com/wiki/Sales
Hence, the PC versions of Doom and Doom II have likely sold over 4 million copies combined

For Doom 1993 Xmas sales, if a gaming PC user has 386DX-33 and above CPU, the user can purchase a faster VGA-compatible clone.. You can't do this for A3000/030, A3000T/040, or any full 32-bit 68020/68030/68040 CPU accelerated Amiga OCS/ECS due to missing RTG infrastructure.

Unlike the fully enabled 32-bit 386DX/486SX/486DX CPU-equipped PCs, the Amiga gaming scene from 1988 to 1994 was NOT able to build upon the existing full 32-bit 68020/68030/68040 CPU accelerated Amiga OCS/ECS install base due to missing RTG infrastructure until CyberGraphX's 1995 release, but that's too late for 1993-1994 time period.


For example
https://archive.org/details/bub_gb_hqQJaNzN9IcC/page/n603/mode/2up
PC Mag 1992-08, page 604 of 664,
Diamond Speedstar 24 (ET4000AX ISA) has $169 USD.

The real answer is, it depends on the frame buffer's implementation.

Removing the CPU bottleneck factor, AGA dumb framebuffer's 42 fps Quake demo3 320x200 results are not bad.

Your argument is nonsense.

Last edited by Hammer on 06-Jun-2023 at 05:11 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: So? It was NOT the ONLY VGA card which was available at the time. AND, which is even worse, you're comparing it (and only this card) with the AGA chipset which arrived FIVE YEARS LATE (when SVGAs were very common). As I've said before, you mix-up two completely different contexts: a total non-sense! There are many slow VGA / SVGA clones. Read https://thandor.net/benchmark/32

I already knew them, thanks.
Quote:

The PC market is very large with sufficient hardware specs to play Doom. https:ED//doom.fandom.com/wiki/Sales Hence, the PC versions of Doom and Doom II have likely sold over 4 million copies combined

Hammer's PADDING.
Quote:

For Doom 1993 Xmas sales, if a gaming PC user has 386DX-33 and above CPU, the user can purchase a faster VGA-compatible clone. .

Only if you want a very good frame rate. Cheaper cards can give a decent frame rate. Of course, if you buy the cheapest then the experience is bad.
Quote:

You can't do this for A3000/030 or any full 32-bit CPU accelerated Amiga OCS/ECS due to missing RTG infrastructure. The Amiga gaming scene wasn't able to build upon the existing full 32-bit 68020/68030/68040 CPU accelerated Amiga OCS/ECS install base due to missing RTG infrastructure until CyberGraphX's 1995 release, but that's too late for 1993-1994 time period.

First, you clearly don't know how most of the Amiga games were developed at time. Hint: by killing the o.s..
So, it could have been easily possible to develop games using packed/chunky graphics if a (new) Amiga chipset had it in hardware, even if the o.s. had zero support.

Second, Commodore was already working on adding RTG support (for Amiga o.s. 4) and this could have been different from CyberGraphics or Picasso, and easier to implement. To be more clear, adding RTG support to the Amiga o.s. could have been much simpler and without requiring the big infrastructure introduced by CG or Picasso. Hence: it could have been in time for releasing the o.s. AND the packed/chunky chipset, and enable o.s.-friendly 3D games.
Quote:

For example https://archive.org/details/bub_gb_hqQJaNzN9IcC/page/n603/mode/2up PC Mag 1992-08, page 604 of 664, Diamond Speedstar 24 (ET4000AX ISA) has $169 USD.

Which can deliver from 27 to 32 FPS, according to the above link. So, it's a top class ISA SVGA card and quite cheap for what it gives.
Quote:

The real answer is, it depends on the frame buffer's implementation.

As already said, you discovered the hot water.
Quote:

Your argument is nonsense.

And... this is related to what? Can you quote me and PROVE where this "nonsense" is coming from?

EDIT. You added other stuff in the meanwhile:
Quote:

Removing the CPU bottleneck factor, AGA dumb framebuffer's 42 fps Quake demo3 320x200 results are not bad.

Again, that's your WISHFUL THINKING: NOTHING like that was achievable AT THE TIME.

You live on a parallel universe and need a good one, as I've already said.

Last edited by cdimauro on 06-Jun-2023 at 05:18 AM.

@g01df1sh

Quote:

g01df1sh wrote: Whats the point there is already Aros. If they are going to port it again port to Arm

AROS is available for ARM as well, albeit the port isn't complete.

Anyway, porting OS4 to some other hardware platform isn't possible due to lack of resources. Not even the PowerPC port was completed, BTW.


@saimon69

Quote:

saimon69 wrote: @g01df1sh I agree on principle but unfortunately you can count actual AROS devs on a single hand -_-

There are some, looking at the commits. But it's true that most patches come from a single developer.


@V8

Quote:

V8 wrote: @saimon69 Quote: I agree on principle but unfortunately you can count actual AROS devs on a single hand -_- True, but it is not like OS4 or MorphOS have more developers either. Amiga and amiga-like systems are pretty nieche. There are honestly not very many people that want to use them and even less people willing to develop for them. In particular this applies to the three NG platforms.

There's really nothing "NG" on all those Amiga-/like o.ses: they were / are ports or reimplementations of the original one or its APIs.

Even the ridiculous stack-to-be-set-high-enough-otherwise-the-app-crashes "feature" (SIC!) is still there...
Quote:

At least AROS is open source and runs on x86-64 so there is that. The only flavor of Amiga that seems to be actually alive and has a lot of developers for both HW and SW and heaps of excited users is the original commodore ks3.1 running on real amigas or new hw/fpgas. That is where it seems all the development and excitement is happening nowadays. 68k Amiga is the only thing that is really active today and totally eclipses all three other flavors combined by orders of magnitude in pretty much any metric, number of users, developers, hw, activity...

That's normal / expected: Amiga IS 68K (AND its chipset).

The above o.ses are imitations which have only some part in common.

@ppcamiga1

Classic Amiga platforms have access to relatively low-cost ARMv8-based PiStorm / PiStorm32 Lite and Raspberry Pi 3A+ to 4B.

Amiga 500+ ECS clone Minimig VER 1.97 ITX can accept ARMv8-based PiStorm-Emu68 CPU-RTG accelerator like a real Amiga 500/500+.

Modern PC still has legacy ISA support via South Bridge's LPC bus just as a single FPGA Amiga chipset clone is being used like PC's South Bridge legacy support.

Commodity X86-64 is not the only option.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

arm in 2023 is also pc. pistorm is just emulator on pc connected to Amiga.
which makes Amiga keyboard and mouse interface to pc.
it is stupid because amiga keyboard and mouser where copied from pc 40 years ago.
get lost with this shit. real classic amiga use real cpu and not cpu from pc.

@cdimauro

Quote:

Hammer's PADDING.

It's for context i.e. minority gaming PC hardware specs can still generate sufficient ROI for PC game developers due to the large PC market size.

For PC games vs PC install base, the PC's attachment rate is terrible.

Quote:

Only if you want a very good frame rate. Cheaper cards can give a decent frame rate. Of course, if you buy the cheapest then the experience is bad.

For Doom's 1993 Xmas, the PC has superior market access and price range coverage.

Replacing the slow VGA card on a typical 386DX/486 PC clone is not hard.

Quote:

First, you clearly don't know how most of the Amiga games were developed at time. Hint: by killing the o.s.. So, it could have been easily possible to develop games using packed/chunky graphics if a (new) Amiga chipset had it in hardware, even if the o.s. had zero support.

Hint: Doom DOS also "hit the metal". The difference is PC VGA standard has been cloned and improved VGA clones were released.

SVGA clone's advanced features are accessible via Windows drivers.

The PC clone market was able to withstand the exit of IBM and Intel's multiple attempted killings of X86 from iAPX 432 to i860/i960 to IA-64. AMD's second source existence made sure Intel wasn't able to kill X86 for IA64 Itanium like Motorola killed 68K for PowerPC.

The clone Amiga chipsets didn't exist in 1992 and 1994, hence when Commodore died, the Amiga platform died as a mainstream gaming platform. In modern times, FPGA Amiga chipset clones are surviving without Commodore / Escom.

The clone 68K CPU market wasn't mature enough to survive Motorola's 68K exit from desktop markets. In modern times, FPGA 68K CPU clones are surviving without Motorola/Freescale.

PiStorm-Emu68 is just a poor man's Transmeta-style CPU clone for 68K.


Quote:

Second, Commodore was already working on adding RTG support (for Amiga o.s. 4) and this could have been different from CyberGraphics or Picasso, and easier to implement. To be more clear, adding RTG support to the Amiga o.s. could have been much simpler and without requiring the big infrastructure introduced by CG or Picasso. Hence: it could have been in time for releasing the o.s. AND the packed/chunky chipset, and enable o.s.-friendly 3D games.

In modern times, FPGA Amiga chipset clones do not need RTG. Vampire's cloned AGA Doom performance is okay, but they didn't exist in 1994.

Vampire's SAGA supports Amiga's P96 RTG and hit-the-metal OCS/ECS/AGA just as PC SVGA chipsets support Windows GDI/DirectDraw drivers and hit-the-metal CGA/EGA/VGA.

The only reason why Amiga needs OS-friendly RTG is due to PC SVGA chipsets don't have embedded Amiga chipset legacy support while PC SVGA chipsets have PC's hit-the-metal VGA legacy support.

Modern PC GpGPUs still have VBIOS for PC legacy support and modern PC south bridge still has ISA bus via LPC bus.

------
On a side note, Intel's latest initiative is X86S i.e. yet another kill 16-bit X86 attempt after Intel's UEFI Class 3 initiative (remove CSM).

AMD's latest RX 7600/7900 XT/XTX and NVIDIA's latest GeForce RTX 4060/4070/4080/4090 still have VBIOS for PC legacy support i.e. not following Intel's initiative.

AMD's latest X670E chipset still has CSM for PC legacy support i.e. not following Intel's initiative.

The same Intel "kill the 16-bit X86" mentality since the AXP i432.

Quote:

And... this is related to what? Can you quote me and PROVE where this "nonsense" is coming from?

Chunky pixel alone doesn't save the platform i.e.
1. Atari Falcon has chunky pixel mode and it was configurated by ex-Commodore CEO i.e. gimped 16-bit bus. 68030 has a similar IPC to 68020.

2. Chunky pixel's performance advantage is dependent on the frame buffer's actual implementation.

ET4000AX has a 65 Mhz core clock speed, up to 32-bit bus**, and a 40 Mhz memory clock.
**32 bit DRAM or 16 bit VRAM variants. ET4000AX's bottleneck is the 16-bit ISA bus. 32-bit EISA bus and Opti Local Bus are optional.

Last edited by Hammer on 06-Jun-2023 at 06:50 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@ppcamiga1

Quote:

pistorm is just emulator on pc connected to Amiga.

Did you know 68060 translates CISC 68K into RISC-like instructions? Intel followed this design concept with Pentium Pro (1995) while AMD's NextGen has Nx586 (1994, RISC86).

Fund a 68K front-end R&D attachment for ARMv8 CPU or RISC V core IP.

Pistorm-Emu68 still has active Amiga 68K Auto-Config.

Pistorm-Emu68's RTG P96 Broadcom iGPU driver is still 68K.

Pistorm-Emu68's 68K data storage is stored as bare metal in memory below the 4GB logical memory address range.

RISC CPU design won the CPU design debate.

Quote:

which makes Amiga keyboard and mouse interface to PC.

Amiga chipset is still active with Pistorm-Emu68 i.e. it's the primary audio device. LOL.

For the internet with Pistorm32-Lite-Emu68, A1200's Gayle is used for PCMCIA 3Com Ethernet 10Mbit/s ethernet card. I tested this configuration and it works.

Quote:

get lost with this shit. real classic amiga use real cpu and not cpu from PC.

RISC CPU design won the CPU design debate. Deal with it.

Last edited by Hammer on 06-Jun-2023 at 07:21 AM.
Last edited by Hammer on 06-Jun-2023 at 07:18 AM.
Last edited by Hammer on 06-Jun-2023 at 07:13 AM.
Last edited by Hammer on 06-Jun-2023 at 07:06 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Hammer's PADDING. It's for context i.e. minority gaming PC hardware specs can still generate sufficient ROI for PC game developers due to the large PC market size. For PC games vs PC install base, the PC's attachment rate is terrible.

I reveal you another secret: games arrived on PCs even when their market size was 1/10 compared to 1993.
Quote:

Quote: Only if you want a very good frame rate. Cheaper cards can give a decent frame rate. Of course, if you buy the cheapest then the experience is bad. For Doom's 1993 Xmas, the PC has superior market access and price range coverage.

Clap clap clap.
Quote:

Replacing the slow VGA card on a typical 386DX/486 PC clone is not hard.

It was possible even on PCs with 8086, 80186, NEC V20/30, 80286.
Quote:

Quote: First, you clearly don't know how most of the Amiga games were developed at time. Hint: by killing the o.s.. So, it could have been easily possible to develop games using packed/chunky graphics if a (new) Amiga chipset had it in hardware, even if the o.s. had zero support. Hint: Doom DOS also "hit the metal".

Here you show to do not know at all how Amiga games killed the o.s. and how DOS games worked: the latter never killed the DOS and/or BIOS and took complete control of the system, even when using the so called DOS extenders. Which makes sense, if you know how DOS and BIOS were working; in fact there was no need to kill them, and all software, either it be an application or game, happily worked WITH them.

Have you had experience writing Amiga games (NDOS and/or os-friendly)? PC games? Or just studied how they worked? I don't think so: you have no idea. Hence, your statement.
Quote:

The difference is PC VGA standard has been cloned and improved VGA clones were released. SVGA clone's advanced features are accessible via Windows drivers.

And even via BIOS.
Quote:

The PC clone market was able to withstand the exit of IBM and Intel's multiple attempted killings of X86 from iAPX 432 to i860/i960 to IA-64. AMD's second source existence made sure Intel wasn't able to kill X86 for IA64 Itanium like Motorola killed 68K for PowerPC.

PADDING...
Quote:

The clone Amiga chipsets didn't exist in 1992 and 1994, hence when Commodore died, the Amiga platform died as a mainstream gaming platform. In modern times, FPGA Amiga chipset clones are surviving without Commodore / Escom.

PADDING.
Quote:

The clone 68K CPU market wasn't mature enough to survive Motorola's 68K exit from desktop markets.

That's a consequence of Motorola stopped 68K development AND keeping the IPs.

68K clones had already licenses, but since nothing new came from Motorola, they had no chance to evolve in architectural terms. Plus, they were used on the embedded market, which didn't required Ghz class processors.
Quote:

In modern times, FPGA 68K CPU clones are surviving without Motorola/Freescale.

Hobby projects for a niche market.
Quote:

PiStorm-Emu68 is just a poor man's Transmeta-style CPU clone for 68K.

No. Transmeta was quite different. IF you know the (low-level) details on how it was implemented and it worked.
Quote:

Quote: Second, Commodore was already working on adding RTG support (for Amiga o.s. 4) and this could have been different from CyberGraphics or Picasso, and easier to implement. To be more clear, adding RTG support to the Amiga o.s. could have been much simpler and without requiring the big infrastructure introduced by CG or Picasso. Hence: it could have been in time for releasing the o.s. AND the packed/chunky chipset, and enable o.s.-friendly 3D games. In modern times, FPGA Amiga chipset clones do not need RTG.

They need it if you want to take advantage of o.s.-friendly applications which exploited RTG features.

Applications which are just and only using the original chipset are severely crippled.
Quote:

Vampire's cloned AGA Doom performance is okay, but they didn't exist in 1994.

Exactly. And the same happens with PiStorm: it was NOT available at the time.
Quote:

Vampire's SAGA supports Amiga's P96 RTG and hit-the-metal OCS/ECS/AGA just as PC SVGA chipsets support Windows GDI/DirectDraw drivers and hit-the-metal CGA/EGA/VGA. The only reason why Amiga needs OS-friendly RTG is due to PC SVGA chipsets don't have embedded Amiga chipset legacy support while PC SVGA chipsets have PC's hit-the-metal VGA legacy support.

No and see above: RTG is needed to go well beyond the crippled Amiga chipset features and performances.
Quote:

Modern PC GpGPUs still have VBIOS for PC legacy support and modern PC south bridge still has ISA bus via LPC bus. ------ On a side note, Intel's latest initiative is X86S i.e. yet another kill 16-bit X86 attempt after Intel's UEFI Class 3 initiative (remove CSM). AMD's latest RX 7600/7900 XT/XTX and NVIDIA's latest GeForce RTX 4060/4070/4080/4090 still have VBIOS for PC legacy support i.e. not following Intel's initiative. AMD's latest X670E chipset still has CSM for PC legacy support i.e. not following Intel's initiative. The same Intel "kill the 16-bit X86" mentality since the AXP i432.

I've just written an article on that last week: Analisi e riflessioni sulla futura architettura Intel X86-S

It's in Italian, but it can easily translated with DeepL (the most accurate), Google Translate, ChatGPT, etc.

In short: you don't know of what you're talking about. Instead of blatantly read some news you should have read the technical article on Intel's site AND the whitepaper to fully understand (IF you're able to do it, of course) the proposal.

My bet: AMD will embrace it as well. Because:
- it's NOT a new architecture;
- it's just reflecting the ACTUAL status of x86/x64 o.ses & applications.
So, it MAKES SENSE.
Quote:

Quote: And... this is related to what? Can you quote me and PROVE where this "nonsense" is coming from? Chunky pixel alone doesn't save the platform i.e. 1. Atari Falcon has chunky pixel mode and it was configurated by ex-Commodore CEO i.e. gimped 16-bit bus. 68030 has a similar IPC to 68020. 2. Chunky pixel's performance advantage is dependent on the frame buffer's actual implementation. ET4000AX has a 65 Mhz core clock speed, up to 32-bit bus**, and a 40 Mhz memory clock. **32 bit DRAM or 16 bit VRAM variants. ET4000AX's bottleneck is the 16-bit ISA bus. 32-bit EISA bus and Opti Local Bus are optional

Do you understand that you're pretending to apply to me something which actually was and is continuously written by ppcamiga1?

Like kolla, you're not able to follow threads and have correct understanding of the discussions, contexts, and who said what/when/why.

Is it too much asking to have a STRONG understanding of what's going on BEFORE replying to people?

My point was that since the AA3000+ with dsp 3210 has been revived and now exists as real life functional and working systems (even as a DYI hobby project) there’s no need to speculate and fantasize about its capabilities. But maybe that’s less fun?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@ppcamiga1

Quote:

ppcamiga1 wrote: @Hammer arm in 2023 is also pc.

A total nonsense. As usual.
Quote:

pistorm is just emulator on pc connected to Amiga.

Which PC, hamster?
Quote:

which makes Amiga keyboard and mouse interface to pc. it is stupid because amiga keyboard and mouser where copied from pc 40 years ago. get lost with this shit.

Then tell me about your beloved PowerPCs...
Quote:

real classic amiga use real cpu and not cpu from pc.

Define "real" on all such components.


@Hammer

Quote:

Hammer wrote: @ppcamiga1 Quote: pistorm is just emulator on pc connected to Amiga. Did you know 68060 translates CISC 68K into RISC-like instructions? Intel followed this design concept with Pentium Pro (1995) while AMD's NextGen has Nx586 (1994, RISC86).

Another misleading information: there's no RISC on all of them.

If you're interested to really understand how was / is the situation you can search and read the interview to one of the key Pentium Pro developer which explains why and strongly rebut your statements. Concepts and statements which still apply to recent CISC processors.
Quote:

RISC CPU design won the CPU design debate. [...] RISC CPU design won the CPU design debate. Deal with it.

Absolutely no. Again, you don't know of what you talk about!

BTW a discussion like that happened some years ago here and I've proved my statement. You can search it.

I've also written another article (this time in English) which is almost ready and that I plan to complete and publish it on the following weeks. Stay tuned.

@kolla

Quote:

kolla wrote: My point was that since the AA3000+ with dsp 3210 has been revived and now exists as real life functional and working systems (even as a DYI hobby project) there’s no need to speculate and fantasize about its capabilities. But maybe that’s less fun?

Then next time WRITE IT EXPLICTLY instead of hiding your thought to the participants: we CANNOT read your mind!

@cdimauro

Quote:

Talking about ignorance, x64 is used on x86 side since around 20 years.

And it’s around 20 years ago that I was confused about it.

Quote:

Whereas its usage for 21x64, well, let's say that it's not so common on IT literature.

When a series of CPUs 80x86 is referred to as x86, it makes total sense that another series 21x64 is referred to x64 - it was the only CPUs around with x64 in its actual name, right? In the world I operated in, 64bit x86 was referred to as amd64 or sometimes x86-64. Only certain parties pitched the x64 term for it, all of which were quite irrelevant for me in my profession, so yes, it had me confused. For a while. I still find it awkward and never use it myself.

Quote:

Where have you've lived in the last 20 years? On a cave?

It’s “IN a cave”, I see you do this mistake all the time - in as inside, on as on top of.

And where?
HPN, HPC, all kinds of servers and services, academia, research and enterprise.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro Quote: Talking about ignorance, x64 is used on x86 side since around 20 years. And it’s around 20 years ago that I was confused about it.

You stated something different:

"It took me quite a while to realize that when people write x64 they typically no longer mean DEC Alpha 21x64…"

20 years x64 was introduced as a shorter and more general term compared to x86-64. So, when your "a while" materialized?
Quote:

Quote: Whereas its usage for 21x64, well, let's say that it's not so common on IT literature. When a series of CPUs 80x86 is referred to as x86, it makes total sense that another series 21x64 is referred to x64 - it was the only CPUs around with x64 in its actual name, right? In the world I operated in, 64bit x86 was referred to as amd64 or sometimes x86-64. Only certain parties pitched the x64 term for it, all of which were quite irrelevant for me in my profession, so yes, it had me confused. For a while. I still find it awkward and never use it myself.

What you find is irrelevant.

What's relevant is that x64 is very wide spread as a synonym of AMD64, x86-64, EM64T (the first acronym used by Intel for its implementation of the same 64-bit ISA) and Intel64 (last and final Intel's name for it).

And no, it's not about "certain parties": it's largely used. https://duckduckgo.com/?q=%22x64%22
Quote:

Quote: Where have you've lived in the last 20 years? On a cave? It’s “IN a cave”, I see you do this mistake all the time - in as inside, on as on top of.

Thanks for the correction. I'm not a native speaker. I'll try to remember it.
Quote:

And where? HPN, HPC, all kinds of servers and services, academia, research and enterprise.

As a sysadmin, I assume. But you're not so well informed about IT literature (see above).

@cdimauro

Quote:

cdimauro wrote: @kolla Quote: kolla wrote: My point was that since the AA3000+ with dsp 3210 has been revived and now exists as real life functional and working systems (even as a DYI hobby project) there’s no need to speculate and fantasize about its capabilities. But maybe that’s less fun? Then next time WRITE IT EXPLICTLY instead of hiding your thought to the participants: we CANNOT read your mind!

I thought it was obvious, and took it for granted that people who are so passionately involved with everything Amiga that they spend so much time writing long series of articles and potsings on Amiga forums would know already.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@cdimauro

When it comes to literature I rather spend my time with other (non IT) topics, and I certainly have little interest in reading about intel architectures of late. My IT books are mostly on topics of certain programming languages, (obscure) operating systems, manuals of old (DEC, yes, lots - orange and gray), networking and protocols, and of course Amiga and similar.

I see Wikipedia list up Microsoft, Sun and Oracle as the parties who have been using «x64», so yes… mostly irrelevant for me, closest I got was probably when using OpenIndiana for an experimental zfs/nfs4.2 based NAS cluster.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro Quote: cdimauro wrote: @kolla Then next time WRITE IT EXPLICTLY instead of hiding your thought to the participants: we CANNOT read your mind! I thought it was obvious, and took it for granted that people who are so passionately involved with everything Amiga that they spend so much time writing long series of articles and potsings on Amiga forums would know already.

Wrong assumption.

First, I've never shown interest on everything about Amiga neither I'm really interested on all of that.

Second, and more important, the discussion's context was clearly and only talking about Commodore's prototypes. Not even a single statement was reported anything different.

Words are important kolla. You may know and surely know much more about Amiga and English, but knowledge by itself isn't useful on real life.

As an IT guy should know and understand the importance of the so called "Applied Knowledge": being able to use the acquired / accumulated knowledge / expertise in a proper way (which requires understanding the context of a problem, of course).

@kolla

Quote:

kolla wrote: @cdimauro When it comes to literature I rather spend my time with other (non IT) topics, and I certainly have little interest in reading about intel architectures of late. My IT books are mostly on topics of certain programming languages, (obscure) operating systems, manuals of old (DEC, yes, lots - orange and gray), networking and protocols, and of course Amiga and similar.

That's fine. As I've said, I'm not omniscient. And nobody could be and is requested to be.
Quote:

I see Wikipedia list up Microsoft, Sun and Oracle as the parties who have been using «x64», so yes… mostly irrelevant for me, closest I got was probably when using OpenIndiana for an experimental zfs/nfs4.2 based NAS cluster.

That's the reason why I've provided a link to just searching "x64": to show that even Wikipedia is reporting a limited information.

In fact, x64 is used well beyond Microsoft, Sun, Oracle.

An example, directly coming from Intel: Introduction to x64 Assembly