Quote:

Dandy wrote: DPaint runs on Amigas without using the OS??? How does it do that?

It obviously doesn't. Running something under the OS doesn't mean it can't access the hardware directly. Accessing hardware isn't the same as running without an OS.

@Dandy

Next he's gonna tell me my '69 MF135 is not to be used with modern farm implements either, in which case I'm screwed

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@KimmoK

Ooooooh, you are SO going to hell.
Quote:

But with extra HW one can attach xena to PCIe, serial, USB etc. ports and slots. For example, you could have xena driven DAQ capture card that uses PCIe lanes for data transfer and xena for control. Hardly a killer feature, though.

NO NO NO NO NO!

Can't you read what the expert wrote.

Quote:

a so called "local bus" connected to the CPU....... This is the only way to interface the XCore with the rest of the computer.

Now, do not use your head for anything more than a hat holder, or suffer the consequences.

EXPERT!

_________________
Tinkering with computers.

@Signal

ops!! that wasn't nice. tell him your sorry :)

_________________
nox's in the house!

Cheap hardware like the Raspberry Pi... something to get more people to use AOS 4, to get it to ex-Amigans and new users who can't afford/aren't willing to pay for current hardware.

As a hobby AOS 4.1 already offers a good time.

With more developers and users motivation will most likely be better at Hyperion and perhaps more can contribute to the development of AOS 4.

I know that there are interested ex-Amigans out there that doesn't use MOS either or AROS for that matter but they are buying things that they enjoy using if the price is right and the product is interesting enough.

Retro is selling and RunInUAE are good enough for those who wants to run some old games while using it for other things (nostalgia). There must be a reason to buy it... it's hard to convince someone to pay for current hardware to run something that is worse (in many ways) than the competition.

Something that looks like a new A1200, a laptop or similar for a good price (I think that the case/keyboard matters).

The OS alone isn't enough.

@KimmoK

Quote:

KimmoK wrote: @persia >what advantage (if any) hardwiring to the X1000 has. -Low latency between the host computer and xcore.

I don't think so, since they have to engage some protocol for communicating. The XCore doesn't have a proper bus, but only I/O pins which should be programmed in some way.
Quote:

-Ability to build dumb prototype cards for PCIe, when xcore handles part of the debugging & control.

According to source which Michael posted, it's not a real PCIe slot.
Quote:

-xena might be able to communicate (+watchdog) the CPU even when USB is already dead, unlike xcore on your Mac.

You can do the same by putting the XCore into a real PCIe card, connected to the computer of course.
Quote:

Nothing earth shaking.

Exactly. I don't see any added value, compared to the several solutions which can be already bought, and at little price.
Quote:

(normally, if xcore is put on motherboard, it most likely is used as programmable bridge chip, as an alternative to FPGA chip. It was said that xcore was put on x1000 motherboard for users to explore, but the idea is still handicapped by the lack of native Amiga tools. Similar to LatticeXP on SAM, IMO. Old thread about ideas: http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=27379&forum=33#668515 )

An FPGA with a sufficient amount of LE would have been much more interesting and useful.
Quote:

@thread It would be cool if AREXX & automation capability would be better used on Amigan offerings. Made as simply usable as note writing on a PC.

Python is much more versatile for this kind of staff.
Quote:

IMO: Amiga speech capability is good example of revolutionary standard OS feature of a 1985 computer. A feature that in 2015 is STILL non-existent / not used at all in computing.

"Hey, Cortana!"

@Dandy

Quote:

Dandy wrote: @cdimauro Quote: cdimauro wrote: @Dandy I cannot convince you if you don't give some clear statement regarding what you want to do with an AmigaOne X1000 and its XCore. [phantasy mode on] First I'd like to explore the possibilities in how far the A1X1k/xCORE can be used as 'brain' of an manufacturing cell, including using CAD/CAE for the design process and CAM for controlling CNC machines and robots. Secondly, I'd like to design, manufacture and build a model steam locomotive (H0 scale) and use the A1X1k/xCORE to control it on the model railway. This would include the transmission of video signals from pen cams on the loco to the monitor, as well as of all the steering commands from the computer to the loco on the model railway. If the A1X1k/xCORE can live up to the advertised capabilities, it should basically be possible to do all this with an A1X1k... [/phantasy mode off]

You'll find a lot of fun transmitting the video signal from the loco to the monitor, since the XCore can transfer only 15MB/s to the CPU/memory.

Anyway, do you think that your project cannot be realized using some other solutions (like the one which I reported, for example)?
Quote:

Quote: cdimauro wrote: The XCore cannot directly access the memory, and for doing this there's a so called "local bus" connected to the CPU, that can be used to access the memory, but it's very slow (15MB/s, if I remember correctly), and I think it'll also long latencies. You're talking about the memory of the A1X1k - and not about the xCORE's own RAM, aren't you?

Exactly.
Quote:

Quote: cdimauro wrote: This is the only way to interface the XCore with the rest of the computer. I know.

It seems that not everyone is able to understand it, even if it's clearly reported on the machine's web page...
Quote:

Quote: cdimauro wrote: And of course it has some GPIO pins that are also connected to the so called Xorro bus (a modified PCI-Express AFAIK). typically 8 GPIO Pins = 1 GPIO channel (port)

OK, thanks for the info.
Quote:

Quote: cdimauro wrote: With such configuration, it cannot do something interesting, apart some automation. That might depend on what you think is interesting...

Something which can be realized with a better synergy of the XCore with the main system. The problem here is that the communication channel between the two is too much limited (see above: 15MB/s of bandwidth).

So, it's much more convenient taking some external GPIO board, and connecting it to a PC with an USB2 connector, for example.
Quote:

Quote: cdimauro wrote: I'm sorry for that. I hope that you solved your problems. Not yet. 'Solving' the problem might take another few years (including some more surgeries at the Charité in Berlin) - if possible at all... (but there seems to be a 70% chance that it can be solved)

I wish you the best for solving it.
Quote:

Quote: cdimauro wrote: IMO they gave the illusion to the amigans that the machine is like an Amiga with custom hardware, whereas the reality is much different. Marketing to sell the machines... Hmmm - I thought from the beginning that this might be an interesting feature. Back in 1989, when I got my CAD analyst education and did my CAD "homework" on my A 500 at home (including the generation of working CNC programs), I was dreaming of a machine with such features...

But nowadays you can have more spending much less.

Consider, also, that SDKs for such boards usually support only Windows and/or Linux. You can have hard time trying to use something different. I'm not saying that it's not possible, but only that the experience can be worse.

@KimmoK

Quote:

KimmoK wrote: @thorham >AOS is a dead end in terms of moving forward What is not? Every SW needs to be rewritten after a few years if they are actively maintained + new features developed, otherwise maintenance starts to take too much effort etc. (IIRC, that was once taught on some ISEB certification course)

That's not true: you don't rewrite it every time.
Quote:

To implement some modern stuff to AOS will require re-implementing a lot of it.

That's a different thing.
Quote:

@cdimaurio: >The XCore cannot directly access the memory, That is true for regarding the main memory. I think in theory it would have been possible to develop a board that uses conexium to DMA to motherboard main memory etc. But it would have been futile for the low bandwidth xcore in use.

Exactly. It makes sense if you can offload some concrete work from the CPU, but this isn't the case.

Do you know how Amiga's coprocessors worked? That means concrete offloading.
Quote:

> and I think it'll also long latencies. I hope that is not true, as it would remove more than half of the usability of the integrated xcore. IIRC, latencies are not specified / benchmarked yet anywhere,

Do some simple math. If you can transfer only 15MB/s, it means that it takes around 67ns to transfer just one byte.
Quote:

but IIRC, xcore can trigger an interrupt to get PA6T attention.

It doesn't change the picture.
Quote:

>This is the only way to interface the XCore with the rest of the computer. True for a xena implementation on nemo board.

Well, some genius hasn't the same opinion...
Quote:

But with extra HW one can attach xena to PCIe, serial, USB etc. ports and slots. For example, you could have xena driven DAQ capture card that uses PCIe lanes for data transfer and xena for control. Hardly a killer feature, though.

Especially if you need to introduce some extra hardware.

Why should anyone has to complicate his life when there are plenty of much cheaper and easier to use solutions?

P.S. Please, don't edit your existing comments adding new stuff: just write new comments. Thanks.

@Dandy

Quote:

Dandy wrote: @cdimauro Quote: cdimauro wrote: @Dandy I don't speak German (nur bisschen, tut mir leid), so I translated it with Google. Oooops - as below your avatar the info "From: Germany" is given I thought you are German. Sorry...

No problem. I moved from Italy two years ago, due my new job, and my kids make jokes to me, for my not even mediocre German knowledge...
Quote:

Quote: cdimauro wrote: Taking the sentence that you marked in bold, the second part (which I've underlined) contradicts the first. Hmmm - to my understanding it does not really contradict. The second part just makes restrictions. I understand it that way, that AOS behaves like an RTOS, as long as all runs 'normal'. But due to design limitations of the OS there is no guarantee that it can respond in realtime, as soon as 'unnormal' events occur. Perhaps it will respond in realtime, perhaps not. That's why it is frequently referred to as "close to RTOS".

But a RTOS was/is designed just to address those cases, which a normal o.s., like the Amiga one, cannot handle.
Quote:

Quote: cdimauro wrote: Do you know of any 64-bit Amiga or post-Amiga (like I call the non-Commodore machines which run the Amiga o.s., or some "cousin")? I don't, except AROS with is available in the x64 "flavor". Yes, I do. The A1X1k has PA6T-1682M cpu, which is 64-bit: Quote: Wikipedia on PWRficient: "PWRficient is the name of a series of microprocessors designed by P.A. Semi where the PA6T-1682M was the only one that became an actual product. PWRficient processors comply with the 64-bit Power Architecture, ..." Just that AmigaOS cannot take advantage of it yet... Furthermore the PA6T-1682M is dual core, where the AmigaOS also cannot capitalise on yet... Quote: cdimauro wrote: I think that he wrote 32-bit only to emphasize that such machines were (strictly) bound to 32-bits. Hmmm - the post-Amiga (as you call it) "Sam" from ACube Systems is also 32 bit. So we have to be careful - the number of bits of the cpu is also not very well suited to distinguish between "Amigas" and "post-Amigas" (or "NG-Amigas).

Yes. BTW, currently also MorphOS runs on 64-bit processors, but like AOS it runs only in 32-bit mode.

AROS is the only one which really runs on and makes use of a 64-bit processor.
Quote:

Quote: cdimauro wrote: You are a rare bird nowadays, since it's pretty common to have such stuff already integrated and immediately working. Some applications aren't even used until you don't run them, so they don't impact on the "fast" o.s.. They are there: click and run them IF you want to do; or forget them. Yes, for mainstream OSses. But I think we both agree that AmigaOS is far, far away (old 'Slade' title, btw.) from being a mainstream OS and most likely will never become mainstream again. So why bother with integrating these things in the OS and waste precious resources on it, if you can have them via 3rd party apps? Better implement things like 64 bit-ness and SMP correctly... I'd prefer to have those essential 3rd party apps included on the OS-CD/DVD and to have an installer that allows me to chose if I want to have the TCP/IP stack, the browser, the mailer and/or the USB stack installed together with the OS.

You can take them from third-parties, but a user wants them. Icaros (an AROS "distro"), for example, integrates a lot of stuff on top of AROS because users want and need them all (even more).

However implementing things like 64-bit and SMP for AOS means losing its backward compatibility with existing applications (both PowerPC and 68K). That's why you'll not see a 64-bit version of it, and only a multicore (not SMP) support will be added.
Quote:

Quote: cdimauro wrote: Lean, well, with hard disk which are already measured in TB from years, doesn't make sense. And a modern o.s. (not only Windows) uses around 10-20GB, which is a ridiculous amount compared to the available sizes. So I don't care too. Hmmm - yes, HDs are available in TB sizes and cheap. But still every bit takes time to be loaded from/written to an HD...

Try an SSD: it completely changes your life.

However in this case common sizes are measured in terms of hundred of GB, and not TB (too much expensive).

@Signal

Quote:

Signal wrote: @KimmoK Ooooooh, you are SO going to hell. Quote: But with extra HW one can attach xena to PCIe, serial, USB etc. ports and slots. For example, you could have xena driven DAQ capture card that uses PCIe lanes for data transfer and xena for control. Hardly a killer feature, though. NO NO NO NO NO! Can't you read what the expert wrote. Quote: a so called "local bus" connected to the CPU....... This is the only way to interface the XCore with the rest of the computer. Now, do not use your head for anything more than a hat holder, or suffer the consequences. EXPERT!

Cherry-picking only part of my comments and ignoring even what the official web site reports (and which perfectly matches with what I've written) shows that your only purpose here is to troll.

And that's the only thing that you are able to do, of course.

@cdimauro

Quote:

cdimauro wrote: @Signal [quote] Cherry-picking only part of my comments and ignoring even what the official web site reports (and which perfectly matches with what I've written) shows that your only purpose here is to troll. And that's the only thing that you are able to do, of course.

From your post #1102
Quote:

From here: "Xena has 64 of these configurable I/O lines. In the AmigaOne X1000 we have one quarter of these connected to the CPU local bus for direct communication with the system, whilst the other three-quarters are connected directly to pins on the Xorro slot for communication with the outside world."

And from your post #1095
Quote:

This is the only way to interface the XCore with the rest of the computer.

You are the one that added the word ONLY to what A-eon stated, and now YOU treat it as FACT.

Of course that makes me a troll.

OR, how about your 15MB/s transfer rate?

I have no idea where you got 15MB/s, is that another one of those "Some guy said" facts?

Did you take into consideration that each transfer is 16 bits?

Oh, nevermind what is, just keep making up your own facts. You're good at it.

And thanks for the Troll designation, I can add it the FACT that I'm on Hondo's list.

Also, thanks for all the laughs, and if I can pull some facts out of my ass I'll be sure to report them to you.

_________________
Tinkering with computers.

@Signal

Quote:

Signal wrote: @cdimauro Quote: cdimauro wrote: @Signal [quote] Cherry-picking only part of my comments and ignoring even what the official web site reports (and which perfectly matches with what I've written) shows that your only purpose here is to troll. And that's the only thing that you are able to do, of course. From your post #1102 Quote: From here: "Xena has 64 of these configurable I/O lines. In the AmigaOne X1000 we have one quarter of these connected to the CPU local bus for direct communication with the system, whilst the other three-quarters are connected directly to pins on the Xorro slot for communication with the outside world." And from your post #1095 Quote: This is the only way to interface the XCore with the rest of the computer. You are the one that added the word ONLY to what A-eon stated, and now YOU treat it as FACT.

Again, cherry-picking. I report the more complete text:

"This is the only way to interface the XCore with the rest of the computer.

And of course it has some GPIO pins that are also connected to the so called Xorro bus (a modified PCI-Express AFAIK)."

This perfectly matches with the information reported on the machine (motherboard) site.

You can continue to climb on the mirrors, but those are FACTS.
Quote:

Of course that makes me a troll.

You provided no contribute to the discussion, apart childishly trying to make jokes against me.
Quote:

OR, how about your 15MB/s transfer rate? I have no idea where you got 15MB/s, is that another one of those "Some guy said" facts?

Here vidarh reported some information from XMOS (so, not me: it's the company which realized the chip) about how to implement an SDRAM memory controller for the XCore (since it has no memory controller at all), which gives 30MB/s of bandwidth.
Quote:

Did you take into consideration that each transfer is 16 bits?

Of course. In fact 30 / 2 = 15, since we are talking about 16 bits instead of the 32 normally transferred.
Quote:

Oh, nevermind what is, just keep making up your own facts. You're good at it.

Thanks.
Quote:

And thanks for the Troll designation, I can add it the FACT that I'm on Hondo's list.

You had only what you deserve.
Quote:

Also, thanks for all the laughs, and if I can pull some facts out of my ass I'll be sure to report them to you.

Well, since now you never reported a single fact against what I've written. So I can sleep well (as usual).

Good night.

@Thorham

Quote:

Thorham wrote: Who ever said you shouldn't use an OS on your Amiga (it's actually impossible to not use an OS on an Amiga, except for the A1000)?

I understood your posting in that way, that you don't like the AmigaOS, but love the old hardware. That's why I stated that hardware without an OS is pretty useless for me.

Quote:

Thorham wrote: Just because AOS is a dead end in terms of moving forward, doesn't mean it's useless.

O.K. - then I misunderstood you.

_________________
Ciao

Dandy
__________________________________________
If someone enjoys marching to military music, then I already despise him.
He got his brain accidently - the bone marrow in his back would have been sufficient for him!
(Albert Einstein)

To Dandy:

Let me put it this way: AOS is fine for what it is, and it allows people to use their Amigas properly (including myself). Nothing wrong with that. The problem is that AOS is a dead end in terms of moving it forward because of bad design choices from the past that AOS is now stuck with. The moment you get rid of this crud is the moment you throw backward compatibility out the window. It's a problem that's exceedingly hard to solve if it's solvable at all.

So basically it's not so much about not liking AOS, but more about recognizing that it's very, very hard to truly move it forward.

Last edited by Thorham on 07-Aug-2015 at 07:21 AM.
Last edited by Thorham on 06-Aug-2015 at 08:10 PM.

@Thorham

Quote:

bad design choices from the past

I guess that goes for f.ex. lack of memory protection.

However, that would have required MMU even on "entry level" systems (and increased memory usage a bit), so would it have been a good choice to make all Amiga systems more expensive in the past? Not necessarily.

Sometimes the choice might be bad, but in reality, there's no other alternative.

Needless to say, MacOS had (in some ways) even worse legacy, and was eventually rewritten (with limited backwards compatibility) for the current major version number release that was released 14 years ago, and much of the compatibility has been dropped since then.

As has been said by others, that would be the only "real" chance for AmigaOS (and any significant market penetration would still be very doubtful). And would that still be "Amiga"? And would there be any point in being "Amiga"?

Is OSX as "Mac" as MacOS 9? I'd say it's at least completely different. The company is definitely more "same" than would be in the case of AmigaOS.

I think Amiga needs to get more mobile and less expensive. Lots of people will be quick to shut anyone down that has a different idea. The Amiga would have never been if anyone had listened to that kind talk. The Amiga was successful in its early days because of vision and a willingness to continue despite the nay sayers. Desktop is OK for me, but not OK for a modern system. Nowadays people want to take their technology with them. There have been success in other os's fairly recently. Android is a good example. I do not like Android as there is no common environment and some apps you just can't seem to shut down, and the buttons have pictures that treat you like a juvenile, but you are supposed to know what picture stands for what when there is no standard. No I do not think Amiga OS needs to be a cell phone, but I think to be successful it needs to be at least mobile.

@Jupp3

Quote:

Needless to say, MacOS had (in some ways) even worse legacy, and was eventually rewritten (with limited backwards compatibility) for the current major version number release that was released 14 years ago, and much of the compatibility has been dropped since then.

thoram is rather right. for a thousand and tenth time: mac os had a user base it could impose whatever they liked and stay in business. does morphos has this atut? does aros or os4 have it? no. so what?

Last edited by wawa on 06-Aug-2015 at 09:27 PM.

@Jupp3

Quote:

Jupp3 wrote: @Thorham Quote: bad design choices from the past I guess that goes for f.ex. lack of memory protection. However, that would have required MMU even on "entry level" systems (and increased memory usage a bit), so would it have been a good choice to make all Amiga systems more expensive in the past? Not necessarily. Sometimes the choice might be bad, but in reality, there's no other alternative.

There were alternatives:

- do not publish the internal structures used by the o.s. (e.g.: use handles, with getters and setters when using "opaque" types);
- set by default all task's memory as private;
- introduce APIs to control how tasks share memory between each other.

You don't need an MMU for that, and it doesn't prevent applications to crash or bring down the whole system, but on a system with an MMU it can be transparently used by the o.s. and applications will benefit immediately without hurting compatibility.
Quote:

Needless to say, MacOS had (in some ways) even worse legacy, and was eventually rewritten (with limited backwards compatibility) for the current major version number release that was released 14 years ago, and much of the compatibility has been dropped since then. As has been said by others, that would be the only "real" chance for AmigaOS (and any significant market penetration would still be very doubtful). And would that still be "Amiga"? And would there be any point in being "Amiga"? Is OSX as "Mac" as MacOS 9? I'd say it's at least completely different. The company is definitely more "same" than would be in the case of AmigaOS.

I absolutely agree. But unfortunately the current post-Amiga scene is bounded to a spasmodic need for (full) retro-compatibility with the old software, and that prevents the introduction of new, modern features on post-Amiga o.ses.

For things to change, the old apps should run on ad-hoc sandboxes, with some resource sharing with the host o.s.. Amen.

@KingKong

I agree it should go Open Source we would be way ahead of where we are now which isn't much farther then the 80's.

@Gmeanie

AROS is already open and available even for 68k but i understood people around here don't want it because is not teh real or something like that

_________________
Scarabocchi Binari - Italian AROS Blog
Binary Doodles - English language AROS Blog