@ppc4me

Quote:

ppc4me wrote: @Zylesea >d) e700 - the 64 bit cory by freescale. Currently put on hold Do you have a source? Imho the e700 was more a candidate to go into the next (2007+) generation of PowerQuiccs in 65nm.

A Freescale representive told me so in summer. Maybe they decided to took up development again. But in summer they said it's put on hold.

Regarding the PA-Semi I am pleased to stand corrected by your remarks. I havn't checked all details regarding that chip yet.

_________________
My programs: via.bckrs.de
MorphOS user since V0.4 (2001)

@olegil

Quote:

olegil wrote: I note how Seehund seems to have a weird definition of failure, though. If something sells out, how can it not have been a success? I'm certain they would sell more batches still (would be nice to find some working riser cards and figure out the old PCI graphics card issue, though).

Oh dear.

/me flicks the "it's selling like hotcakes!" switch to OFF on the Boumese Reality Distortion Field control panel.

Seriously. If the stated goal with the "AmigaOnes" had only been to "let Eyetech very slowly shift a teeny weeny trickle of batches to a minority of the sad remnants of the Amiga community", then this venture and this product could possibly have been described as "a success". A totally pointless "success", but an accomplished goal nevertheless. While people predicted this result (cough), it wasn't Eyetech's/AInc's stated goal.

The "AmigaOne" was supposed to be the hardware product where AmigaOS would be reborn. It was supposed to be sold with so much better warranties, testing, compatibility guarantees, guaranteed availability, user support et c. than other hardware, so much better in fact that AmigaOS users shouldn't even be allowed to shop outside the "Amiga hardware market". It was supposed to be what would attract new users once AmigaOS4 had been released. Some even thought it was pertinent to refer to it as "the new Amiga"...

It FAILED. Miserably.

In four years, Eyetech sold less than 1,500(!) units, and to my knowledge that is the vast majority of all Terons ever sold anywhere regardless of trademark! During these four years, THREE Teron/"AmigaOne" models were introduced, of which one was made because the predecessor was totally unsuitable for public consumption, the other because its predecessor proved to be unsellable, but that didn't help. There never was any demand large enough to even warrant a production batch larger than the average test run for other consumer hardware! Once Terrasoft had tested the Terons, they scrapped their far-gone plans to sell them, and Eyetech remained the only vendor outside the Atum/Mai sphere! Virtually nobody outside the "Amiga hardware" pseudo-market wanted a "non-AmigaOneised" Teron in the end. And how many genuinely interested customers are waiting for an "AmigaOne" today? 10? 15? 30? Anyway it wasn't enough to warrant production of another minimal batch. And now the goddamn northbridge has even been EOL'd and nobody really knows if Mai Logic still exists, so even if anyone would be interested in producing more Terons/"AmigaOnes" at some time, they couldn't, unless there's a box of ArticiaS's rotting in a warehouse somewhere.
All this before AmigaOS4 is even finished, and even that is partly because of the time wasted on finding software workarounds for, er, misbehaving hardware.

What's "weird" about calling this a failure?
Is this something we'd like to see happening again?

_________________
Oh, bother.

@T_Power

Quote:

don't believe anyone called Dave a "caveman", but some of US were called "cavemen" BY Dave. No hassle though, most here have real tough skins. (The "Report" button is for the weak. )

Sorry, I misquoted. I've read through the thread in entirey, not going to do so yet again, but there were enough derogatory statements aimed at Dave or assuming he had no knowledge etc, that I'd posted a warning earler on in the thread once he started participating. The pint being that debates and arguments can be made without being insulting. Yes, Dave made a few semi-insulting comments as well, but I was more susprised/dismayed by a few user comments myself.

Quote:

Any inside news from Hyperion as to when update #4 will be released?

Heh, I wish!

_________________
Are we not done with the same silly arguments and flames yet??!

@Hammer

Quote:

PearPC is not the fastest FOSS PPC emulator i.e. refer to SheepShaver.

Yes, and I believe you and I had this discussion already somewhere on AW, where I posted the relevant figures instead of a one line response..ie refer to....complete arguments vs one-line responses.

_________________
Are we not done with the same silly arguments and flames yet??!

@Samwel

Quote:

Just to make this clear. Eyetech did NOT make a loss in selling the A1's. The price was calculated to give exact 5% profit & 5% for future development which means 10% profit per board as Eyetech themselves didn't make any boards.

I didn't know that. I don't suppose you also have handy if there was also a hardware fee going to AI as well? I thought I'd seen something to that effect but can't find it/may be wrong..?

Even so, 10% is a _very_ slim margin. One thing goes wrong and you're in the red.


edit - err, total screwup, write black instead of red, wouldn't that be nice?

Last edited by wegster on 17-Jan-2006 at 07:51 PM.

_________________
Are we not done with the same silly arguments and flames yet??!

@COBRA

Quote:

Nobody suggested that the CPU of XBox is a potential CPU for OS4 hardware, or that this particular CPU will make the other PPC chips cheaper/etc. We're talking about where the PPC architecture in general is going and looking at what markets it is currently entering with different chips. The XBox360 was worth mentioning I think, because Microsoft, the company pretty much responsible for the success of the x86, is using PPC chips in their new consoles, in fact all of the three main console companies are going PPC. It basically means PPC is going to dominate the home entertainment market. I don't really understand why discussing this here upsets you.

Actually, I did, in a fairly long post in this thread, as well as PS3. Licensing issues aside, meaning I don't know what they might be in producing a 'title' for either the XBox360 or PS3, it was suggested as a means to bring OS4 to 'the masses,' on effectively inexpesnvie/commodity hardware while remaining PPC in nature, and perhaps generating some $ if the details were able to be worked out.

(Yes, I'm aware they are different CPUs than a normal desktop, ie Mac G5). See prior thread and Dave's response somewhere between pages 17-21 or so iirc..

And in general though, you're right, any discussion of PPC future should certainly include the fact that at least 2 of the next gen consoles are PPC derived..

_________________
Are we not done with the same silly arguments and flames yet??!

@wegster
Quote:

Quote: Any inside news from Hyperion as to when update #4 will be released? Heh, I wish!

I wish too, sound like many new features in update #4.
One I need is screen dragging, because moving back and forth from the A4000 to the A1 I often frustrate myself trying to drag screen on the A1.

Quote:

Even so, 10% is a _very_ slim margin. One thing goes wrong and you're in the black

With all the things that have gone wrong, all in Amiga-Land should be rich.
Hehe! "...in the black" means in positive.

Cheers,
Tim

@T_Power

Quote:

With all the things that have gone wrong, all in Amiga-Land should be rich. Hehe! "...in the black" means in positive.

rofl, right you are, not a CLUE what I was thinking when I wrote that, obviously I meant in the red, changing

_________________
Are we not done with the same silly arguments and flames yet??!

@Stephen_Robinson

Quote:

He talks a lot of sense this man.

I agree with a lot of things Dave has said. Whether I want to believe them or not is a whole different story.

I'm one of these people that sits on the fence and mainly just watches, hence my join date and post count.

Steve

_________________
A1200T - OS4.0,OS3.9: 603e PPC 200mhz,060 50mhz, 256mb ram, FastATA MK-III, BVision, 160gb,20gb HDDs

A1200 - OS3.1: Blizzard IV 030, 64mb ram, 400mb HDD

OS4.x - Flying the AMIGA flag

@wegster

Quote:

wegster wrote: Actually, I did, in a fairly long post in this thread, as well as PS3. Licensing issues aside, meaning I don't know what they might be in producing a 'title' for either the XBox360 or PS3, it was suggested as a means to bring OS4 to 'the masses

I agree, but I'm afraid chances M$ will permit an "official" port of OS4 to their console are less than zero, even if we consider how poor is our OS, feature-wise, compared to Windows.
On the other hand, PS3 could be a very different affair... well at least that's what I hope.

Quote:

And in general though, you're right, any discussion of PPC future should certainly include the fact that at least 2 of the next gen consoles are PPC derived..

Again, I agree. I also understand that Revolution's going to use a PPC-derived CPU too

@CodeSmith

Quote:

About the Z80 line though, Zilog still seems to be making them. I know because I'm slowly teaching myself digital electronics, and my "course project" is a simple 80s style SBC, so I've been doing research (so many cool chips have been discontinued

I was looking for a "real" Z-80, something pin compatible with the original. I couldn't find it. Yes, the Z-80 ISA is still available, but if you had a Z-80 board, you have probably had to change it over that time, to adjust to the changes in the available parts. Yes, certainly less than having to re-code everything, though you probably wind up, over the years, moving to a more "microcontroller" than "microprocessor" design, and changing to use all the extra stuff, too.

For example, today, it would be kind of silly to make a Z-80 with the old 8/16 bit external bus, when you could easily put RAM and Flash on-board for essentially the same price (eg, the cost of the smallest die available and the package).

Quote:

I must say that I'm slowly becoming more and more disturbed by this thread... are we *really* avoiding x86 *just* for 'religious' reasons?

Yes. That's my take on it. At one point in time, x86 was a demonstrably convoluted architecture, and well behind the curve of other CPUs. Of course, there were a bunch of them back then. Sometimes later, it was pretty much just x86 and 68K. Then everyone using 68K for anything beyond "personal computer" switched to RISC.

As a result, the 80's mindset would absolutely point to "personal computer" + "RISC" as A Good Thing. When, in truth, RISC was really just "the toolkit for 90s CPUs", and every chip, eventually, moved toward RISC: x86 (from NexGen to K7 to P4), 68K (in the modern embedded ColdFire incarnation), general embedded chips (from ARM, the world's most popular ISA at the moment, MIPS, PowerPC-embedded, MSP430, etc).

I suspect there's also a small degree of cognative dissonance. The Amiga industry, as a whole, was opposing The PC for so long, it had to be difficult to in a way admit defeat, even though that defeat really had nothing to do with The Amiga. After all, the 68K was dropped by Motorola, in lieu of IBM's processor (which isn't particularly any more like or dislike the 68K, other than the fact IBM adopted something similar to the 88110 local bus, itself an evolution of the '040 bus, when designing the first PowerPC). It doesn't make a great deal of sense to care about CPU specifics, particularly in moden times. Nobody's using segment registers anyway (though you still can, on the x86 or the PPC).

Even as an Engineer, it took me awhile to really accept that x86 was the only remaining architecture that made sense: 10 years after C= failed I had two different PowerPC machines sitting next to my Amiga and PClone. This was driven, though, not so much by cognative dissonance as my desire to see engineering win over marketing. And it ultimately did.. just in AMD, not Motorola/Freescale or IBM.

-Dave

@hazydave

>Even as an Engineer, it took me awhile to really accept that x86 was the only remaining architecture that made sense

So...in your opinion IA64 will go away?

@Turrican3

Quote:

Turrican3 wrote: I agree, but I'm afraid chances M$ will permit an "official" port of OS4 to their console are less than zero, even if we consider how poor is our OS, feature-wise, compared to Windows. On the other hand, PS3 could be a very different affair... well at least that's what I hope.

Console manufacturers don't want an OS on these things, the only reason they want to sell them is so they can sell games, they definately do NOT want people buying these things for any other porpose, they lose money on 'em.

Last edited by T_Bone on 17-Jan-2006 at 10:21 PM.

_________________
"If you want to tell people the truth, make them laugh, otherwise they'll kill you." - Oscar Wilde

@Leo

Quote:

Because there's no PPC processor maker focusing (interested) on desktop anymore.

Spot on!

Well, this was the case BEFORE Apple switched to x86 so we know there will definitely be no suitable chips for long now.

I wouldn't be surprised if OpenPPC.org were not just a cynical ploy to get rid of (What should have been Apples) old stock of G3/G4/G5 by IBM.

Look how much trouble Elbox have had securing a basic chip for their Dragon production.

Much as I'd love to see PPC go on into the future, it's not going to happen except in embedded devices -

PPC got sick when Apple was forced to ship Intel servers.

PPC died as a usable, high performance, general purpose desktop platform the moment Apple switched. If you can't accept this then you can delude yourself!

It may take six months, or 2 years but soon, everyone in the PPC desktop market will have exactly the same problems Alan had with G4 and Elbox are having with CF.

The crazy thing is that IBM and Freescale couldn't be bothered to keep their one and only flagship desktop customer happy.

@syrtran

Quote:

First, I want you to know that I am a big admirer of your work (looks over at 3000 sitting unpowered next to A1 that is slowly replacing it).

Thanks...

Quote:

That being said: Quote: I think the big problem is that Hyperion's chasing this dragon of PPC, and the hardware's not something they really do control well. They could have done a straight port to the PC in much less time... You've repeated this a couple times in this thread. I just want to say: WRONG! (Boy, first I'm, er, "discussing" with Bernie, now, you. I'm taking shots at all my heros, today )

Hey man, you're entitled to your opinion, and don't let anyone knock you down for it. But hey, you're also

WRONG!

Keep in mind, I'm one of the few people on the plant who's brough up new desktop-class operating systems on new hardware, cold. I never actually did this at Commodore/Amiga (of course, the Los Gatos guys did) -- we had a working AmigaOS for new hardware, working hardware for a new OS. But since then, several times.

Quote:

To start off, I'm a professional programmer. I've been a professional programmer for 25 years. At one point I was sysadmin on a small IBM mainframe (a 4361, if anyone cares). I even got to do sysgens. I started programming in high school in 1972 on an 1130.

I wasn't in High School in 1972, or using computers. A year later, I started teaching myself programming, first on a programmable HP desktop calculator, next on a CDC Cyber 72 on my Dad's account (he was a department head at Bell Labs).

Progamming lets you get close, but make no mistake -- OS Systems programming is unlike anything else. For one, it's the only place in programming where the hardware is a real issue.

Quote:

And I still get torqued when non-programmers decide that a software project should've happened faster.

You should. But just in case you don't know, I've been programming since 1973, I have a degree in Mathematics/CS from Carnegie Mellon (it's a double degree, I also did Electrical Engineering, and nearly made a triple with Cognative Psychology, only I would have had to stay an extra semester to schedule things, versus what actually happened: underloaded last semester, just graduate Robotics and Compiler Design).

I know OS design as well as most of the people I've worked with. A few know it better; they're usually the guys I'd prefer to have working on my project, but they're not always available. They don't know hardware. So I do the hardware. At some point, probably when I retire, I'm going to design a CPU, then a system around that CPU, than an OS (based on a hardware supported object model I've had in mind for 10-something years), and then put it all in a book.

Quote:

Even though I've been a hardware hacker on and off those past 30 years (I built my own -working- parallel port card and 4k RAM card for a Heathkit H8 I built way back in '80 - from individual components, I might add), you don't see me telling you that the PIOS-1 should've appeared 6 months sooner. You don't hear me telling you that you take too long developing your digital R/C controls (you work for Futaba?). That's because I -don't- know how long it will or should have taken. While I -understand- the process, I don't -live- it.

Good for you. I have lived it; probably more than you, unless you actually have been designing operating systems. Before Commodore, my job at General Electric was software-only; I was maintaining a compiler/simulator for the ISPS language. I didn't do any of this at Commodore, though I did a few OS-level hacks (SetCPU, for example) that no one else managed to get working. At Scala, I did nothing but software, working with Mike Sinz on the OS technology, as well as designing the object description language and the compiler and database for MMOS. At Metabox, as well as designing the Metabox 1000 main board, I wrote various pieces of the OS.

So please, don't speak to what you have no clue about. And again, unless you've actually been designing operating systems and hardware, you may know programming, but you don't know what I do.

Quote:

This is the kind of attitude that I would expect more from the PHBs you had to deal with back in the "good ole' days" than from the h/w guru who did the dealing.

Again, you don't know the breadth of my experience. I don't make specific comments about what I don't know, in this instance, possibly better than anyone else on this board.

Quote:

What you are suggesting is that the OS4 port could have been rewritten, almost from scratch, without endian problems, without CPU dependencies, without Agnus/Alice dependencies, without CIA dependencies, without your own Zorro 3 dependencies, in less than five years.

Without question. And your poke at me it telling. For one, this isn't even remotely "from scratch". Endian problems are solved by the HAL -- Microsoft/Dave Cutler did it, after all. This is not rocket science, but it is proper engineering. CPU dependencies are largely kernel and compiler issues. There are no Agnus/Alice dependencies outside of the graphics, and that's already been done, several times. One guy -- Oliver Bausch -- did his own RTG library, even a year or so before the other guys were shipping non-Amiga graphics cards. And 3.x made it even easier. The CIA dependencies are non-existant, particularly after the "Jumpy the magic timer" technology was designed for 2.0... a simple bit of recoding, and your new timer does the job. There ARE no Zorro III dependencies anywhere in the code: the Zorro III support module is independent of all else. You'll find much of that's prevalent in AmigaOS code.

Five years! Again, the orginal didn't take five years, why should a port. At Amiga Technologies, Andy Finkel and I had a PowerPC Amiga proposal. The OS port was to take slightly over a year, given a full staff. About 25%-30% was devoted to HAL.

On another project, Andy, Carsten Scholte, me, and the rest of the team at Metabox implemented an AmigaOS clone (like Amiga enough to tap things like Voyager and MUI without drastic problems, but also fixing many problems), on the Coldfire 5307/5407, in about two years. This didn't include everything you got on an Amiga, but it also came with the Web brower (V), email, graphics with integrated video windowing, DVD, DVB, and PVR applications, analog TV tuner in a window (I wrote the tuner driver myself), TCP/IP included, etc.

In other words, you don't have the slightest concept how long this kind of development takes. You don't do this work, at least from what you've told me. And you don't appreciate how much work the kind of engineers C=/Amiga, Andy, and myself hire. When I get four hours of sleep, it's a quiet night.

Quote:

Linux took longer than that.

Linux is a really bad example, and the fact that you mentioned it illustrates your lack of understanding. "Linux the Kernel" was written by one guy, as a hobby, over a few years before he even released it. Until recently, it's been hobby code; all volunteer. Now, go look at a Linux distro -- the last few Skal sent me occupy several CDs or even a full DVD, and if that's the modern definition of Linux, you have about 50x more code than any AmigaOS release, including what Hyperion is trying to do.

Quote:

Beos took longer than that.

I guess you're not familiar with the difference between DESIGNING AN OS FROM SCRATCH and simply porting one. And I do mean simple. BeOS was released to developers in 1995; they started in 1991, from scratch. In the process, they designed one new computer (based on the AT&T "Hobbit" microprocessor and the DSP3210), scrapped it, designed another computer (Joe Palmer's original 66MHz BeBox), and ported it. So there you have it -- that whole 4-year process includes a port from one hetergenous multiprocessor system to a radically different homogenous multiprocessor system.

Is this helping? Are you starting to understand the dramatic difference.

Quote:

Heck, even Windows NT took longer than that.

Well, do you really want to know the story. The NT group (composed of many VAX/VMS people from DEC) was assembled in November of 1988. The were initially targeting the Intel i860, but running code in a simulator. By the end of December, they had substantial kernel bits working; by April of 1989 they had a CLI-type system up and running, still in simulation. This team consisted of the five guys from DEC and one original Microsoft guy.

By late 1989, they had about 35 people, but keep in mind, this was all kind of from-scratch stuff. Microsoft didn't have a TCP/IP stack, for example. At the end of '89, they did their first port -- from the i860 to MIPS R3000. In less than three months, they were running well on real MIPS hardware (this included a flip from little to big endian mode). In 1990, they started the '386 port, and by September, they had designed and implemented much of the Win32 top layer. Only then did the group get huge -- Microsoft decided to back NT rather than OS/2, and assigned 300 programmers to the project.

By the time (1995) they got to porting to the PowerPC, they were porting in a month or two. Actually, they spent twice as long waiting for IBM to deliver hardware as they did actually coding the port. NT's HAL basically means that, unless you have a bug in your code, there are not endian issues, no alignment issues, etc. Some of the more basic concepts of this were used in AmigaOS all along.

Quote:

And those are from companies/groups that had something Hyperion doesn't: lots of programmers.

Well, it's foolish to take on a project you know you can't finish. I don't know how many programmers they have. Microsoft started NT with 6 people, and while they grew in number, the basic OS stuff was done, again from scratch (eg, there's lots of design time) in less than two years.

Quote:

Where's SkyOS? Where's AROS? Where's Syllable? These projects have taken even -longer- than OS4. Why? Lack of resources. Lack of programmers, money, and time.

Open Source projects such as AROS only require interest, money doesn't actually help, even if it did exist. More importantly than anything is knowledge -- there are many people who think they know how to design or even port an OS. Most of them are wrong.

Quote:

I don't know how much of the history of OS4 you know about, but there were times in the last 5 years that Amiga Inc. appeared to have -no- money to give Hyperion. During those times, work on OS4 obviously stalled.

I wasn't even remotely suggesting that their hardware issues were their only problem. But they have had hardware issues, they continue to have hardware issues, and the Hyperion people -- the guys actually doing the work -- have said as much.

To suggest their choice of hardware, based on a custom PPC motherboard that also had to be brought up, was wise is the proclamation of the ignorant. If you're going to do a port, your best bet, time and money-wise, is to port to something that already works, and is available in volume. That's one reason NT's first port didn't take long: NEC had all the MIPS-based PCs you could ask for.

Quote:

I still respect you, but don't go telling programmers how to program, especially when it's not putting food on the table.

Having had more OS experience than, I believe, any of the programmers involved, at least before they began the job, I believe I'm qualified. I have no reason to believe you have a clue about this kind of work -- you've demonstrated ignorance of the nature of it so many times in this one message, it's not funny. And I'm really not here to jump down your throat.. but I am from Jersey, we call it as we see it, and in the long run, you'll benefit from honesty more than P.C. nonesense.

-Dave

[Edit: fixed quote]

Last edited by _Steve_ on 17-Jan-2006 at 11:23 PM.

@hazydave

Yeah - But can you complete Turrican?

_________________
Put the needle to the record when the drum beats go like this...

@hazydave

Hey Dave! I salute the Uncle of Amiga!

First; It has got to be a tribute to your work that there is currently a ten year old A4000T on Ebay threatening to sell for £1000! It's amazing how much A1200s and A600s sell for - not just to 'crazys' like us but to ordinary people who remember the glory days of the Amiga.

Second; (Sorry it's just as OT as the First) Have you checked out Dennis from the Hollands' minimig over on Amiga.org. Basically this guy has in one year, learned verilog, studied the commodore books and UAE/Fellow and has produced a working OCS A500 (+scandoubler) on a XILINX FPGA (Sans 68K!).

Amiga.org:Minimig

I was wondering what you think of this achievement (if you haven't already seen it!)

@wegster

Quote:

Quote: PearPC is not the fastest FOSS PPC emulator i.e. refer to SheepShaver. Yes, and I believe you and I had this discussion already somewhere on AW, where I posted the relevant figures instead of a one line response..ie refer to....complete arguments vs one-line responses.

It should be noted, however, that SheepShaver is only a PPC *usermode* emulator. Most significantly, it doesn't support the PPC MMU; This would make its code a good choice to run existing PPC apps on a non-PPC OS, but it won't be any good for running any non-obsolete PPC OS on a non-PPC machine.

apple did go x86 and the ppc area is over now. now soon there will be a problem to get ppc chips... i hope just hope there will be enough for the ack ppccard. so i can use my a1200 with aos4.0.... when aos5.0 will be here in the future, i expect it will be powered by an x86 like the new intel macs are today...... if not aos is dead by that time...

_________________
I was a ZX Spectrum owner....

@d0c

Apple have never used the chips that will be on the ACK ppc card.