Dave Haynie (lead engineer of C= Amiga) opinion on Amiga Successors

[uncharted]

bhoggett wrote:

If you want to read people ranting on ad-infinitum about things they know absolutely nothing about, read an Amiga forum. Nice to see some things never change.

But doesn't that give a nice warm fuzzy feeling inside? :-D

[Hammer]

Sad really can`t agree with him. My G-3/600MHZ Pegasos2 running Mplayer on MorphOS, plays Divx video, that my Athlon 1GHZ WinXP locks up and can`t handle.

Just don't run DIVX-HD on G3 processor.

Cheaper is not always better. Apple makes the switch to jump on the money train. Eating their own words that the PPC is better. Sony doesn`t seem to want to switch. They seem to produce tons of PPC machines that sell less than $150. One word "CELL"

Pitiful VMX performance (e.g. Alias cloth simulation). Hence SJ’s statement on Cell being less effective than G5.

Refer
http://www.research.ibm.com/cell/whitepapers/alias_cloth.pdf

Using 8-way SPU’s 5 time over 1-way Pentium IV is unwised since this didn’t factor in Intel’s Tigerton.

In "Porting to the Cell Architecture"

Porting to the Cell Architecture

The most notable features of the Cell processor architecture are the IBM® Power-based RISC
CPU core (PPE), the 8 SIMD vector units (SPEs) and the impressive bandwidth that can be
leveraged to obtain maximum performance from the architecture.
In order to port the prototype cloth simulator to the Cell processor, the code and data structures
had to be restructured. First, changes had to be made to allow for parallelization of the algorithm(1).
Even for more conventional multi-core targets, this is a challenging process. With the Cell
architecture, there is the additional requirement to manage memory transfers between the main
processor (the PPE) and the SPEs. Second, where feasible, scalar code was re-written to take
advantage of the vector support in both the PPE and the SIMD units.

It seems that Pentium IV’s code was not optimized, since the original code needed to be  restructured for vectorization and parallelization i.e.
such optimizations also benefit Pentium IV side (1) i.e. for SSE2/SSE3 and SMT.

In Alias|Wavefront Maya 6.0 Zoo Render benchmarks, Opteron @2Ghz easily beats Intel Nocona Xeon @3.4 GHz. In most cases, Yonah is clock to clock similar to Athlon 64 X2.  

[A3KOne]

You know... one thing that amazes me is that the fanatics will start off saying how fast and responsive and powerful their Amiga with an 030 (powered by a squirrel in a cage) processor is, and then will use Cell processors or some other incredibly overpriced -yet obsolete by the time it is mass marketable- CPU as justification for sticking with PPC tech.

Reality:

AmigaOS ported and running native on a $300 PC would run circles around any PowerPC alternative currently available.  

A specifically chosen aftermarket Amiga seller (maybe one on each continent) could build a single system based around a specific motherboard and gfx chip combo specified by Hyperion, and the hardware support issues would be gone.  Future upgrades and additional gfx card alternatives could be handled via RTG drivers released for the specific cards.  The Amiga 1000/500/2000/3000/1200/4000 did not support a vast lineup of gfx cards. Commodore did not sell motherboards stand alone unless it was a replacement part. With the price of PC hardware being as low as it is, you could replace a motherboard 4 or 5 times for the cost of a single PPC A1.

It is really simple.  An official Amiga x86 package is selected and released.  It comes with everything in the box to work when you plug it in.  It could have a custom case, or there could even be multiple solutions released...IE: an A1200 style system with a fully integrated motherboard and a big box system with the supported gfx card on a high speed pci or agp bus.  
Upgraded gfx cards could come later as drivers are written for newer hardware.  Amiga dealers could sell the off the shelf PC GFX cards at the same price as the PC dealer + a nominal fee for the driver.  You would not have to support every chipset on the planet.  The systems would be low cost and Amiga users could purchase them for $500 or less, loaded and ready to go.  Since the hardware would be bog standard PC stuff, if a user desired to purchase the hardware and build their own system, all they would need to do is buy the OS.  If they are upset because their Funkytronic xs5000 gfx card is not supported, they will have to deal with it.  AmigaOS 3.1 did not support anything but AGA/ECS. At least OS4 has RTG built in for future expansion. The reality is, the prebuilt system will be just as cheap as do-it-yourself because the Amiga builder will be buying the components at an OEM price. He can sell it and make a profit without killing the consumer.  Lot of successful PC companies started this way...Dell comes to mind.

Some people will immediately scream WHAT ABOUT PIRACY?!?
Well...newsflash! Piracy may or may not happen...it probably will to some degree.  The fact is, if you sell 100,000 copies of OS4 and 10,000 are pirated, you have lost some money, but it is still better than selling 5,000 copies with no piracy.  In all likelihood the same number will be pirated regardless.  The l33t h4xX0rZ will be l33t h4xX0rZ regardless.  The difference in piracy numbers will be nominal...the difference in sales will not be.


100% of new users could not possibly care less about what CPU the computer uses. They just want it to work and work well.  An x86 Amiga could and would work well.

80% of Amiga users would immediately be happy with this solution.  They would have a fast an powerful Amiga.

10% of Amiga users would be happy with this solution once they used it and saw how fast/powerful it is.

10% will not be happy unless Jesus raises Jay Miner from the dead so he can design a new Amiga complete with new custom chips, that is then built by CBM/Amiga at their plant in Germany using chips fabricated by MOS technologies.

That is the truth as I see it.

[Tigger]

 Commodore did not sell motherboards stand alone unless it was a replacement part.

Actually as someone who bought thousands of motherboards from Commodore, I'm going to have to disagree with you on this point, however everything else is pretty much right on.
    -Tig

[A3KOne]

Tigger wrote:

 Commodore did not sell motherboards stand alone unless it was a replacement part.

Actually as someone who bought thousands of motherboards from Commodore, I'm going to have to disagree with you on this point, however everything else is pretty much right on.
    -Tig

I think you misunderstood what I was saying (or trying to say) about the motherboards.

There were not 500 different motherboards.  Sure there were revisions, but they were the same motherboards with bug fixes...just replacement parts. There were a couple of exceptions such as the short lived A3000 030/16.  You would not buy a new motherboard from CBM to put in your A3000 because it was a huge upgrade. You might replace the buster because of flaws in the original.  There was never a "WoW! I gotta run out and buy that new motherboard for my Amiga 500!"
Nor would you buy an Amiga motherboard to install it in the new case from brand-X and build your own...unless you were a hardware hacker.  It was not like PC's today.

I doubt you ever had to install new drivers because you changed motherboards.

[Oliver]

Have to agree with A3KOne about x86 amigas.

edit- actually, I know a lot of people these days whos home computers are mostly just for fun.  People who work all day with windows, and don't really want/need to use windows when they go home at night.  Some of these friends still turn on their old classic Amigas just to play and tinker.  I asked one of my friends why he used his A3000 in this fashion, instead of his all powerful PC sitting right next to it, and he just said "nah, windows: use it all day".  I think these people would be happy with an x86 Amiga OS.

[coldfish]

I cant agree with Dave Haynie more!
 
As a bonafide C= hardware engineer, is there anyone in Amiga-land who can counter his argument with more authority?

It's kinda nice when it comes from straight the horses mouth isnt it?

[DonnyEMU]

Responding to the apples and oranges of Amiga comparisons to PCs. Some people are just religious. I bought my Amiga long ago in 1985, after I saw it's multi-tasking, it's amazing graphics design software, and after seeing it genlock and realizing that I could use this as a great video editing and production platform.

We were told this is what the Amiga was all about. What passes for Amiga today is just an OS and those amazing capabilities have been dwarfed by others. There is no Amiga equivalent to Media Center or Apple's Front Row on the platform. The Amiga doesn't know what 1080i or 1080p is in hardware support these days.

AMD and ATI and NVidia are already doing what mr. R suggests, in that they have a 64 bit cpu in it's own native mode that does all of this stuff already. To say that it's not happening is just not really the case.

As far as TV stations using Amigas etc. They aren't using PowerPC based ones they used the other ones, I worked for several of them. But this isn't really happening anymore because they can't get replacement parts and they are all becoming antiquities thanks to companies like NewTek Scala, who did this stuff on the Amiga and now do it on modern PC hardware.

If the Amiga is to regain it's nitch in these areas, a lot of hardware and software inventiveness would have to happen. The old companies surely ain't coming back for something they have done elsewhere including ancient powerpc macs..

Aros is a great OS people need to write programs for it and innovate with it, if the spirit of the machine is to survive. Apple's new Intel machines are 4x-5x faster than the powerpc versions because the development is just that far ahead. The platform needs cheap cool new cutting edge hardware and a new direction, if this is to come back.

[A3KOne]

Oliver wrote:
Have to agree with A3KOne about x86 amigas.

edit- actually, I know a lot of people these days whos home computers are mostly just for fun.  People who work all day with windows, and don't really want/need to use windows when they go home at night.  Some of these friends still turn on their old classic Amigas just to play and tinker.  I asked one of my friends why he used his A3000 in this fashion, instead of his all powerful PC sitting right next to it, and he just said "nah, windows: use it all day".  I think these people would be happy with an x86 Amiga OS.

You know, if done correctly, an x86 Amiga could turn the tables on Windows and do some of the same things to Windows that has been done to Amiga.

Think about it... implement Wine on AmigaOS and have Windows apps running in a windows compatible API...or make a sandbox environment where Windows can be booted within AmigaOS.  It would be easier than UAE because there would be less hardware emulation required.  With the bright minds in the Amiga community, it would be interesting to see a WINE port.  I bet it would be cleaner and more usable than the Linux version.  Amiga Installer could be used to take care of all the config issues so it would not be as complicated as the Linux version.

This would be important because the AmigaOS would, in a sense, be tied to the hardware.  The difference in the x86 AmigaOS would be that the extent of the hardware dependence would be minimal... Chances are that if someone wanted to upgrade speed, a new motherboard would exist that would be 100% compatible.  The only legacy hardware would be the gfx card.  support for integrated ethernet and ac97 would not be a big deal since there are not that many commonly used chipsets.  When the time approaches when a significant speed upgrade is needed, Hyperion or Amiga or whoever could find a motherboard that is reasonably close, release a patch or two, and make that motherboard the certified Amiga board.  There would probably be numerous PC hardware manufacturers willing to provide docs to get the additional business.  Again, the key is quality and price, not quantity.  There is no need to support every single motherboard combo.

If there is an A1200 style machine, that is even easier.  There are a limited number of Mini motherboard providers.  Amiga companies seem to like strategic partnerships and are good at making them.  There would be an opportunity to have one that is worthwhile.  Via makes good boards and they are reasonably priced.  There are others...

To me, it seems insane to go the PPC route at this juncture.
In 2000, I was all for it, but now it is the wrong thing.  PC hardware has advanced a great deal. Legacy ISA is gone.  The PC is much closer to Amiga in hardware philosophy than it used to be.  When the AmigaOne was first announced, the most common Mac was a G3 or a slow G4. Speeds of 1Ghz were rare in the PPC world. The AmigaOne was already behind, but not that much.  The PC world was not that far over 1Ghz, either...2Ghz systems were extremely expensive.

In 2005, 3Ghz on a PC is cheap and the AmigaOne is still stuck at 600-800Mhz.  Even Apple has bailed on PPC.

Hyperion said a long time ago that they were moving the OS to a much more portable language.  It was said that porting to other architectures would be easy.

If that is the case, it is time to consider it.  We need to swallow our pride and realize that the PC world evolved past us.  It is foolish to forsake powerful, inexpensive hardware, for overpriced pocket calculators.  Compare the Amy05 to the Dell Axim x51v, which is a 624Mhz PDA with integrated 802.11b, bluetooth, 256Mb rom, 64Mb sdram,16Mb video ram and is multimedia capable, and pocket calculator is not far from the truth.  

I know the zealots are probably fuming by now... I used to be one of you and I would have been fuming too.  I finally woke up and realized that computers are tools to allow me to do things. Amiga just happens to be a more fun tool to use for many.  The OS is what you see and what you use...the hardware is only as imporant as its performance and price. The fact is that the current hardware options AmigaOS has  are outdated and have little to no support.  At least main stream PC hardware would have a warranty that would be easy to take advantage of.  If you are in the USA and have trouble with your AmigaOne (which appears to be common), good luck getting it fixed.  By the time you pay to ship it overseas to someone who can fix it, you could have bought another PC motherboard...and like it or not, the Amiga is a PC...personal computer.  Some people have forgotten this.

Hyperion...please...if you read this. Move the desktop OS to x86. Read what I said and consider it...I believe that is the only hope for a successful return of Amiga.


edit- one other thing...I believe the only hope for success is if the system has the name Amiga.  There is still value in the tired old girl, believe it or not.  While Aros and Morphos are worthy efforts and nicely done, they will never experience the widespread acceptance that Amiga potentially could with the release of a new hardware/software platform, even if it is x86 based.  Aros and Morphos are great for hobby computers if that is all you want.  If there is any hope of a return to levels where major software houses port recent software, it lies solely in the hands of the name that started it all...Amiga...and in my opinion the AmigaOne and Amy type names need to go the way of the dodo.  Any new hardware/software platform should return to the naming scheme from the old days... something like...Amiga 1100...Amiga 510...Amiga 610...Amiga 2200...Amiga 3300...etc.

[yester64]

Mmm.. what i have to say to this is that all follow up incarnations of Amiga, are not real Amigas to me.
I admit, that people working on their own to get it to the point of beeing Amigalike, but there is no developed hardware to it.
Besides, what it would need is software like EA's Amiga Deluxe Paint etc.
To me, the Amiga died that very day Commodore went bankcrupt. From that point only if you were a diehard you sticked to the Amiga.
Linux is nice as an alternative, but it really lacks userfriendly support. If someone has no freaking knowledge about compiling he/she has to stick with what ever comes with the disk.
And the linux community, well most have the "read the stupid howto faq" attitute, which makes you hate linux pretty much.
In my case, i sold my Amigas long ago and have no software from the time (i had a lot). The only Amiga, i call an Amiga, would be the original Amiga. But then you have a 12 year old hardware (or older) and don't have upgrades.
Like i said, if you a diehard and you really love the machine and like to work with it, its cool.
But otherwise you better stick with a PC (or a Mac).
MorphOS is a nice attempt, but somehow its just teasing the Amigafeeling.
Don't get me wrong. Its nice to see development. But i believe there has to be a leading hardware before and even the Pegasos board is not a leap ahead of times, considering that PCIx is to replace AGP and even Apple says goodbye to PPC G series. :cry:

[Piru]

@yester64

If someone has no freaking knowledge about compiling he/she has to stick with what ever comes with the disk.

Hmmm, and he/she can't use binary packages?

Last I checked I could install and keep uptodate linux without any compiling.

And the linux community, well most have the "read the stupid howto faq" attitute, which makes you hate linux pretty much.

Yeah well, at least the faq is there to read. Agreed some of the community are a bit rude, but I've always found help when I needed it.

Still they do have a point: If you're never going to try to solve the problems yourself, you're never going to learn either. And, if you're not ready to learn, then Windows XP is better for you... :-)

[Oliver]

@yester64

Can't really agree with your take on the need for Amiga hardware development, in order for it to be an Amiga.  The old hardware, in most facets, has been superceded many times over, and what people seem to miss most is a damn good operating system.  Needs for specialised harwdare can always be met with periferals, when the core architecture is sufficient.

The OS was fast and efficient.  It was simple for normal users, and also allowed advanced users to take a hands on approach.  It was not overloaded with vast amounts of code to autoconfig and auto cache your machine to death.

The old Amiga hardware provided performance and features which were unseen at that time.  Furthermore, there was a lot of competition in the hardware and OS markets at that time.  That level of competition doesn't exist at the moment.  As the x86 hardware has had so much development, it would be foolish not to leverage that.  I spoke with a microelectronics researcher, who worked for Intel last year, who said something equivalent, in regard to transistor technology.  SiCMOS is not the most advanced transistor technology around, but when so much investment and development has gone into it, it just makes sound commercial sense to keep using and developing it to its limits.  This is done to great success, regardless of SiCMOS not having the greatest potential of all transistor technology, but just because its so well established.  Same thing applies to other technologies.  If there is good hardware available, that already has a lion's share of the market, then it makes commercial sense to leverage it.  o/w it's just too hard to compete.  To develop a complete platform of custom hardware to compete with highly specialised multibillion dollar industries just isn't feasible.  As there is currently powerful hardware available, which is so widely used, there just isn't the need (or the opportunity) for a custom hardware platform.

Furthermore, users demand compatibility as much as anything else.  This is most easily and cheaply achieved by using a common basis for a new platform.

Plenty of people dislike windows, and personally I know a lot of ex-amigans who were interested in the idea of a modern amiga, but just couldn't justify the investment in specialised architecture.  Just like Haynie said with his argument of a threshold for investment.  If I could run a well developed and commercially supported Amiga OS on my windows box, absolutely I would, and I know a large number of others who would too (maybe around a hundred friends/aquaintances have said as much to me).  Custom hardware just doesn't come into it.  I don't care if it's 68k, AGA, Zorrow or whatever: I use the OS and apps, I look at the screen, listen to the speakers, and touch the keyboard and mouse.  To my thinking, the experience of the OS is probably the most critical factor in making a computer an Amiga.

If an OS seriously outperforms others on the same architecture, with a good set of features, it could really generate interest.

Well, I know this has all been said before.  Sorry, I couldn't help myself.  Actually, this is what I thought years ago, before the AOne was ever made.  From what has been said on this forum, the AOne seems to be a reasonable platform in its own right, but one can not possibly say its a commercial success.  I also seriously doubt that 'Amiga anywhere' will be raking in the cash either, and certainly doesn't do anything great for technology or users.

[Oliver]

Piru wrote:And, if you're not ready to learn, then Windows XP is better for you... :-)

 :lol:  :lol: Hahahaha  :lol:  :lol:

Actually, I really believe that intellectual laziness is one of the main reasons for windows success.  If something doesn't work, just:
1) Delete everything and reinstall windows, cos chances are that windows has destroyed itself anyway
2) Buy all the latest hardware and software, cos 800MHz/256MB just simply aint enough to run a word processor

Hardware and software vendors just love the perpetual re-investment which is driven by shoddy design.

[Crumb]

@thecrow

Even my p2 266 plays divx at decent rate.

Yeah! Sure! I watch 640x480 DivX on my pentium75 too ;-P

[dammy]

edit- one other thing...I believe the only hope for success is if the system has the name Amiga. There is still value in the tired old girl, believe it or not. While Aros and Morphos are worthy efforts and nicely done, they will never experience the widespread acceptance that Amiga potentially could with the release of a new hardware/software platform, even if it is x86 based. Aros and Morphos are great for hobby computers if that is all you want. If there is any hope of a return to levels where major software houses port recent software, it lies solely in the hands of the name that started it all...Amiga...and in my opinion the AmigaOne and Amy type names need to go the way of the dodo. Any new hardware/software platform should return to the naming scheme from the old days... something like...Amiga 1100...Amiga 510...Amiga 610...Amiga 2200...Amiga 3300...etc.

Which has been proven untrue by A1 unit sales.  The "Amiga" market is not a financially viable market.  

Dammy

[Oliver]

Which has been proven untrue by A1 unit sales.  The "Amiga" market is not a financially viable market.  

Dammy

Actually, A3kOne only said it would be necessary to use the Amiga name, he didn't say it would be sufficient.  The AOne is not a product along the lines of what he was advocating.