@Hans

Quote:

Unless you have tests that you can confirm is on 1080i H.264 AVC video ("Full HD" does NOT mean 1080i to everyone), with no frame-skipping, no audio skipping, and no downsampled output, then "pooh poohing" my turion 64 comment is a bad idea. Yes, it's a different processor (and no Hammer, lack of SSE128 isn't a deciding factor here, especially when we're talking about G4 PowerPCs), but the complexity of the decoding is the same in both situations. Don't expect a 1.5 GHz PowerPC G4 to perform miracles.

Well, 1.5 GHz PowerPC G4 has a FSB issue.

My downclock 1.5Ghz Core 2 Duo P9500** (in single core mode) smacks my old K8 Turion MT-34 @1.8Ghz in H.264 media playback. AMD K8 is not competitive against Core 2. **Penryn SSE4 includes video related instructions i.e. H.264 is one of it's targets.

Last edited by Hammer on 26-Sep-2009 at 03:06 PM.
Last edited by Hammer on 26-Sep-2009 at 03:00 PM.
Last edited by Hammer on 26-Sep-2009 at 02:42 PM.

_________________
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)

@feanor

I'm a PowerPC fan myself, but I am sort of skeptical about the future of PowerPC on the desktop as the leaders (IBM and Freescale) of that technology don't seem interested in that space, which seems to make it hard to develop for if you have that sort of friction. It would be great to have a standard motherboard configuration with the expected peripherals and interfaces that makes the board easily expandable. I think one of the problems in the Amiga community is the expectation to support every sound, graphics and network card out there, but I think if you can get great support on the most popular ones, you can then get support for the others as time goes by. A lot of collaboration between the Amiga community and the developers would make your sort of endeavor possible and successful even if Amiga is considered a hobby now. Bringing back the computer hobby excitement of the mid-70s to early-80s can be successful in this day and age of said market; you just have to figure out how to approach it.

Last edited by AmigaMac on 26-Sep-2009 at 03:49 PM.

_________________

@billt

Quote:

Is this what AMCC refer to as Titan for some time now?

Nope, this is a new IBM design.

_________________
Whyzzat?

@feanor

First, welcome (again) and thanks for your support of the Amiga community.

You mention Debian. Well perhaps you could also help us a little here. We have a member here, Geri, who is working on updated Linux kernels for the A1 hardware and i am working on building the installer ISO for the A1 and possibly the Sam in future. So a man who likes his PPC and Debian is good to have around!

Now for the PPC board the first priority is quality control and everything working. And once you have that then you can make a batch at the lowest cost possible. On the A1 a capacitor/resistor component error caused USB hot plugging not to work. IDE and net conflicted. And IDE DMA modes malfunctioned caused by bad hardware, unavailable documentation for drivers or a combination of both.

Now, unfortunately even though Eyetech dumped their original AmigaOne design (PredatorSE IIRC) in place of a full stand alone ATX board, it was in no way perfect. If it was fine like the Sam mostly is then there would be no issues. But I have a point here, Eyetech set out to make batches of about a 1,000 each. Now, for the A1 board where it wasn't proven, a short run was a safer option. If a run of 1,000 boards cost $1,000 each to manufacturer that's a million dollar investment. Now, I don't know the business like you, but I think if Eyetech could have made 10,000 boards instead and brought the cost per board down to $100 each then an even better price could have been offered to the user.

Back then we had about 10 times the users or more interesting in investing in an OS4 machine and if a good one could have been made at a cheaper price then the platform would have been in better shape now. Actually, even with those problems on the original, a cheaper price would made have them almost redundant. In the low hundred dollars we wouldn't have cared! But that was then and this is now. If even that was possible.

However, it would be good, to work in reverse, and first propose the lowest price possible with a large bulk order. More people would be interested this way and more people likely to pay a lower price. To cover yourself you could do what happened with the A1, where we had to prepay. In the case that a large number of users were interested and this was possible I'm sure Amiga users would be happy to invest. Now since you are also interested in non-Amiga markets, hey, maybe this kind of thing could be possible? A modern "PowerStation" machine to fill the gap for Haiku, Linux and Amiga platforms would be cool, but since that is G5 territory let's not go there yet.

Another possibility, if it could be realised, would be to design a board without a CPU! Now that would be cheap! This board would be similar to the A1 with a MegArray connector. However, this time it would be compatible with a Mac CPU card. Just grab the fastest Mac G4 you can find and put the CPU in place! Fastest G4's possible and without needing a PPC on the boards you could make them really cheap!

Just need to scrap a few Macs there.

Now, we we know PowerPC hardware is getting rarer. PA Semi looked like a good PPC CPU and Amiga even made an announcement about using one in a machine a while back, until Apple decided to buy them out and kill the PowerPC. in all markets. Note that was a dual core 64-bit CPU at 2GHz. I'm not even going to compare with an x86, but that was two years ago, and if we can't get close to that there is a problem. A PowerPC CPU that can't compete with itself! We also saw the PowerStation. The first ones I saw were running 4 cores at 4Ghz! Enough CPU power to even kill a PC! But now the speed is dropped to half. What is going on? It seems that for PowerPC the speed is getting lower instead of higher as in a normal CPU progression. Has the curse of the 500Mhz ceiling come back with a vengeance? The only thing PowerPC seems to be increasing here is entropy. For a CPU design that was more modern than the current offerings that doesn't make sense. But Intel do seem to have a Mafia on the market.

So, for our market, we'd need at least 2Ghz and AltiVec. Otherwise, we couldn't make use of your library. Hyperion were talking about SMP support but I could see that breaking more things like the new memory system did, even so we'd welcome it! The rest seems fine and the price looks good from what I can tell. But, the "modern" PCI slots, well we need more! At least four slots I think. Amiga users want expansion. And whatever is in place of AGP these days. We don't need onboard VGA, or is it called DVI these days? External gfx should be better for you and faster for us even if we have to buy it.

I think that's about all I have to say on the subject. Well, goodbye.

@Hypex

Well your post wants me to start screaming ........

Eyetech could have ordered 1.000.000 Ter^H^H^H A1s, that still wouldn't ahve brought the price down to anywhere near 100$ (70 Euro or so). A simpler (less discrete parts) and smaller design like the Peg2 might have been made at 200$ if it had been produced in such numbers. With real numbers it ended a product starting at 300Euro (about 500$) and that was useing G3 modules probraly left over from the Pegasos1-project (which was a failure in buisness terms).

Back than there was no chance that the market would have swallowed 10K A1s at any price, cos there was no OS4, it wasn't even clear even and when there would been an OS4. Add to that there weren't even close to 10k potential next-gen-Amiga costumers.

Starting with a low price (for a mass production) and then gradually increasing it to comply with actual demand ? Now if there was ever a bullet-proof way to alienate your costumers this gotta be it !!

But you should really be smacked for even suggesting another pre-pay !

Oh and btw. did you ever investigate what a megarray connection costs even if bought in big numbers ?

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Hypex

Quote:

You mention Debian. Well perhaps you could also help us a little here. We have a member here, Geri, who is working on updated Linux kernels for the A1 hardware and i am working on building the installer ISO for the A1 and possibly the Sam in future. So a man who likes his PPC and Debian is good to have around!

I'm not sure how I can help, but Debian is high on my list of support for the board, so we might as well join forces.

Quote:

Now for the PPC board the first priority is quality control and everything working. And once you have that then you can make a batch at the lowest cost possible. On the A1 a capacitor/resistor component error caused USB hot plugging not to work. IDE and net conflicted. And IDE DMA modes malfunctioned caused by bad hardware, unavailable documentation for drivers or a combination of both.

I agree there. Documentation and schematics will be definitely available somewhere, I'm not going to say it's going to be "open source" but I was thinking something like a "shareware" kind of registration, where one could get the schematics for a small fee, and that would allow users to perhaps make a custom version -for non-commercial uses. Eg a user with access to a hw fab lab, might want to do a version with more PCI-e slots, or adding a gfx card on board, etc. Ideas are welcome.

Quote:

Now, unfortunately even though Eyetech dumped their original AmigaOne design (PredatorSE IIRC) in place of a full stand alone ATX board, it was in no way perfect. If it was fine like the Sam mostly is then there would be no issues. But I have a point here, Eyetech set out to make batches of about a 1,000 each. Now, for the A1 board where it wasn't proven, a short run was a safer option. If a run of 1,000 boards cost $1,000 each to manufacturer that's a million dollar investment. Now, I don't know the business like you, but I think if Eyetech could have made 10,000 boards instead and brought the cost per board down to $100 each then an even better price could have been offered to the user.

Economies of scale dont' work that way unfortunately, if I produced 10 times the quantity, I wouldn't have 1/10 of the cost/board. Surely it would be cheaper, but not that cheap :)
In fact, 1000 * $1000 = 10000 * $100 :)

Quote:

Another possibility, if it could be realised, would be to design a board without a CPU! Now that would be cheap! This board would be similar to the A1 with a MegArray connector. However, this time it would be compatible with a Mac CPU card. Just grab the fastest Mac G4 you can find and put the CPU in place! Fastest G4's possible and without needing a PPC on the boards you could make them really cheap!

The CPU-less board has crossed my mind, but that is even harder. Users would have to try to find their own CPUs, or worse, I should try ebay or every computer graveyard, for old mac CPU cards. What if I fail to find the cards I need? I'd be left with useless boards.

Quote:

So, for our market, we'd need at least 2Ghz and AltiVec. Otherwise, we couldn't make use of your library. Hyperion were talking about SMP support but I could see that breaking more things like the new memory system did, even so we'd welcome it! The rest seems fine and the price looks good from what I can tell. But, the "modern" PCI slots, well we need more! At least four slots I think. Amiga users want expansion. And whatever is in place of AGP these days. We don't need onboard VGA, or is it called DVI these days? External gfx should be better for you and faster for us even if we have to buy it.

you can't make an omelette if don't break the eggs! Anyway, I don't know about many PCI slots, if the board includes everything by default (SATA2, USB2(3?), gigabit ethernet, AC97 sound with SPDIF, PCI-e 16x slot for gfx card) of what use are the many PCI slots? If at all, I will probably include just one plain PCI slot, and more PCI-e 1x slots.

@Kronos

Quote:

Add to that there weren't even close to 10k potential next-gen-Amiga costumers.

Noting what Acube is reportedly doing for SAM440 sales, I think we can say it's clearly under 1,000 today. The current situation is far below any possible economic viability threshold. Lets look at what possible customers are out there. 68K guys are happy either with their functional 68K Amigas or with UAE. Retro gamers are not going to spend that type of money when they have UAE on their dual/triple/quad core beasts.

That leaves those wanting a "modern" Amiga. That goes into those who can afford high price systems (Sam440) and those who can not or will not. Those who can not or will not look towards either a cheapo Mac for MOS or some sort of EFIKA (1 or MX) for MOS/AROS, or those who rather recycle or multiboot x86 for AROS or Anibus-OS.

I just can't see the demand for it to warrant an investment of that size.

_________________
Dammy

@feanor

duel core is the future, we need the hardware to make the software.

why not use a hardware H264 processor like a Davinci media processor ?
http://focus.ti.com/docs/prod/folders/print/tms320dm642.html

Future proof parts , availability for 2 years.

Good luck with your project it will be an exciting ride :)

_________________
Download 499.26 Mbps, 659.94 Mbps Upload :)

@damocles

damocles wrote:

"I just can't see the demand for it to warrant an investment of that size."

"If you build it they will come...."

The fact of the matter is there is a large portion of the market who were/are not interested in the SAM boards because they are low-end boards. Many people have asked for 1+GHz boards. Anyone that owns a Peg-2 and many that own AmigaOnes simply don't have a huge incentive to get a SAM board, because they are slower in many areas (ok not all, but cpu is definately slower).

Anyway, I would hope that properly priced and marketed high-end board would do just as well as the SAM board, if not better.

But yes, you are right, still a relatively small but stable and growing market. OS4 development continues so that will keep things moving from the software side, too.

_________________
AmigaOS 4.x Beta Tester - Classic Amiga enthusiast - http://www.hd-zone.com is my Amiga Blog, check it out!

This is something I would definitely be interested in - I use my A1-XE every day, but I'm always worried about it dying one day. If another motherboard came along which was faster, I'd replace the A1 in a flash. Especially if it had USB2.0 (and was supported by the OS of course), because this is the one limiting factor I've found with the A1 in day-to-day use - transfers to large USB drives take aaaaages

_________________
RobTheNerd.com | InstallerGen | SMBMounter | Atoms-X

After all that happened, and the long threads about the mac mini, the future of the powerpc, and the lack of processing power that is so evident we are suffering, I though the people here started to be A BIT more smart and learned what we really need.

But they are voting 8610, instead of the 8640D. Incredible and depressing.

@vision

Quote:

After all that happened, and the long threads about the mac mini, the future of the powerpc, and the lack of processing power that is so evident we are suffering, I though the people here started to be A BIT more smart and learned what we really need. But they are voting 8610, instead of the 8640D. Incredible and depressing.

Well, I can't blame them, it's much cheaper and it's already 3x FASTER than the 1Ghz G4. So, go ahead and get the fastest Mac mini, the 8610 will still perform faster all around. After all, it the 8610 performs good, we might move to the 8640D or produce a 2nd model.

@feanor

Quote:

it's already 3x FASTER than the 1Ghz G4

I'd imagine that sort of speedup might be possible on some very specific tests. However given it has the same core as the G4 I don't think the speedup would not be anything nearly as big on most code. It will be faster simply due to the faster memory system but clock for clock I doubt it'll even get 50% for the most part. i.e. I don't expect a 1GHz 8610 to really challenge a 1.5GHz Mac mini.

Last edited by minator on 26-Sep-2009 at 09:23 PM.

_________________
Whyzzat?

@Hans

Quote:

Thanks for that. Seeing as the demo was with a DIVX file, I doubt that is could play 1080 HD H.264 AVC.

Debatable, but irrelevant. Why would you want to play full HD anyway? A Bluray player will do a much better job.

_________________
Whyzzat?

@minator

Quote:

minator wrote: @feanor Quote: it's already 3x FASTER than the 1Ghz G4 I'd imagine that sort of speedup might be possible on some very specific tests. However given it has the same core as the G4 I don't think the speedup would not be anything nearly as big on most code. It will be faster simply due to the faster memory system but clock for clock I doubt it'll even get 50% for the most part. i.e. I don't expect a 1GHz 8610 to really challenge a 1.5GHz Mac mini.

very specific tests = linux kernel compile: depends on CPU speed, memory speed, disk speed, etc. Same OS, OpenSUSE 11.1. The disk was pretty much the same speed. Quite a heavy and demanding job.
MPC8610: 1/3 of the time of a Pegasos II 1Ghz.

Also check http://freevec.org/content/libfreevec_104_benchmarks_updated for more benchmarks between the two. I'm sorry, I don't think the mac mini stands a chance, even if it does it will only be for computational tasks NOT requiring continuous memory access.

@minator

Quote:

I don't expect a 1GHz 8610 to really challenge a 1.5GHz Mac mini.

So, as usual, we are f*cked before even try, because I already have a Macmini at 1.5 GHZ and it hardly can play 720p at h264.

So forget full HD or any other MODERN (but CURRENT, not even future) computing task, like editing video, mixing HQ music, rendering 3D, playing web based 3D games, etc...

@feanor

Quote:

Well, I can't blame them,

Well, maybe you won´t, but I'll do: Our needings are SO EVIDENT, SO BASIC, and we had already so many years to notice and discuss about them, that even a 8 years old kid would know what to ask for.

The proof is that all interesting people have left Amigaland already, only retarded seem to stay here (and I will include myself if the situation continue as it is).

Last edited by vision on 26-Sep-2009 at 09:46 PM.

@vision

"because I already have a Macmini at 1.5 GHZ and it hardly can play 720p at h264." - with osx i can imagine. but you will be suprised to see how fast this mini with morphos. hd vids (720p) just plays fine according to the ones who could try. i can imagine that morphos with a 8610 can play full hd vids.

@vision

Quote:

vision wrote: So, as usual, we are f*cked before even try, because I already have a Macmini at 1.5 GHZ and it hardly can play 720p at h264.

I have a MPC8610, and 720p h264 plays without a hickup -on Linux, with the internal DIU display even. Full HD, that's different but MorphOS/AmigaOS are lighter so they may fair better I cant' say.

Quote:

So forget full HD or any other MODERN (but CURRENT, not even future) computing task, like editing video, mixing HQ music, rendering 3D, playing web based 3D games, etc...

If you really want all that, what are you doing with AmigaOS? esp with regard to 3D, everything happens on Windows and to a lesser extent MacOS X.

Quote:

Well, maybe you won´t, but I'll do: Our needings are SO EVIDENT, SO BASIC, and we had already so many years to notice and discuss about them, that even a 8 years old kid would know what to ask for.

Ok, here is the situation. You -not you specifically, the community- are offered after YEARS of neglect with a possible hardware upgrade that is ~3x faster than your current solution. With potential -if business goes well- to move to a faster system like MPC8640D or some QorIQ CPU in the future. And you complain that the CPU chosen by your peers is not the PC killer that you would personally like. Think of that, just a bit.

@minator

Quote:

minator wrote: @Hans Quote: Thanks for that. Seeing as the demo was with a DIVX file, I doubt that is could play 1080 HD H.264 AVC. Debatable, but irrelevant. Why would you want to play full HD anyway? A Bluray player will do a much better job.

Not debatable at all. Decoding H.264 AVC requires more processing power than DIVX; this is a fact, and others have confirmed this too. I would like to be able to play videos that come off my HD camera in full HD, without having to buy a bluray player, and burn the videos to a bluray disk first.

Hans

_________________
http://hdrlab.org.nz/ - Amiga OS 4 projects, programming articles and more. Home of the RadeonHD driver for Amiga OS 4.x project.
https://keasigmadelta.com/ - More of my work.

@feanor

No, you are mistaken: I don´t want a pc killer. I just want a REGULAR computer. My approach when I make deliverings is simple: I want a computer to make things with it. USEFUL things, so I guess which hardware allows me to do those useful things, and I of course I ask for a computer that is up to some MINIMAL requirements. Let's forget even what Amiga was really about : A computer ahead of its time, that allowed people to render 3D images, to compose real music, draw with multiple colors, and enjoy all kind of graphical effects when the people (like me) was playing with an Spectrum. We are now just asking for a DECENT computer compared with the ones that everybody can buy for just 300 euros/dollars, for god's sake!

I know we must pay more, I know there aren´t so many choices, and I know it is very hard, slow, and in a very very small scale to produce this new hardware, but if some option is available and affordable. WHY NOT USE IT?

And I am not complaining because the most chosen (until now) is already faster and I would want even more, but because we have another (faster) choice, but some (not all) people take slower one, why on earth??? for 100 euro?? sorry but I find it SOO STUPID


Quote:

If you really want all that, what are you doing with AmigaOS?

Now? you really want to know?? I think it is the same like the mayority of people here:

Just waiting! Waiting... waiting, WAITING, WAITING !!!! already for years. Waiting and standing this stupid situation to end thanks to someone with clever ideas and a bit (not much is required) of money to understand what is Amiga about and why it still could make sense to relive it.

Back in the nineties I made a lot of things with my 1200, appart from playing some of the best games in my life, enjoy demos , learn all kind of graphical tricks and achievements or draw my digital art pieces, I designed my first 32 bit videogame which allowed me to become professional and start my career in the industry. But it's been more than 10 years since that "glory" (and the reference for most of you) days, we are entering year 2010, don´t forget it.

Now, sadly I have to stick to Windows Xp (and starting with 7) to do any serious task, and they are on another league not only because of the hardware and the wonderful software and tools they have, but because the OS is already about controlling the computer with voice, multitactile screens and real 3D with or without classes. WITH A COMPUTER THAT SIMPLY COULD PLAY HD VIDEO, WE WILL KEEP BEING AT LEAST 1 GENERATION BEHIND MOST OF THE PEOPLE, EVEN TECHNOLOGICAL NEWBIES, Can you get that??

And let's not talk even about notebooks, which make the experience a lot more enjoyable, but I don´t even dare to ask for mobile Amiga cause nobody seem to find a way to build them.

I think you should be the one thinking what is the better choice for your bussines, because if tomorrow Hyperion unveils its secret project and it can do what your system can´t or offfer exciting features that you wont, or simply Morphos can squeeze more power (or equal) from the mac mini than your board will deliver, you will end in the middle of (and paid for) a production that nobody will want. You have been warned.

Last edited by vision on 26-Sep-2009 at 11:05 PM.
Last edited by vision on 26-Sep-2009 at 10:56 PM.
Last edited by vision on 26-Sep-2009 at 10:49 PM.