| [News] Dave Haynie on Zorro... | ANN.lu |
|
| Dave Haynie on Zorro... : Comment 1 of 59 | ANN.lu |
| Posted by RoboShawn on 18-Dec-2001 10:18 GMT | | Nah. Dave is saying that Zorro 20 is best and all you competitors is bad innit. |
|
| Dave Haynie on Zorro... : Comment 2 of 59 | ANN.lu |
| Posted by AmiDelf on 18-Dec-2001 10:42 GMT | Its so pathetic to just go for todays standard. Why cant Merlancia do something? Zorro 5 would be great, it could go further and make a new standard, a Amiga standard, wich the PC could follow..
I dont want my Amiga or AmigaONE to be a PC.. It is already to much PC.. with these PCI expansion boards for Amiga 1200 and Amiga 4000.
-amidelf |
|
| Dave Haynie on Zorro... : Comment 3 of 59 | ANN.lu |
| Posted by RC on 18-Dec-2001 11:20 GMT | Are you people insane, or deaf? Zorro is a dead technology. If any of you even understand the electronics involved, why don't you bloody write a proper proposal outlining why Zorro is a better technology then PCI is, and why/how it should be modernised. I would be interested to see the results of such a masterpiece of literary arts...until then, please stick to your own profession, and don't tell engineers that we are all wrong. Words mean things. Explanations are in order.
RC. |
|
| Dave Haynie on Zorro... : Comment 4 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 11:48 GMT | In reply to Comment 3 (RC):
Don't waste your time, RC... some Amiga users are so f*cking stupid that they're not even capable to admit that the Amiga technology IS OUTDATED now. Period. |
|
| Dave Haynie on Zorro... : Comment 5 of 59 | ANN.lu |
| Posted by Peter Gordon on 18-Dec-2001 11:48 GMT | In reply to Comment 2 (AmiDelf):
You idiot! You're just saying you don't want PCI because its a crappy "PC standard" and you want Zorro 5 because it would be wikkid!
PCI isn't even a "PC standard" its actually an independant standard, and its a damn good one, too!
Plus, it is supported by thousands of companies around the world, ensuring a continuing supply of high-spec, low-cost cards. Even if you made some "kick-ass" Zorro 5 bus, who would support it? I'll tell you who: nobody.
Any cards made for it would be made in production runs barely into 3 digits. They would be prohibitively expensive. Standards are basically a good thing, provided that they are GOOD standards in the first place, and PCI is just that.
Stop living in a fantasy world! |
|
| Dave Haynie on Zorro... : Comment 6 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 12:01 GMT | In reply to Comment 3 (RC):
I agree. And when you speak in terms of 'good' or 'better' you should always include economy and rational business thinking as a heavyweight argument.
The PCI is the only widely accepted standard there is, and it is not bad at all! All new hardware (and components) is based on PCI technology, often integrated in a single chip, which requires little or none extra components to function. Motherboard manufacturers and expansion card manufacturers can easily and very rapidly develop new products based on PCI-based hardware components, simply by choosing the desired chip and mounting it on a circuit board. If you for some irrational reason should decide NOT to use PCI hardware, then you would have to design every single chip/feature from scratch, because all existing hardware is based on PCI technology. That is, you would have to design new disc controller chips, audio chips, IO chips, network chips and so on, because you have rejected to use existing components. So in the end of the day (or perhaps 'in the end of the decay', since such development efforts would take some time), you would end up with a computer (or a set of OEM chips) which do the same thing as their PCI counterpart, several years after their PCI counterparts hit the market, and to a price that is so high that no one could ever do anything useful with it (but don’t worry, no one would WANT to do something with it, since everyone is using PCI based hardware except you).
Not using PCI would be totally irrational and pointless. Trying to build a business on a custom designed 'standard' would be nothing but a successful suicide. |
|
| Dave Haynie on Zorro... : Comment 7 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 12:29 GMT | In reply to Comment 5 (Peter Gordon):
I wouldn't say PCI was particularly good, but balanced against
the alternatives and the obvious price benefits for peripherals,
it's the best choice for modern hardware. |
|
| Dave Haynie on Zorro... : Comment 8 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 12:37 GMT | In reply to Comment 7 (Anonymous):
But PCI *IS* good. PCI is the best there is! |
|
| Dave Haynie on Zorro... : Comment 9 of 59 | ANN.lu |
| Posted by AmiDelf on 18-Dec-2001 13:13 GMT | In reply to Comment 4 (Anonymous):
I admit that, but what I want to get thru here, is that Amiga should come with something standard as well, ... Why do we need to allways follow PC standards.. Some of them are right, others can be improoved..
-amidelf |
|
| Dave Haynie on Zorro... : Comment 10 of 59 | ANN.lu |
| Posted by Matthew Garrett on 18-Dec-2001 13:30 GMT | In reply to Comment 2 (AmiDelf):
PCI is not a PC standard. Apple use PCI in the Mac. IBM use PCI in the RS/6000. Sun use PCI their entire product range. HP use PCI in their PA-RISC Unix workstations. SGI use PCI in their entire product range. Compaq use PCI in their Alpha systems. The vast majority of these used their own expansion bus design at some point (IBM had MCA, Alpha had Turbo Channel, Sun had SBUS, Apply had NuBus) before deciding that PCI was better. PCI even won in the PC world because of its superiority over the alternatives. PCI is a standard throughout the computing world, and the reason for this is that it's good. A Zorro varient is 1) unlikely to be as good, and 2) guaranteed to be significantly more expensive. |
|
| Dave Haynie on Zorro... : Comment 11 of 59 | ANN.lu |
| Posted by AmiDelf on 18-Dec-2001 13:35 GMT | In reply to Comment 10 (Matthew Garrett):
What I want, is that Amiga comes with a better standard.. PCI and AGP are good.. I dont say theire bad.. I just say that an Zorro update would be a great way to go.
Also, what I find interesting, is Zorro's soundcards, like Repulse, DelfinaDSP and many more.. There are NO sound leaks at all..
Try to hear the difference between DelfinaDSP and Soundblaster PCI LIVE for example! SB might have more features and is better technical designed, but the sound is worse...
-amidelf |
|
| Dave Haynie on Zorro... : Comment 12 of 59 | ANN.lu |
| Posted by Martin Blom on 18-Dec-2001 13:48 GMT | In reply to Comment 3 (RC):
Nah. It's just that those of us who have a clue can't be bothered to waste time
on all these senseless discussions. |
|
| Dave Haynie on Zorro... : Comment 13 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 14:10 GMT | In reply to Comment 11 (AmiDelf):
OK, I'll bite.
Explain to me HOW you propose to make PCI and AGP better in your "Zorro V"?
Better in WHAT SPECIFIC ways?
What shortcomings of PCI et al will you seek to overcome?
Details, please, or else shut the fsck up. |
|
| Dave Haynie on Zorro... : Comment 14 of 59 | ANN.lu |
| Posted by EmGee on 18-Dec-2001 14:14 GMT | This is one of the reasons Amiga never went that big; proprietary hardware interfaces/busses. Even to use a normal printer (a Star LC-10 for my A500 back then, pprrrrrrrrrrrrrrrrrrrr :)) you had/have to remove the pin14 wire, else you could blow up your CIA chips.
Ok the Zorro slots where more advanced then the ISA on the PC in the 80's/begin 90's but as Dave Haynie stated PCI was/is good, and architecture independant. So why re-invent the wheel again, if you can use a good and CHEAP to manufacture bus/slots ???
If Amiga wants to be really modern on the hardware side they should join the Hypertransport or 3GIO groups, whatever the nextgen IO standard will be for the rest of the world. But PCI/AGP is good for starters, way to go Amiga/Eyetech!!
EmGee |
|
| Dave Haynie on Zorro... : Comment 15 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 14:52 GMT | In reply to Comment 11 (AmiDelf):
I don't know for certain, but I bet that the Delfina card is based on a PCI audio chip, with some additional PCI to Zorro bridging hardware. Hasn't that been the case for most of the amiga hardware for some time now?
And you know, sound quality has very little to do with the bus, at least as long as the bus can transfer data at a sufficient speed without interrupts.
And who ever said that soundblaster live is a professional soundcard? You know, you can find soundcards in all price ranges. Some are very cheap with soundquality according to its price. But actually, sundcards like that suits the needs of most of the computer users out there (especially since most of the PC speakers on the desktops around the world hardly are of hifi quality anyway). Some soundcards offers hifi or studio quality to a price that suits professionals. Some soundcards have a strong brand (read: soundblaster), which makes it possible for the manufacturer to increase the price without necessarily increasing the performance/quality. But soundblaster live IS a good soundcard! It's well up in the upper region in the consumer segment.
But this discussion is pointless anyway, since ALL soundcards (and OEM sound chips for that matter) have a PCI interface. The PCI interface is often integrated in the audio chip itself. |
|
| Dave Haynie on Zorro... : Comment 16 of 59 | ANN.lu |
| Posted by William F. Maddock on 18-Dec-2001 15:20 GMT | You know, I'm not an engineer, so take this for what it's worth. But I don't just see some "know-nothings" spouting nonsense. I also see "engineers" acting like arrogant assholes refusing to look forward---as though they know everything and nothing can be improved further so just shut up and be a good little sheep.
In case you've never noticed it, innovators are never sheep. They don't look at what's out there and say "oh, weell, that's the way it is." They look at what's out there and say, "How can I improve this mess?"
Anything designed and made by man is a mess, so it can always be improved. Even standards that are better than the rest are not perfect and can be improved. That goes for PCI as much as anything else.
Who's to say that the standard that eventually takes over from PCI---because PCI is not forever---won't be a group of ex-Amiga engineers with a better idea? But it'll never happen if you don't drop your arrogant attitudes, get off your duff, and start designing.
That's one reason I'm so glad that Dave's working for Ryan. Dave doesn't sit back and spout. Dave designs. Maybe the rest of you pretenders could follow suit. Quit whining about why something won't work and start designing something that does work.
Jesus is Lord. |
|
| Dave Haynie on Zorro... : Comment 17 of 59 | ANN.lu |
| Posted by Matthew Garrett on 18-Dec-2001 15:35 GMT | In reply to Comment 11 (AmiDelf):
The Delfina is based on a standard PC sound chip, the CS4231, which was designed to work on the ISA bus. Noise in the signal will be generated by a combination of poor-quality d/a conversion and interference from the machine internals. This is not determined by the bus, but instead is determined by the quality of the components and how well designed the card is. The only real use for the soundcard to use the bus is to obtain digital data representing the sound. For CD quality sound, this required little other than for the bus to be able to provide 150K of data per second in a fairly sustained fashion. Providing it can do that, the bus architecture that delivers this data is irrelevent as far as the quality of the resulting sound goes. |
|
| Dave Haynie on Zorro... : Comment 18 of 59 | ANN.lu |
| Posted by Bladerunner on 18-Dec-2001 15:41 GMT | In reply to Comment 16 (William F. Maddock):
But you won`t start development for a new Amiga beginning with the bus architrecture, will you?
It sounds like as you invent a new car, you will start with inventing the tires for it.
and as you claimed dave haynie.. he said in a few postings here, that he would use parts that are there
before he re invent it, when it fits for his work. And he said, that PCI is still fine....
And please Remember, a new bus architecture has to be supported for many plattforms, not only the Amiga. |
|
| Dave Haynie on Zorro... : Comment 19 of 59 | ANN.lu |
| Posted by Matthew Garrett on 18-Dec-2001 15:48 GMT | In reply to Comment 16 (William F. Maddock):
>Who's to say that the standard that eventually takes over from PCI---because PCI >is not forever---won't be a group of ex-Amiga engineers with a better idea? But >it'll never happen if you don't drop your arrogant attitudes, get off your duff, >and start designing.
Don't you think the time would be better spent designing something that offers some greater degree of advantage over the status quo? The primary requirement for an expansion slot bus to provide is the ability to transfer data over it. PCI does that. 32 bit, 33MHz PCI does that sufficiently quickly that the consumer market is satisfied. 64 bit PCI exists, as does 66MHz PCI. They'll be there once the consumer market needs more. You can plug a card into a PCI slot, close the machine, install the drivers and it'll work. What more do you want? The reason for not spending time on designing something "better" than PCI is that the improvement would be fairly insignificant and the peripherals would be more expensive. To make things better, most of them would just be Magical Mystery Bus to PCI bridges with a standard PCI ethernet chip/SCSI controller/sound chip/TV decoder stuck on. It's just not worth the effort when there are so many other aspects of computers that could be improved upon for significantly greater end result. |
|
| Dave Haynie on Zorro... : Comment 20 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 15:58 GMT | In reply to Comment 8 (Anonymous):
You obviously can't read. From a technical point of view, there are
better buses than PCI, but PCI wins because it's good enough and
has definite economic advatages. |
|
| Dave Haynie on Zorro... : Comment 21 of 59 | ANN.lu |
| Posted by William F. Maddock on 18-Dec-2001 17:19 GMT | In reply to Comment 19 (Matthew Garrett):
So basically, you're assuming that there will never be a need for greater throughput through the bus than what we now have available. Things will forever fit nicely into the architecture we already have and then come to a screaching halt, right? How does history bear up such an assumption?
In this thread there are already people asking for a new bus architecture that would blow PCI out of the water. They're only calling it ZORRO something in order to give it a name. The first group that survives designing a stable and usable bus that is far superior to PCI will probably be the one that benefits from beating PCI, so start designing.
There are other engineers in the world that are just as talented as anyone here, but who are not sitting back claiming that it can't be done, or it'll never catch on. They're designing, maybe in their garages or their basements, and in their spare time, but they're designing. Do you want them to get the money, or do you want to get the money yourself? Stop whining, start designing.
Jesus is Lord. |
|
| Dave Haynie on Zorro... : Comment 22 of 59 | ANN.lu |
| Posted by Adam Kowalczyk on 18-Dec-2001 17:43 GMT | In reply to Comment 16 (William F. Maddock):
William F. Maddock, I think you are mistaking realism as arrogance. It will take more than divine intervention for an ex-group from any company to create the next PC replacement, bus replacement, etc. This is not the 80's or early 90's when being different was possible from a hardware perspective. I do agree with you that something will supercede PCI and to be successful, the new Amigas should be prepared to adopt that technology as well.
Amiga is aiming for hardware independence with their AmigaDE and this is the way it should be to protect themselves from change that happens at an astounding rate. I don't believe that the next Amiga has to be superior from a hardware point of view, it should at least be equal. The new Amiga machines will be successful if Amiga adopts the strategy of a lean, efficient OS that isn't a resource hog like Windows. As much as the original Amiga hardware technology was leading edge in graphics and bus technology, it was the OS that allowed everything to work together so well. A good example of this on the PC side is Linux. Linux really makes PC hardware shine and easily outperforms Windows running on the same machine. I'm confident the Amiga OS would do the same thing when applied to standard hardware and even outshine Linux in some ways.
If Amiga and it's affiliated companies want to be successul, they need to keep a close eye on the pulse of the PC world. Where would Apple be if they had not adopted PCI? Don't forget good designers at Apple sometimes gets a little ahead of the market. Remember the G4 cube? A very elegant piece of hardware, yet a market failure. I think Steve Jobs has to give up on selling cubical machines (remember the NeXT machine?). I am an engineer, and good engineers don't re-invent the wheel just to be different. In today's market you need to put together systems that get the job done for the end customer in a timely cost-effective manner.
William, do not misunderstand me as I do believe everything can be improved. But there is reality to face as well as shareholders. Show me something real that I can touch and use and I will buy it. Start talking of new busses and other technology that doesn't exist yet and both my wallet and desktop will be empty. I really hope the bPlan and Eyetech projects succeed. I will undoubtedly purchase one. I'm also glad to see Dave Haynie still very much involved in the Amiga community.
Jesus may be the Lord, but which OS does he use? |
|
| Dave Haynie on Zorro... : Comment 23 of 59 | ANN.lu |
| Posted by Mike Veroukis on 18-Dec-2001 17:45 GMT | In reply to Comment 2 (AmiDelf):
>Its so pathetic to just go for todays standard. Why cant Merlancia do
>something? Zorro 5 would be great, it could go further and make a new
>standard, a Amiga standard, wich the PC could follow..
You'd be better of investing in time-machine technology so you can go back to 1985 and re-live the Amiga's golden age. You're obviously stuck in the past and oblivious to the modern computing world.
- Mike |
|
| Dave Haynie on Zorro... : Comment 24 of 59 | ANN.lu |
| Posted by Anonymous on 18-Dec-2001 17:51 GMT | In reply to Comment 11 (AmiDelf):
>What I want, is that Amiga comes with a better standard..
>
Hey, AmigaOne and Pegasos just came with a Slot1 CPU connector
to take us beyond the suxxy and slow ZIFs. :-) |
|
| Dave Haynie on Zorro... : Comment 25 of 59 | ANN.lu |
| Posted by Harry Kok on 18-Dec-2001 17:52 GMT | Well, from what i've picked-up about PCI, it is interrupt-driven.
AGP seems to have DMA but i've heard that it is buggy.
Someone pointed out that you need to think beyond what is
currently available, lets speculate on that.
Say, you want to do some audio-work with your computer,
and you want several inputs/outputs on that thing,
on seperate cards..
The thing with audio is, you can't allow much delay-time
between in and outs. 1mSec is barely acceptable.
You simply can't do that without proper dma.
So, it may be a good thing to design a dma bus wich,
layout wise, sits in line with the PCI bus.
So you can stick pci-cards in, but also special dma cards.
For this the main memory needs to be as close to the bus as possible.
On PCI, that is not the case I guess no?,
there's a northbridge between them or something like that.
So in order to make such dma possible, there will also be the need
for some new kind of chipmem (yes, there it is), which sits right
next to the bus. And has no 'Blackouts' longer that a few
microseconds. That chipmem could also be located on a card.
The bus beeing just wires with a busmaster/interface to the cpu.
(or perhaps a pci/dma bridge card wich does all that)
Just speculating..
"But this is all...Academic" |
|
| Dave Haynie on Zorro... : Comment 26 of 59 | ANN.lu |
| Posted by Matthew Garrett on 18-Dec-2001 17:52 GMT | In reply to Comment 21 (William F. Maddock):
>So basically, you're assuming that there will never be a need for greater
>throughput through the bus than what we now have available. Things will forever
>fit nicely into the architecture we already have and then come to a screaching
>halt, right? How does history bear up such an assumption?
No, I'm saying that the existing PCI implementation present in most machines is sufficient for consumer use and that already existing (and used at higher levels) extensions are extremely likely to be sufficient for several years. The past 6 years of PCI usage seem to bear that out rather well - it's been around since the late 486 era, PCs have gone from 60MHz to 2000MHz in the same time period, and it's still not generally a limiting factor. We aren't going to need something better for a long time.
>In this thread there are already people asking for a new bus architecture that
>would blow PCI out of the water. They're only calling it ZORRO something in
>order to give it a name. The first group that survives designing a stable and
>usable bus that is far superior to PCI will probably be the one that benefits
>from beating PCI, so start designing.
Blow PCI out of the water how? What features do you believe are going to be required within the next 5 to 10 years that aren't currently provided by PCI and can't be added to the PCI specification without too much trouble?
>There are other engineers in the world that are just as talented as anyone
>here, but who are not sitting back claiming that it can't be done, or it'll
>never catch on. They're designing, maybe in their garages or their basements,
>and in their spare time, but they're designing. Do you want them to get the
>money, or do you want to get the money yourself? Stop whining, start designing.
You're missing the point. If they're designing replacements for the PCI bus, they're wasting their time. The PCI bus is not the limiting factor. Memory bandwidth, on the other hand, is something that even consumers are hitting their heads off (which is why RAMBUS and DDR SDRAM get attention). Build a faster memory bus and the world will come to your door. Don't reinvent something that's not going to enhance anybody's computing experience. |
|
| Dave Haynie on Zorro... : Comment 27 of 59 | ANN.lu |
| Posted by Matthew Garrett on 18-Dec-2001 18:08 GMT | In reply to Comment 25 (Harry Kok):
>Well, from what i've picked-up about PCI, it is interrupt-driven.
>AGP seems to have DMA but i've heard that it is buggy.
PCI supports DMA. DMA modes on hard drives would be rather difficult to do otherwise :) |
|
| Dave Haynie on Zorro... : Comment 28 of 59 | ANN.lu |
| Posted by Mike Veroukis on 18-Dec-2001 18:09 GMT | In reply to Comment 16 (William F. Maddock):
>In case you've never noticed it, innovators are never sheep. They don't look
>at what's out there and say "oh, weell, that's the way it is." They look at
>what's out there and say, "How can I improve this mess?"
True, but it's worth noting that the market of today is not the market of 1984. When the Amiga was built the market was still up for grabs. There were no standards so everything had to be designed from scratch. Amiga was not innovative to be original, it was innovative because you just had to be those days. There were no standards for anything hence Paula, Agnus and Zorro. If nVidia, ATI and PCI were around at the time it's quite possible that the Amiga would not have custom hardware at all. If anything Amiga failed to see the market potential for their custom chips. If they really wanted to be innovative and bleeding edge they would have marketed their chips to other companies instead of keeping them to themselves. Amiga could have been the nVidia of today but they didn't see the market potential.
>Anything designed and made by man is a mess, so it can always be improved.
>Even standards that are better than the rest are not perfect and can be
>improved. That goes for PCI as much as anything else.
I doubt any engineer would argue that they could not improve on PCI/AGP. Also, no engineer would argue that putting a man on Mars isn't impossible, however, it's really hard and you need a hell of a budget. Aside from the fact that even if you hired some space aliens to build the most advanced bus supporting light-speed transfer rates, you'd still have to convince all the PCI users to convert... You'd probably lose out in the end. Engineering is just one obstacle you face.
>Who's to say that the standard that eventually takes over from PCI---because
>PCI is not forever---won't be a group of ex-Amiga engineers with a better
>idea? But it'll never happen if you don't drop your arrogant attitudes, get
>off your duff, and start designing.
Yes, PCI will one day be replaced, but for that to happen limitations will have to arise in the PCI standard first. Whatever those limitations might be will determine what the next generation expansion bus will be like. Creating a new bus "just because" isn't much to go on I'm afraid. So far I haven't read one single limitation or problem with PCI except that it was not designed by Dave Haynie. Unless you can come up with a real reason based on technical issues then forget it.
- Mike |
|
| Dave Haynie on Zorro... : Comment 29 of 59 | ANN.lu |
| Posted by Johan on 18-Dec-2001 18:26 GMT | In reply to Comment 25 (Harry Kok):
Ok. PCI does support DMA, of course it does.
A proper DMA design has a need to tell the CPU when the transfer is done though and that is what the IRQs are for.
Zorro does it in exactly the same way, you know. You set up the DMA transfer, start it and do something else until the IRQ arrives.
When it arrives, the DMA transfer is done.
AGP has a slightly specialized version of DMA, more geared towards the tasks of 3D graphics. AGP is by the
way just an 66MHz PCI card with this system. (Though AGP 2x uses DDR, and AGP4x uses 133 DDR IIRC).
Most of the complaints people have on PCI is usually just software problems. Using windows as a refrence for proper HW usage
is not the way to go, you know? |
|
| Dave Haynie on Zorro... : Comment 30 of 59 | ANN.lu |
| Posted by Brecht [darklite] on 18-Dec-2001 22:04 GMT | In reply to Comment 11 (AmiDelf):
>What I want, is that Amiga comes with a better standard.. PCI and AGP are
>good.. I dont say theire bad.. I just say that an Zorro update would be a
>great way to go.
You just want Zorro because it's *not* PC. But like I said before, if you want to change the bus, why not replace DIMM's, IDE & SCSI, Serial, Parallel, USB, Firewire with overexpensive crappy alternatives? Those are all >>PC<< standards too after all...
Get a
Also, what I find interesting, is Zorro's soundcards, like Repulse, DelfinaDSP and many more.. There are NO sound leaks at all..
Try to hear the difference between DelfinaDSP and Soundblaster PCI LIVE for example! SB might have more features and is better technical designed, but the sound is worse... |
|
| Dave Haynie on Zorro... : Comment 31 of 59 | ANN.lu |
| Posted by Brecht [darklite] on 18-Dec-2001 22:07 GMT | In reply to Comment 11 (AmiDelf):
(oops, sorry for previous post)
>What I want, is that Amiga comes with a better standard.. PCI and AGP are
>good.. I dont say theire bad.. I just say that an Zorro update would be a
>great way to go.
You just want Zorro because it's *not* PC. But like I said before, if you want to change the bus, why not replace DIMM's, IDE & SCSI, Serial, Parallel, USB, Firewire with overexpensive crappy alternatives? Those are all >>PC<< standards too after all...
Surte, someone could design Zorro5, that could be slightly better than PCI/AGP ass, but when PCI2 (or whateven standard will follow up PCI) is here, you're, again, nowhere.
>Also, what I find interesting, is Zorro's soundcards, like Repulse, DelfinaDSP
>and many more.. There are NO sound leaks at all..
Sound leaks?
>Try to hear the difference between DelfinaDSP and Soundblaster PCI LIVE for
>example! SB might have more features and is better technical designed, but the
>sound is worse...
Please, try harder. Maybe it also depends on the speakers used? Drivers? ... |
|
| Dave Haynie on Zorro... : Comment 32 of 59 | ANN.lu |
| Posted by Brecht [darklite] on 18-Dec-2001 22:17 GMT | In reply to Comment 16 (William F. Maddock):
>You know, I'm not an engineer, so take this for what it's worth. But I don't
>just see some "know-nothings" spouting nonsense. I also see "engineers" acting
>like arrogant assholes refusing to look forward---as though they know
>everything and nothing can be improved further so just shut up and be a good >little sheep.
Maybe those "assholes" are tired of explaining the same stuff over and over?
>In case you've never noticed it, innovators are never sheep. They don't look
>at what's out there and say "oh, weell, that's the way it is." They look at
>what's out there and say, "How can I improve this mess?"
>Anything designed and made by man is a mess, so it can always be improved.
>Even standards that are better than the rest are not perfect and can be
>improved. That goes for PCI as much as anything else.
Unlike when Jesus was still around, nowadays, you need money to do some decent R&D. Money ey? Isn't that one of the things Amiga developers do not have?
>Who's to say that the standard that eventually takes over from PCI---because
>PCI is not forever---won't be a group of ex-Amiga engineers with a better
>idea? But it'll never happen if you don't drop your arrogant attitudes, get
>off your duff, and start designing.
I'm sure they'll be glad to start designing, if you are willing to cover all costs.
>That's one reason I'm so glad that Dave's working for Ryan. Dave doesn't sit
>back and spout. Dave designs. Maybe the rest of you pretenders could follow
>suit. Quit whining about why something won't work and start designing
>something that does work.
Wasn't Dave one of the "assholes" that pointed out that trying to design a bus architecture to replace PCI is plain silly?
>Jesus is Lord.
Yeahyeah, and I got 2 cats at home, noone cares. |
|
| Dave Haynie on Zorro... : Comment 33 of 59 | ANN.lu |
| Posted by Joe "Floid" Kanowitz on 18-Dec-2001 22:52 GMT | In reply to Comment 22 (Adam Kowalczyk):
Adam Kowalczyk said:
>Remember the G4 cube? A very elegant piece of hardware, yet a market failure. I
>think Steve Jobs has to give up on selling cubical machines (remember the NeXT
>machine?)
Actually, Jobs just needs to stop selling cubical machines for 2-4x the price of an equivalent iMac... |
|
| Dave Haynie on Zorro... : Comment 34 of 59 | ANN.lu |
| Posted by Mike on 18-Dec-2001 23:21 GMT | Yeah, new standards all around...
I think we also need a better hard drive interface, no more SCSI III crap. We need something more obscure than a Dovorak(sp?) keyboard layout. Instead of CDs, I want all my data engraved in cheese. Running off of electricity is so 1990's, we should use lasers, some mirrors, and some smoke, maybe water... |
|
| Dave Haynie on Zorro... : Comment 35 of 59 | ANN.lu |
| Posted by Mike Veroukis on 18-Dec-2001 23:41 GMT | In reply to Comment 34 (Mike):
>I want all my data engraved in cheese.
Yeah, but it's gotta be real Greek Feta! I hate that cheap imitation Feta! Yuck!
- Mike |
|
| Dave Haynie on Zorro... : Comment 36 of 59 | ANN.lu |
| Posted by NeRP on 18-Dec-2001 23:45 GMT | Amiga has to go PCI. Designing a new bus now would take too long, and
the Amiga needs help now. Like someone posted, adding PCI to your
Amiga doesn't make it more "PCish". Everyone use it. Even my old EGS
Spectrum gfx card was made out of a PC chipset, and the card itself is
basically PC to Amiga bridge. The X-Surf ethernet card is a perfect
example of that. It is literally a PC ethernet card attached to a
zorro brige. I agree wholeheartedly that the Amiga needs to keep it's
individuality, but PCI under AmigaOS runs exactly the same as a zorro
slot to the end user. We need an Amiga running with a fast PCI based
graphics card and something like a Sonnet Cresendo 7200 G4 card in our
Amigas right away, then we can start walking instead of crawling
forward. |
|
| Dave Haynie on Zorro... : Comment 37 of 59 | ANN.lu |
| Posted by Alkis Tsapanidis on 19-Dec-2001 00:23 GMT | In reply to Comment 24 (Anonymous):
Those are made for adapters with one or 2 ZIFs you idiot....
The cpu itself is on a zif card... |
|
| Dave Haynie on Zorro... : Comment 38 of 59 | ANN.lu |
| Posted by Adam Kowalczyk on 19-Dec-2001 00:53 GMT | In reply to Comment 33 (Joe "Floid" Kanowitz):
Point well taken about the price of a G4 cube being twice the price of an iMac. Amiga distributors in the US can learn a lot about pricing from this example. It will be hard to buy a product someone if it's double the cost it should be even if it is in a fancy translucent or stainless steel case.
Adam |
|
| Dave Haynie on Zorro... : Comment 39 of 59 | ANN.lu |
| Posted by amigammc on 19-Dec-2001 02:07 GMT | In reply to Comment 16 (William F. Maddock):
>Who's to say that the standard that eventually takes over from PCI---because
>PCI is not forever---won't be a group of ex-Amiga engineers with a better
> idea?
I doubt it. The after-PCI standard has already been drawn, it's the Serial ATA, and it will be in all computers within 2 years, and it's almost 10 times faster than current PCI. |
|
| Dave Haynie on Zorro... : Comment 40 of 59 | ANN.lu |
| Posted by amigammc on 19-Dec-2001 02:20 GMT | In reply to Comment 21 (William F. Maddock):
>So basically, you're assuming that there will never be a need for greater
>throughput through the bus than what we now have available.
I don't think he's ever said that
>In this thread there are already people asking for a new bus architecture that
>would blow PCI out of the water. They're only calling it ZORRO something in >order to give it a name.
In this thread that are some (quite ignorant) people who would employ their time better by reading mainstream magazines rather than posting nonsense on ANN.
PC World, January 2002 issue, Page 69: Good Bye PCI, Hello Arrapahoe
"[...] the aging PCI and AGP bus standards won't be fast enough at shuttling data between your PC components. [...] codenamed Arrapahoe, that's up to ten times quicker than today's fast PCI-X bus. [...] When it's coming? Early 2004."
Ok, now we'll have people complaining that early 2004 is at least 13 months away. Those people, of course, are free to design a bus standard faster than Arrapahoe and position it in the market in less than 13 months. |
|
| Dave Haynie on Zorro... : Comment 41 of 59 | ANN.lu |
| Posted by Mattthew Garrett on 19-Dec-2001 03:30 GMT | In reply to Comment 39 (amigammc):
>I doubt it. The after-PCI standard has already been drawn, it's the Serial ATA,
>and it will be in all computers within 2 years, and it's almost 10 times faster
>than current PCI.
No. Serial ATA is the next-generation IDE interface. Arapahoe is the PCI replacement, but it's not going to displace PCI at the consumer level for several years yet. |
|
| Dave Haynie on Zorro... : Comment 42 of 59 | ANN.lu |
| Posted by Nathaniel Downes on 19-Dec-2001 05:51 GMT | In reply to Comment 25 (Harry Kok):
PCI just demands that there is an interrupt controller of some sort. I actually designed a PCI/AGP chipset for PowerPC that would have given an Amiga-like interrupt system. But, there's was no interest, so I've focused on my integrated chipset which can work parallell from the chipset.
PCI, as a bus connect goes, is decent, a bit slow, but not too slow. There are new standards appearing, but PCI will be with us for a good while longer yet. |
|
| Dave Haynie on Zorro... : Comment 43 of 59 | ANN.lu |
| Posted by Nathaniel Downes on 19-Dec-2001 05:55 GMT | In reply to Comment 32 (Brecht [darklite]):
I did not see Dave say that, actually. He said that if there is a need for it, to do it. But in this case, when the AMI bus was being drafted, PCI did the same job, so why fight the flow when you can use it instead? |
|
| Dave Haynie on Zorro... : Comment 44 of 59 | ANN.lu |
| Posted by Nathaniel Downes on 19-Dec-2001 05:57 GMT | In reply to Comment 39 (amigammc):
Uh....
I hate to break it to you, but Serial ATA is the next generation of HARD DRIVE INTERFACE STANDARDS. Using it to replace PCI would be like me trying to start my car by inserting a squirrel into the keyhole. |
|
| Dave Haynie on Zorro... : Comment 45 of 59 | ANN.lu |
| Posted by Nathaniel Downes on 19-Dec-2001 05:59 GMT | In reply to Comment 41 (Mattthew Garrett):
Aaraphone is *one* post-PCI standard. Another one is Hypertransport. And unlike Aaraphone, Hypertransport is here today. nVidia already includes a HT interface on it's nForce chipset. Other chipsets from VIA and SIS are coming with HT included. |
|
| Dave Haynie on Zorro... : Comment 46 of 59 | ANN.lu |
| Posted by megol on 19-Dec-2001 07:54 GMT | In reply to Comment 45 (Nathaniel Downes):
Hypertransport is designed for chip interconnection, not for communication with expansioncards. PCI-X is the most probable replacement for PCI as it is backwards compatible and gives a 2x speedboost compared with todays fastest PCI standard. |
|
| Dave Haynie on Zorro... : Comment 47 of 59 | ANN.lu |
| Posted by Nathaniel Downes on 19-Dec-2001 14:54 GMT | In reply to Comment 46 (megol):
PCI-X is a pretty poor expantion bus, it's limited to 1 or 2 slots, total. Basically it's you combine multiple PCI slots together into 1 physical port. |
|
| Dave Haynie on Zorro... : Comment 48 of 59 | ANN.lu |
| Posted by Graham on 19-Dec-2001 16:45 GMT | In reply to Comment 39 (amigammc):
You are thinking of 3GIO.
Serial ATA is the followup to the current parallel ATA standard for connecting IDE devices.
Instead of a 40-pin connector and 40/80 wire ribbon connecting up to 2 devices, it will have a 9-pin interface (IIRC) and cable to connect a single device. There will be 4 to 6 of these SerialATA connectors on the motherboard, and the controllers will be part of the chipset, or on PCI cards.
The first Serial ATA is expected to run at around 150MB/s per interface, growing as needed to around 1GB/s in several years time.
Graham |
|
| Dave Haynie on Zorro... : Comment 49 of 59 | ANN.lu |
| Posted by Graham on 19-Dec-2001 16:55 GMT | In reply to Comment 47 (Nathaniel Downes):
That is the problem with speeding up PCI - you lose devices because of timing problems.
PCI will be with us for many more years. 32/33 PCI is a mainstay and all motherboards will have such a bus. Some will have 64/66 PCI slots (usually two), and next year that may become PCI-X (64/133 IIRC). PCI-X allows you to have 1 slot at 133MHz (1GB/s), 2 slots at 100MHz (800MB/s) or 4 slots at 66MHz (533MB/s). I would expect the latter to become popular when the technology filters down from high end servers.
Hypertransport has no physical connector defined. API are defining one though, but personally I see it as an (excellent) chip to chip interconnect technology - low pin count, and damn fast. And it works.
3GIO/Arapapapapaahoe is the follow-up to PCI. There is a good chance that it will build upon HyperTransport anyway, considering that HT is doing so well. The idea will be to reduce the size of the physical connector primarily, as well as speed up the bus. Expect a connector half the size of a standard PCI connector, but running at 8x - 32x the speed... in 2004 though :( |
|
| Dave Haynie on Zorro... : Comment 50 of 59 | ANN.lu |
| Posted by RLFrost on 19-Dec-2001 17:50 GMT | OK Guys,
Here are the advantages of Zorro in a nutshell. It's hard to understand why these advantages aren't obvious, but here we are.
1. Zorro is a heck of a dashing guy. His dress is stylish and without peer in
the industry. Black is the perfect color for a bus or an outlaw...who can
doubt it?
2. Zorro has a mask. Sometimes his mask is 0x0000742bh, and other times it is
of another material, but it is always there to hide his identity from those
who would do him harm.
3. Zorro has a heck of a horse. I mean this guy is strong, fsst, and smart. The
ride is smooth, and the backtalk (sound leaks) are minimal.
4. Zorro's sidekick is very efficient. His accent may leave something to be
desired, but he is always on top of the situation and the information that
he provides his boss is accurate and true.
5. Zorro always kicks the bad guy's a*^!
6. Zorro's not cheap.
Hope that helps,
RLFrost |
|
| Anonymous, there are 59 items in your selection [1 - 50] [51 - 59] |
|
- User Menu
-
- About ANN archives
- The ANN archives is powered by #AmigaZeux. It was updated daily (news last: 22-Oct-2004; comments last: 18-May-2005).
ANN.lu was created, previously owned and maintained by Christian Kemp, www.ckemp.com.
- Contribute
- Not possible at this time!
- Search ANN archives
- Advanced search
- Hosting
- ANN.lu was hosted by Dreamhost. Sign up through this link, mention "ckemp" as referrer and he will get a 10% commission on any account you purchase.
Please show your appreciation for any past, present and future work on ANN.lu by making a contribution via PayPal.
|
