Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: Pyromania on September 08, 2015, 11:25:11 PM
-
http://www.atarimusic.net/index.php?option=com_content&view=article&id=263:the-falcon030-vs-the-amiga1200&catid=60:atari-hardware&Itemid=214
It's on an Atari site so not sure what I was expecting. Conclusion was that the machines were pretty much equal which is false. The A1200 was much better of course except for the DSP on the Atari which had limited 3rd party software support.
-
I thought I heard that the falcon was a more powerful machine than the a1200?
I could have heard wrong.
-
Seems like a fair and balanced article to me and a solid read.
-
My goodness you Amigans are dumb sometimes.
Im an Amigan And i know the Falcon was WAY more powerful than a the A1200
-
http://www.atarimusic.net/index.php?option=com_content&view=article&id=263:the-falcon030-vs-the-amiga1200&catid=60:atari-hardware&Itemid=214
It's on an Atari site so not sure what I was expecting. Conclusion was that the machines were pretty much equal which is false. The A1200 was much better of course except for the DSP on the Atari which had limited 3rd party software support.
Not bad comparison. I'd love that A1200 had a DSP like Falcon has (or that coders took advantage of Delphina's 56002 DSP as much as Atari coders do with their 56001).
Here you have some videos showing the work of a brilliant programmer running on a standard 16MB Falcon in action:
https://www.youtube.com/watch?v=WpwlZgQPCpk
https://www.youtube.com/watch?v=jjgWx3DE1CY
https://www.youtube.com/watch?v=RTxwfRl_I0U
-
An extremely fair and balanced article. The Flacon was horribly crippled by its memory bus, but other than that it has the chunky gfx mode, the 16bit audio, a DSP and a 68030 which were all features missing from the Amiga 1200.
The A1200 was good in spite of its design, not because of it.
-
Not sure why he keeps making reference to pc and mac sound cards too, saying that by the 90's the Amiga's sound was lagging behind. Maybe he doesn't know that 16-bit sound cards have always been around for the A1500/2000/3000/4000 machines. Even the wedge-cased Amiga's got 16-bit sound cards...these connected via serial/parallel, pcmcia, or clockport.
With the standard 68EC020 cpu onboard the A1200, the bus is 24 not 32-bit. Adding a cpu upgrade with fastram gives 32-bit. Unless he was referring to AGA? If he was then the falcon bus width should have been listed as 16 not 32.
Check this (very long) thread out if you fancy some Atari vs Amiga discussion/arguments/flame-wars. They talk about the Falcon vs. A1200 too later on in the thread IIRC:
http://atariage.com/forums/topic/29957-atari-st-vs-amiga/page-1?hl=fastrobplus amiga 1000
Enjoy :)
-
Check this (very long) thread out if you fancy some Atari vs Amiga discussion/arguments/flame-wars. They talk about the Falcon vs. A1200 too later on in the thread IIRC:
http://atariage.com/forums/topic/29957-atari-st-vs-amiga/page-1?hl=fastrobplus amiga 1000
Enjoy :)
I read a sampling from about 10 pages.... seems most liked the Amiga! You can watch the Amiga grow with great add-on's over the years in that thread while the Atari has pretty much stayed stagnant! But CRAP! That thread went on for 9 years! :rtfm: Thanks for the link Paul!
-
Amiga rules them all.
:)
-
Wasn't the Falcons 030 crippled by 16 bit bottle neck connections or something, making it slower?
-
Amiga rules them all.
:)
Yeah! We ROCK! :laugh1:
Hey Atari.... make like a fuse and BLOW! :lol:
(http://www.racketboy.com/images/amiga-defender-comparison-434x400.jpg)
Defender of the Crown (Top to Bottom: Amiga, Atari ST, DOS (CGA) )
--------------------------------------------------------------------------------
-
Yeah! We ROCK! :laugh1:
Hey Atari.... make like a fuse and BLOW! :lol:
Defender of the Crown (Top to Bottom: Amiga, Atari ST, DOS (CGA) )
--------------------------------------------------------------------------------
Even the C64 graphics looked better than the DOS (CGA) version
-
A stock A1200 is a crippled machine.
You could easily spend three times its price on expansions alone. Some say that's good design because it allows expansions to grow your system as you needs change. I say it was nothing but a cash grab.
You'd think the CPU is a big let down. But actually no, that wasn't the biggest brain fart decision.
Seriously: shipping with 2 MB chip RAM? Not upgradable? C'mon. Even 2 MB of fast RAM revamps the performance and possibilities.
Then adding even a 20 MB hard drive improves it further.
Just those two changes as standard improves the software and OS possibilities.
And sound: the same chip for 10 years?
Please. Of course a stock ATARI was better than a stick A1200.
-
On paper, and for specific audio workloads, the Falcon was better. But for all intents and purposes, in real world applications across a broad spectrum of workloads, the A1200 was better.
The article does not take into account the TCO of each system. Bang for buck the A1200 was a much better investment. The market dynamics proved it to be so.
-
Even the C64 graphics looked better than the DOS (CGA) version
Yeah, Much Better!!.
-
On paper, and for specific audio workloads, the Falcon was better. But for all intents and purposes, in real world applications across a broad spectrum of workloads, the A1200 was better.
The article does not take into account the TCO of each system. Bang for buck the A1200 was a much better investment. The market dynamics proved it to be so.
As a floppy-only 2 Mb system? What real world apps and broad spectrum of workloads would be possible?
-
The Falcon is more powerful, of course. Only a raging fanboi would say otherwise. And no, I'm not an Atari fan (the horror!).
-
With the standard 68EC020 cpu onboard the A1200, the bus is 24 not 32-bit. Adding a cpu upgrade with fastram gives 32-bit. Unless he was referring to AGA? If he was then the falcon bus width should have been listed as 16 not 32.
Nope, the standard A1200 bad a full 32bit bus, although one that was shared with the AGA chipset (if you didn't have a FastRam card). The 24-bit thing is just the size of the address space, this is what limited the size of A1200 memory expansion (the ones that lacked a CPU) cards to just 8MB.
Adding a full 68020 or '030 would give you the full 32bit address space and let you install more than 8MB of RAM.
-
An extremely fair and balanced article. The Flacon was horribly crippled by its memory bus, but other than that it has the chunky gfx mode, the 16bit audio, a DSP and a 68030 which were all features missing from the Amiga 1200.
The only thing Falcon feature that was truly missing from the A1200 was the chunky video modes.
The 68030 on a 16 bit data bus was a terrible idea. It doesn't appeal to the low end or the high end customer, although it is better than an Atari ST.
Both computers were less than perfect because of the companies collapsing around them. Atari shipped Tos and a beta of MultiTos with the Falcon as proof. The A1000 got away with shipping a beta OS because the computer had more potential than others, but in 1992 neither the Amiga or Falcon had any long term potential.
I've not seen any development nightmares from Atari, but that is expected because Jacks sons Sam, Leonard and Gary were integral to the business. While commodores primary business was infighting and politics, they just happened to make computers as a side line. We at least know why AGA was cobbled together in haste and (the AA3000 at least) didn't end up with DSP and 16 bit sound.
p.s. The falcon case looks truly horrible.
-
The Falcon is more powerful, period. Don't be a bunch of fanbois, guys.
-
For Amiga long ago should build 030/040/060/PPC processor card with DSP and 16-bit audio codec with mixer Paula and Line output and input and microphone/headphones/CD-input/Digital etc.
Cards on the clock port / Zorro is nothing special, inefficient.
-
I always read that a 030 on a 16bit bus is not a good idea.
This might be true, although I did my tests on an A500+ with an ACA500 and my blizz030.
AIBB gave me the exact same results as on my A1200 except some results like Line test because of AGA.
Propably a 50MHz 030 is not fast enough to bottleneck a 16bit bus?
-
I always read that a 030 on a 16bit bus is not a good idea.
This might be true, although I did my tests on an A500+ with an ACA500 and my blizz030.
AIBB gave me the exact same results as on my A1200 except some results like Line test because of AGA.
It's not the same thing, in both cases the 68030 in your blizzard 030 is connected to a 32 bit bus. Only chip ram access will be 16 bit and most of the tests won't be touching it at all. I assume you left the 32 bit memory attached to the blizzard 030, remove that and you'll see the speed difference.
The 68030 in a falcon has no way of accessing any memory at 32 bit. In the Atari world ST-RAM is equivalent to 16 bit chip ram and TT-RAM is 32 bit fast ram, the falcon (without an accelerator) only supports ST-RAM.
So to get good performance you have to pay just as much for as an accelerator as you do on the A1200. Which makes the choice of 68030 rather strange. It's obviously a 68000 design with a 68030 shoe horned in as they only connect 24 address lines. The A1200 on the other hand can get 32 bit fast ram with just a cheap trapdoor ram upgrade. Although the A1200 is limited to 24 address lines unless you buy an accelerator.
-
I always read that a 030 on a 16bit bus is not a good idea.
This might be true, although I did my tests on an A500+ with an ACA500 and my blizz030.
AIBB gave me the exact same results as on my A1200 except some results like Line test because of AGA.
Probably a 50MHz 030 is not fast enough to bottleneck a 16bit bus?
Just try running a VXL*30 without a ram32. Or any other a500 or a2000 68030 accelerator without 32 bit ram.
The integer math and even the floating point go way up but all the custom chip transfers, the graphics and all the ram transfers stay slow.
You only get a 10-20% increase in speed from an 68030 without 32bit ram over a stock 7 mhz 68000. The upside is you can run more things (requiring 68020+).
The Graphics in a500 or a2000 even with an accelerator stay slower than the A3000s graphics. Check out AIBB tests over at amiga hardware database.
-
... although I did my tests on an A500+ with an ACA500 and my blizz030. ....
That's a nice configuration!!!!! :D
-
The Falcon is more powerful, of course. Only a raging fanboi would say otherwise. And no, I'm not an Atari fan (the horror!).
Stock yes on paper. But their OS is just a load of Tosh. What's all this I read about:
While the Falcon has its operating system in ROM, the multi-tasking version ships on floppy. There is a good reason for this (Other than it wasn't ready when the Falcon first shipped!), many professional applications would not work well with this new system, especially timing critical programs like Cubase and Logic. For these to run properly, the more common ROM based operating system was used while power users who needed the new functions could install the new operating systems on their hard drives. Even with this installed, there was always the option to boot into the single tasking OS with the use of a control panel applet, which could also configure other options on the machine, like the advanced sound system and the new high speed serial port.
Sounds like a compatibility nightmare to me.
-
An extremely fair and balanced article. The Flacon was horribly crippled by its memory bus,
Don't forget that the Falcon was also horribly crippled by it's SW (read TOS) which allready was in legacy hell as it wasn't really planned for multitasking or HW expansion.
-
Stock yes on paper.
Not on paper, it really is more powerful:
https://www.youtube.com/watch?v=WpwlZgQPCpk
https://www.youtube.com/watch?v=jjgWx3DE1CY
https://www.youtube.com/watch?v=RTxwfRl_I0U
No Amiga can do that with a 16 mhz 68030.
But their OS is just a load of Tosh.
That's software, not hardware. Doesn't count.
The A1200 is still the cooler machine, of course, but it's not more powerful.
Guys, don't be fanbois, please, it makes us look bad.
-
@Thorham
Lack of chunky mode meant that the A1200 sucked donkey-balls in 3D.
But to your suprise there actually were other uses for computers in 1992 except FPS games !!
Plenty other fields were a stock Falcon was utterly useless compared to a stock A1200.
-
Lack of chunky mode meant that the A1200 sucked donkey-balls in 3D.
Lack of chunky modes is the least of your concerns when dealing with real 3D.
Plenty other fields were a stock Falcon was utterly useless compared to a stock A1200.
Didn't I say that the A1200 is the cooler machine of the two?
-
For Amiga long ago should build 030/040/060/PPC processor card with DSP and 16-bit audio codec with mixer Paula and Line output and input and microphone/headphones/CD-input/Digital etc.
Cards on the clock port / Zorro is nothing special, inefficient.
Falcon DSP bus was quite slow too, slower than ZorroII. I guess that more DSP software could be ported to Delfina.
Falcon had nice 16bit screenmodes although it lacked sprites and its blitter was more primitive than Amiga's one.
Anyway skilled programmers use DSP to accelerate drawing lots of software sprites:
https://www.youtube.com/watch?v=87EMhNkTkeo
they also have a nice NeoGeo emu, btw.
-
Anyway skilled programmers use DSP to accelerate drawing lots of software sprites:
https://www.youtube.com/watch?v=87EMhNkTkeo
Isn't this great?
Sometimes I ask myself why didn't we have games like this on amiga taking advantages of 040/060 :-/
-
Lack of chunky modes is the least of your concerns when dealing with real 3D.
In terms of rendering it's pretty important. Texture mapping hardware and geometry/lighting maths in hardware are what you'd want next, but it would have made a huge difference.
Commodore should easily have been able to produce something with chunky pixels and non perspective corrected texture mapping in hardware (ie primitive PS1) by 1991. But they had to rush to get AGA out the door because they had been chasing AAA for so long. It would have still sold a boat load if it only had 8 bit colour and 8 bit sound, but they would have had to release another machine in 1994 to leapfrog the PS1 though.
Isn't this great?
Sometimes I ask myself why didn't we have games like this on amiga taking advantages of 040/060 :-/
People left on the Amiga seem to be intent on creating web browsers and other productivity apps, rather than writing games. If AGA hadn't been a rush to get something out of the door then it would have had chunky video and the blitter would be more than just the OCS blitter copy and pasted, and you'd have games like that on stock machines.
https://www.youtube.com/watch?v=WpwlZgQPCpk (https://www.youtube.com/watch?v=WpwlZgQPCpk)
https://www.youtube.com/watch?v=jjgWx3DE1CY (https://www.youtube.com/watch?v=jjgWx3DE1CY)
https://www.youtube.com/watch?v=RTxwfRl_I0U (https://www.youtube.com/watch?v=RTxwfRl_I0U)
No Amiga can do that with a 16 mhz 68030.
I think I can live without Quake 2 in low res at 6fps. The DSP helps with all that geometry sure, but the rest of the hardware can't keep up. I'm not sure why anyone would want a 16mhz 68030 with no fast ram on an A1200, apart from an extra 2mhz and the extra cache it would barely be any different.
There seems to be some competition in the thread http://www.atari-forum.com/viewtopic.php?f=68&t=26775&start=950 I'm not quite sure why he's having so much problems with data caches on the Amiga though.
-
Don't forget that the Falcon was also horribly crippled by it's SW (read TOS) which allready was in legacy hell as it wasn't really planned for multitasking or HW expansion.
I was going to add the fact that the OS was atrocious was also a downside... But Mint was quite good, even though nothing ran on it... Yeah I should have said crappy 16bit data bus and crappy OS :)
-
People left on the Amiga seem to be intent on creating web browsers and other productivity apps, rather than writing games.
Ahh, productivity apps. 20 years late, but better late than never. ;)
My only other contribution to this thread is "Stupid Commodore, should've sold the A1200 with an '030 and some fast ram, or at least a higher clocked '020 and some fast ram. Would it have killed them to just add a little stinkin' fast ram to the base system?" *sigh* Hindsight, I think this thread topic comes up every year or two! ;)
-
Can't you guys just admit that the Falcon is more powerful and move on?
-
The Amiga is not just another computer. They're alive, they have a soul. They're full of charisma from the moment electricity flows through their custom chips.
-
The Amiga is not just another computer. They're alive, they have a soul. They're full of charisma from the moment electricity flows through their custom chips.
(https://brainsyndicate.files.wordpress.com/2010/10/thumbs.jpg)
-
The Amiga is not just another computer. They're alive, they have a soul. They're full of charisma from the moment electricity flows through their custom chips.
I'll swallow their soul.
-
The Amiga is not just another computer. They're alive, they have a soul. They're full of charisma from the moment electricity flows through their custom chips.
Spoken like a true poet. :)
-
Actually it is a pretty well written, non biased article. Out of the box the Falcon was the better machine for a variety of tasks, although the A1200 did beat it in some things the Falcon was the better machine. Also keep in mind at the end of the article when he talks about when adding more RAM to the Amiga it then overtakes the Falcon so....if both machines were upgraded then the A1200 would actually be the better of the two. Nothing wrong with the article and the Falcon was a great machine also!
-
IMO, the technical and Amiga understanding of the article author could have been better but it is not deliberately biased. Both the Amiga 1200 and Atari Falcon had bottlenecks in stock form but the Falcon had more processing performance while not being the easiest to take advantage of. The DSP is very difficult to use from a compiler but can be useful for offloading the CPU for some tasks, especially related to audio. The Amiga 1200 is really meant to be expanded. I'm surprised C= didn't remove the CPU and some logic from the motherboard and ship them all with a CPU card (of choice) to save a few dollars and add higher margin upscale models. The low clocked 68EC020 is really slow but was no doubt very cheap for C= at the time. No CPU data cache and slow (chip) memory is a very bad combination and one of the reasons why a simple fast memory expansion provides a huge performance gain.
-
I liked the article too and remember the Games Master magazine comparison from the time. The A1200 was still the machine to get because of the Amiga software support. In terms of hardware the A1200 was a disappointment and you could argue that the Falcon was better hardware out of the box. The A1200 should have had a high density disk drive (if not a small hard drive), an upgraded Paula and fast ram as standard in 1992. By 1993 a 030 version should have been available. By 1994 a CD-Rom bundled 030 machine with AAA or chunky capable AGA/Akiko to support the CD32 software market and keep Commodore afloat.
-
IMO, the technical and Amiga understanding of the article author could have been better but it is not deliberately biased. Both the Amiga 1200 and Atari Falcon had bottlenecks in stock form but the Falcon had more processing performance while not being the easiest to take advantage of. The DSP is very difficult to use from a compiler but can be useful for offloading the CPU for some tasks, especially related to audio. The Amiga 1200 is really meant to be expanded. I'm surprised C= didn't remove the CPU and some logic from the motherboard and ship them all with a CPU card (of choice) to save a few dollars and add higher margin upscale models. The low clocked 68EC020 is really slow but was no doubt very cheap for C= at the time. No CPU data cache and slow (chip) memory is a very bad combination and one of the reasons why a simple fast memory expansion provides a huge performance gain.
as usual, within amiga subject, i cant really find a scratch in your reasoning..
-
I've been fostering a hypothesis for a little while now that Commodore and Atari are actually one large bipolar schizophrenic company, Commotari
And when Amiga came along it made matters even worse; Both "Jekyll" and "Hyde" were in love with her.
-
as usual, within amiga subject, i cant really find a scratch in your reasoning..
C= produced the Amiga 4000CR (Cost Reduced) motherboard without a CPU to save a few dollars so why not the Amiga 1200? Maybe the majority of 1200 owners were using the CPU on the 1200 where few 4000 owners were? Maybe C= had an oversupply of 68EC020 processors they bought cheap and wanted to get rid of? My logic is good for not having much inside information ;).
I've been fostering a hypothesis for a little while now that Commodore and Atari are actually one large bipolar schizophrenic company, Commotari
And when Amiga came along it made matters even worse; Both "Jekyll" and "Hyde" were in love with her.
Considering how many Amiga guys were originally Atari guys, it makes me wonder what would have happened if Atari had been able to create a better work environment and keep their talented employees. Amiga would have been an Atari product and would have benefited from a better dealer sales network and no competition from the Atari ST/Falcon. There probably would have been better games support but maybe less "computer" support and expansion (depending on how much freedom Jay Minor would have been given). The Atari ST/Falcon did get good computer support and expansion but how much of it was to compete with the Amiga? The other what if is if the original Amiga developers had been allowed to develop the Amiga as they wanted. C= failed to keep the original Amiga employee talent as well.
-
...depending on how much freedom Jay Minor would have been given.
Tale as old as time:
Hey you there, where did you learn to carve like that?
Here and there my Pharaoh.
You should totally come work on my pyramid. Onemhotep, bring this artisan with us.
Later...
Artisan! You're chiselling the story wrong! I want you to tell it like I did the whole thing by myself.
Artisan! You're using too much blue! Lapis doesn't grow on reeds you know.
Later...
Onemhotep! Have you seen the artisan lately?
-
Not having HD diskdrive would'nt have been a problem if the A1200 came with a harddrive as standard, with all versions.
Chipset should have been improved too, with Chunky & VGA with integrated scandoubler.
-
As the A4000 was high end. You could have made the A1200 in a larger case it would be cheaper to upgrade.
2.5inch drives were expensive back then.
-
The A1200 should have had a high density disk drive (if not a small hard drive)
The A1200 was sold both without and with a hard drive preinstalled, the latter was called A1200HD (with a 20MB drive) and A1200HD/40 (with a 40MB one): http://www.bigbookofamigahardware.com/bboah/product.aspx?id=15
Varthall
-
It matters little that the Falcon was slightly quicker, had better sound capabilities and a DSP.
When the A1200 cost £399 and the Falcon cost a whopping £999 when they launched in the UK. Even if you add in a Blizzard 1230/50 + some fast RAM, and a small HDD, the A1200 would still have come out cheaper by a few quid.
So IMO, in terms of cost versus performance the A1200 is the better machine.
-
Tale as old as time:
Hey you there, where did you learn to carve like that?
Here and there my Pharaoh.
You should totally come work on my pyramid. Onemhotep, bring this artisan with us.
Later...
Artisan! You're chiselling the story wrong! I want you to tell it like I did the whole thing by myself.
Artisan! You're using too much blue! Lapis doesn't grow on reeds you know.
Later...
Onemhotep! Have you seen the artisan lately?
Jay Minor was given freedom when creating the Amiga "game machine" and had the foresight to make it an expandable "computer". He would likely not have had as much freedom if he was working for Atari or C= where the Pharoh would likely have ruined his creativity. Creativity is only possible with freedom.
It matters little that the Falcon was slightly quicker, had better sound capabilities and a DSP.
When the A1200 cost £399 and the Falcon cost a whopping £999 when they launched in the UK. Even if you add in a Blizzard 1230/50 + some fast RAM, and a small HDD, the A1200 would still have come out cheaper by a few quid.
So IMO, in terms of cost versus performance the A1200 is the better machine.
Good point. It's really cost/performance that matters.
-
It matters little that the Falcon was slightly quicker, had better sound capabilities and a DSP.
When the A1200 cost £399 and the Falcon cost a whopping £999 when they launched in the UK. Even if you add in a Blizzard 1230/50 + some fast RAM, and a small HDD, the A1200 would still have come out cheaper by a few quid.
So IMO, in terms of cost versus performance the A1200 is the better machine.
£999? Wow, I had no idea! That's a lot of dough! Would have been a Microbotics 68030 accelerator though as the Blizzard wasn't out in 1992/3. Correct me if I'm wrong though. MX1230 or something was it?
-
C= produced the Amiga 4000CR (Cost Reduced) motherboard without a CPU to save a few dollars so why not the Amiga 1200?]
It was the other way round. https://en.wikipedia.org/wiki/Amiga_4000#A4000-CR_version
The first Amiga 4000s were shipped without the cpu present on the motherboard but with a card in the cpu slot. They then started shipping them with the cpu on the motherboard and without the card, to save money on the cpu card.
It's much cheaper to put the cpu on the motherboard than ship it on a card, So if they had done that with the A1200 you would have to pay more for the same speed cpu.
£999? Wow, I had no idea! That's a lot of dough!
I'm pretty sure it was closer to £599.
-
£599 probably - the US prices were:
"The Falcon030 is available in four variations -- 1 meg of RAM and no hard
drive, 4 meg of RAM and no hard drive, 4 meg of RAM and an 80 meg hard
drive, and 14 meg of RAM and an 80 meg hard drive. Retail prices range
from $799 to $1899. Actual street price is generally $100-200 less."
-
It was the other way round. https://en.wikipedia.org/wiki/Amiga_4000#A4000-CR_version
The first Amiga 4000s were shipped without the cpu present on the motherboard but with a card in the cpu slot. They then started shipping them with the cpu on the motherboard and without the card, to save money on the cpu card.
Interesting, so the 4000CR with CPU would be cheaper for buyers wanting a 68EC030 but I would have thought the percentage wanting a higher end CPU would be growing and decreasing the advantage of adding an unused CPU on the motherboard. Many more people probably used the low end CPU in the 1200 than in the 4000 which probably explains why it is present. C= must have purchased these low clocked EC processors for dirt cheap. Too bad they didn't put a 68020/68030 socket with replaceable oscillator on both the 1200 and 4000 motherboard but that would have cost a few cents more.
-
Interesting, so the 4000CR with CPU would be cheaper for buyers wanting a 68EC030 but I would have thought the percentage wanting a higher end CPU would be growing and decreasing the advantage of adding an unused CPU on the motherboard.
If they wanted a higher end CPU they would buy the A4000/40.
You could argue there would be people who wanted to buy an A4000 without a cpu at all and add a third party one, but that would likely cause issues with people buying one and finding out it wasn't actually a stand alone computer. You'd also likely add more cost by having multiple SKU's. If you are buying an 060 board you can afford the low cost of the EC030 on the board.
It would likely have been cheaper for commodore to only ship the EC030 versions and sell anything faster as dealer upgrades.
Atari banned retailers selling Falcons via mail, unless the customer asked for written permission from Atari that they then passed on to the dealer. That would have severely hampered the uptake, which would have reduced the amount of software being created. It was hard enough for commodore to get developers to make AGA specific games.
-
I'm at Scotch Corner motorway services at the moment. All I wanted was a sandwich and a bag of crisps... So I got to the counter and the lady said 'It's cheaper if you buy a drink with it because there's a Meal Deal.'
So I had to buy a drink too despite not wanting one. Just doesn't make sense to me (besides ripping people off).
-
An extremely fair and balanced article. The Flacon was horribly crippled by its memory bus, but other than that it has the chunky gfx mode, the 16bit audio, a DSP and a 68030 which were all features missing from the Amiga 1200.
The A1200 was good in spite of its design, not because of it.
Yes, an interesting article. indeed.
If Atari had made the falcon with a full 32bit data bus. the story would be totally different...
-
Not sure why he keeps making reference to pc and mac sound cards too, saying that by the 90's the Amiga's sound was lagging behind. Maybe he doesn't know that 16-bit sound cards have always been around for the A1500/2000/3000/4000 machines. Even the wedge-cased Amiga's got 16-bit sound cards...these connected via serial/parallel, pcmcia, or clockport.
With the standard 68EC020 cpu onboard the A1200, the bus is 24 not 32-bit. Adding a cpu upgrade with fastram gives 32-bit. Unless he was referring to AGA? If he was then the falcon bus width should have been listed as 16 not 32.
Check this (very long) thread out if you fancy some Atari vs Amiga discussion/arguments/flame-wars. They talk about the Falcon vs. A1200 too later on in the thread IIRC:
http://atariage.com/forums/topic/29957-atari-st-vs-amiga/page-1?hl=fastrobplusamiga1000
Enjoy :)
Question:
Looking at the surface of the A1200 motherboard. you see prints next to the 020 cpu. looking like the oulines of a 6881/2 fpu. if one found a full blown MC 020 and a fpu... I wonder if it would be possible to replace the cpu and add an fpu directly to the motherboard. and would it be possible to replace the crystal with, lets say 16mhz....
I have wondered about his for many years....
-
Question:
Looking at the surface of the A1200 motherboard. you see prints next to the 020 cpu. looking like the oulines of a 6881/2 fpu. if one found a full blown MC 020 and a fpu... I wonder if it would be possible to replace the cpu and add an fpu directly to the motherboard. and would it be possible to replace the crystal with, lets say 16mhz....
I have wondered about his for many years....
Yes Sir! :)
http://eab.abime.net/showthread.php?t=69974
I know I saw a detailed blog once but can't find it ATM....
Here's some additional Googling....http://www.google.com/search?hl=en&source=hp&biw=&bih=&q=amiga+1200+fpu&btnG=Google+Search&gbv=1
-
£599 probably - the US prices were:
"The Falcon030 is available in four variations -- 1 meg of RAM and no hard
drive, 4 meg of RAM and no hard drive, 4 meg of RAM and an 80 meg hard
drive, and 14 meg of RAM and an 80 meg hard drive. Retail prices range
from $799 to $1899. Actual street price is generally $100-200 less."
Nope, I remember the UK magazine ads for it. Pricing in the UK almost never reflects the actual dollar to pound currancy conversion. Due to additional taxes etc, if something in the US is $799, it will more often than not end up being £799 (ie about 1.5 times more than what it costs in the US), if not a LOT more. This was the case for many years, and for many products, still is.
-
Yes Sir! :)
http://eab.abime.net/showthread.php?t=69974
I know I saw a detailed blog once but can't find it ATM....
Here's some additional Googling....http://www.google.com/search?hl=en&source=hp&biw=&bih=&q=amiga+1200+fpu&btnG=Google+Search&gbv=1
sweet.. i didn't know that. :D
now. next question.. is it possible to replace the cpu with a proper MC 020?
and thirdly.. is it possible to overclock the system, if so. by how much?
20mhz- 25mhz?? :D
-
sweet.. i didn't know that. :D
now. next question.. is it possible to replace the cpu with a proper MC 020?
and thirdly.. is it possible to overclock the system, if so. by how much?
20mhz- 25mhz?? :D
MC020? Hmmm, maybe, But timing of most of the a1200 circuits are synchronous, which would alter other HW operations if you change the CPU clock. I think any clock change to the CPU would require a separate clock (in theory)... that's why accelerators work.
Interesting read here... doesn't really go anywhere and a promising link is dead.
http://eab.abime.net/showthread.php?t=67006
-
Looking at the surface of the A1200 motherboard. you see prints next to the 020 cpu. looking like the oulines of a 6881/2 fpu. if one found a full blown MC 020 and a fpu... I wonder if it would be possible to replace the cpu and add an fpu directly to the motherboard. and would it be possible to replace the crystal with, lets say 16mhz....
Yes, you can add a 68881. It's been done and there's pictures somewhere here on the forum. But it will only run at 14MHz. And no, you can't replace the crystal, since that will throw off all the Amiga's other timings. :p
Edit: Oops, missed Gizmo's response! ;)
-
@gizmo350 and oldsmobile_Mike
I see. pretty much stuck with the 14mhz speeds then. :(
-
And as to replacing the 68ec20 with a full 68020.
a) you wouldn't get any benefits, since the Mobo still has only 24bit addressspace
b) those extra 8 signals might get confusing signals (assuming there even is an 020 with the same package as the ec)
-
pretty much stuck with the 14mhz speeds then. :(
Add an accelerator! :D
-
@gizmo350 and oldsmobile_Mike
I see. pretty much stuck with the 14mhz speeds then. :(
Stuck with 68EC020@14 MHz? Your sig says:
"Amiga 1200 Tower Os 3.9
BPPC 603e+ 040-25/200, 256MBram, BVIsionPPC, Indivision AGA MK2"
An Amiga 1200 motherboard would not be worth modding even if doubling the clock frequency of the 68EC020 was possible. The performance would still be bottle necked by the lack of data cache (which forces memory to be used directly without any cache for data) and slow chip memory (data accesses are not cached so the CPU waits for these slow memory accesses). If you are poor (it appears not from your sig) then add a fast memory card to a stock 1200. If you have enough money for more performance, a good 68030 (or FPGA) accelerator with fast memory could provide several times stock performance while maintaining excellent compatibility. A 68040-68060 would provide significantly more performance but some compatibility is lost compared to a 68020/68030.
@Oldsmobile_Mike
You beat me with the short answer. You must have had your coffee today ;).
-
Add an accelerator! :D
yeah.. i got one.
but i was just going to tinker about with one of the MB i have....
The thought goes:
Route the 28mhz signal from the oscillator directly to the CLK pin on the cpu. That should in theory run the cpu at 28Mhz.
and add fpu and 4/8MB trapdoor ram card....
This should work, in theory at least. and of course unless one need complicated circuits to maintain the timing with the chipset...
Would be fun to test it out, just to see if it would work....
-
@Oldsmobile_Mike
You beat me with the short answer. You must have had your coffee today ;).
Third cup! :D
Yeah, I saw that sig and was scratching my head, too. Maybe he's trolling? ;)
-
Third cup! :D
Yeah, I saw that sig and was scratching my head, too. Maybe he's trolling? ;)
Trolling.... WTF
Is it no longer possible to have any conversation with fellow amigans withoug getting flamed....?
-
Trolling.... WTF
Is it no longer possible to have any conversation with fellow amigans without getting flamed....?
Trolling was my first thought also. There is no next generation Amiga talk so I figured there was a logical explanation. I refrained from using the word troll but I still made a joke of the "inconsistency" :).
-
I would have loved a 56001 DSP on the A1200. I think they missed the boat in just sticking with Paula. And thats why the Atari produced numerous Cubase revisions.
-
Trolling was my first thought also. There is no next generation Amiga talk so I figured there was a logical explanation. I refrained from using the word troll but I still made a joke of the "inconsistency" :).
The whole tread was about A1200 VS Falcon....
what is then wrong in airing some thoughts about the A1200 and what it could be capable about with minimum expense?
-
Trolling.... WTF
Is it no longer possible to have any conversation with fellow amigans withoug getting flamed....?
OMG, dude. Here in the rational world, it is hard to rectify your sig (which lists many highly expanded Amiga's) with your comment of (quote) "pretty much stuck with the 14mhz speeds then." I'm going to assume that English is not your first language, which is just fine, but if I said "Oh woe is me, I'll never be able to expand my Amiga beyond 14MHz!" and then list such powerful systems as in your signature, it's a head-scratching oxymoron, to say the least.
Or maybe we're just all old and cranky and hate seeing people who should know better asking absurd questions.
And now I need another cup of coffee. Cheers! :D
-
OMG, dude. Here in the rational world, it is hard to rectify your sig (which lists many highly expanded Amiga's) with your comment of (quote) "pretty much stuck with the 14mhz speeds then." I'm going to assume that English is not your first language, which is just fine, but if I said "Oh woe is me, I'll never be able to expand my Amiga beyond 14MHz!" and then list such powerful systems as in your signature, it's a head-scratching oxymoron, to say the least.
Or maybe we're just all old and cranky and hate seeing people who should know better asking absurd questions.
And now I need another cup of coffee. Cheers! :D
Oh wow.... I was referring to A STOCK A1200, as mentioned in the post...
What can possibly be done to an A1200 without buying expensive accelerators...
How can you relate that to my expanded a1200? the post was about stock machines....
if you cant understand something as simple as to stick with the tread... then WTF???
its best to stop here.....
-
Well you could mess with the crystal I guess but you would need
that 14 Mhz 020 to go to 28 (a little too hot probably).
What they should have done is put 68EC020/20 Mhz in there and
make it run at 28 Mhz and put at least 2 Mb fast ram as default.
Other mistake was 2.5" IDE connector as 2.5" IDE hard drives were
most expensive at the time , SCSI was much cheaper but 3.5" IDE
were way cheaper.
What they did is make companies who made software stick to 14 Mhz
2 Mb slow Chip Ram as default configuration and floppies.
If they did it that way it would be 4 x faster than stock A1200 and
software would be made for that speed.It's only 1/2 speed of 030/50
but it's 4 x fast as stock.
-
I would have loved a 56001 DSP on the A1200. I think they missed the boat in just sticking with Paula. And thats why the Atari produced numerous Cubase revisions.
I wanted the Amiga to have a DSP at one time also. As my understanding of processors has increased, I think the Amiga would have been, and would be, better off with more general purpose and flexible processors. A DSP would have offloaded the CPU for audio processing but, IMO, a faster or better CPU, fast memory and a larger Amiga custom chip audio buffer would have given us faster computers and better software today. The key difference is that a program can be compiled and instantly takes advantage of a faster CPU and better memory while doing the same with a DSP gives no advantage. Perhaps a DSP (and some Paula enhancements) would have attracted some professional audio enthusiasts and software developers back in the day. Perhaps there would be a few assembler coders trying to use the DSP today to do tricks like the Falcon DSP is used for. The Amiga received better high end software eventually because of high end 68k processors but then lost out because there were no more new 68k processors to take it higher and compete with faster processors. Today, a DSP would never be put in a general purpose computer as they are limited in what they can do and difficult to program. CPUs today generally have DSP like instructions, specialized hardware and/or an SIMD unit which is more flexible. Superscalar, multi-core, multi-threaded and other parallel processing also reduce any advantage a DSP would have. A DSP can still be cheaper and draw less power for consistent and repetitive tasks. FPGA hardware has replaced DSP processing in many cases. The FPGA code could sometimes be classified as a custom DSP processor. Much more processor customization is possible today with cheap FPGAs.
The whole tread was about A1200 VS Falcon....
what is then wrong in airing some thoughts about the A1200 and what it could be capable about with minimum expense?
Nothing. It is not your fault we thought of trolls but the fault of all the trolls attacking in other threads. Our neural network minds are trained by repetition. I don't think Mike was accusing you of being a troll either.
-
The thought goes:
Route the 28mhz signal from the oscillator directly to the CLK pin on the cpu. That should in theory run the cpu at 28Mhz.
and add fpu and 4/8MB trapdoor ram card....
I think it's unlikely that the 14mhz ec020 will run happily at 28mhz.
An fpu doesn't need to run at the same speed as the cpu, but all that surface mount soldering is not going to be fun. It's cheaper to just buy an accelerator (if you have the skills to do the mods then in the time you've spent tinkering with your a1200 you could earn money doing something else).
-
Trolling.... WTF
Is it no longer possible to have any conversation with fellow amigans withoug getting flamed....?
I knew that you were referring to a stock A1200. They have been very rude to you. If anyone was trolling here it was definitely them with their ridiculous trolling accusations.
-
calm down, pyree was not trolling as far as i can tell. i think it was you lot who have got confussed. I knew straight away he meant that you would be stuck to 14mhz even if you tinkered with the mobo, not that he would never been able to boost his machine personally. It is no wonder this site has become such an empty place, everyone is so eager to jump down everyones throats.
-
Other mistake was 2.5" IDE connector as 2.5" IDE hard drives were
most expensive at the time , SCSI was much cheaper but 3.5" IDE
were way cheaper.
SCSI drives allways were expensive when compared to similar speced IDE drives.
3.5" IDE was no option:
- space (yeah I know they do fit, but surely not according to any spec)
- power consumption (add some RAM/CPU and drain a bit power at the ports and you'll have problem)
- most 3.5" drives at that time had problems sending there parameters, not a problem on x86 where those where strored in NVRAM after been read out at setup. 2.5" drive where much more reliable.
-
3.5" IDE was no option:
- space (yeah I know they do fit, but surely not according to any spec)
- power consumption (add some RAM/CPU and drain a bit power at the ports and you'll have problem)
- most 3.5" drives at that time had problems sending there parameters, not a problem on x86 where those where strored in NVRAM after been read out at setup. 2.5" drive where much more reliable.
A low profile 3.5" drive, an uprated power supply and a buffered ide port worked fine. Yes it's not standard, but then plenty of amiga upgrades weren't standard.
-
3.5" drives made in 1993 or later, or older drives >200MB work without problem.
But in 1992 and on a budget C= would have offered drives around 50-200MB and most of them would not have worked.
-
I didn't have problem with most 3.5" IDE drives as long as they
were less or equal to 4 Gb although name wasn't recognized
from firmware partition program just worked on plain Kick 3.0
and original FFS.Off course I meant that they should have made
bigger case , stronger PSU and better support for those from
the start.
-
But in 1992 and on a budget C= would have offered drives around 50-200MB and most of them would not have worked.
I'm not convinced that most 3.5" drives had broken identify responses, but if commodore had placed an order then the manufacturer would fix it.
What really kills your argument is that the A4000 shipped first in 1992 with 3.5" drives and they worked, so they would work in an A1200.
They used 2.5" drives for size and power reasons.
-
The 68030 in a falcon has no way of accessing any memory at 32 bit. In the Atari world ST-RAM is equivalent to 16 bit chip ram and TT-RAM is 32 bit fast ram, the falcon (without an accelerator) only supports ST-RAM.
no.
at least, strictly speaking:
in Atari world how many bits RAM have does not determine it's name. Name is determinate if RAM is accessible by custom chips.
ST-RAM is ram that is accessible by custom chips (video, audio) and
TT-RAM is accessible only by CPU and not by custom chips
TT-RAM usually is 32bit but e.g. on Atari ST you can have TT-RAM (if you have more than 4MB) but it will be accessed also in 16bit and will be not accessible by custom chips (I think that it is same on e.g. Amiga 500/600: FastRAM would be no faster than ChipRAM coz it will be also 16bit and will be no accessible by custom chips).
So to get good performance you have to pay just as much for as an accelerator as you do on the A1200. Which makes the choice of 68030 rather strange. It's obviously a 68000 design with a 68030 shoe horned in as they only connect 24 address lines. The A1200 on the other hand can get 32 bit fast ram with just a cheap trapdoor ram upgrade. Although the A1200 is limited to 24 address lines unless you buy an accelerator.
memory read/write speed of stock Atari Falcon is faster than stock Amiga 1200:
Falcon ST-Ram R/W 32bits (MB/s): 5.345 / 6.488 (nembench)
Amiga ChipRam R/W 32bits (MB/s): 3.020 / 5.330 (SysSpeed (http://amiga.resource.cx/perf/sysspeed.pl?amiga=A1200&amiga=A4000&testgroup=cpu&testgroup=ram&testgroup=gfx&order=name))
Amiga access ChipRAM RAM in 32bit at ~7MHz and Falcon access ST-RAM in 16bit at 16MHz. Videl (video chip) access ST-RAM in 32bits burst mode. ChipRAM in Amiga 1200 is 250ns and in Falcon 120ns.
BUT on Amiga you could easily (and cheaply!) add FastRAM and outperformance Falcon in memory performances (with FastRAM A1200 should be two times faster?)!
Which makes the choice of 68030 rather strange.
Reason for 68030 is probably MMU so MiNT (MultiTOS) could use memory protection.
and btw MMU in Falcon made X68000 emulator possible :)
Cho Ren Sha - https://www.youtube.com/watch?v=voiRnr72YhQ
Pac-Mania - https://www.youtube.com/watch?v=HELyQzBSB9M
Galaga 88 - https://www.youtube.com/watch?v=kOH_uMRSrDY
and in future maybe NeoGeo if Anima summon some dark magic :D
https://www.youtube.com/watch?v=hxPCYeHtg60
-
Don't forget that the Falcon was also horribly crippled by it's SW (read TOS) which allready was in legacy hell as it wasn't really planned for multitasking or HW expansion.
GEMDOS / TOS from start have support for Accessories.
Accessories are programs that are available from main application so TOS from start could run few programs at same time. Basic for multitasking was there from start but roas was long!
Eventually Atari got three major multitasking OSs: MagiC!, Geneva and MultiTOS (MiNT). Latest have even memory protection.
More about evolution and insides about TOS: http://www.fultonsoft.com/category/atari-st/revisiting-gem-for-the-atari-st/
Plenty other fields were a stock Falcon was utterly useless compared to a stock A1200.
not sure why Falcon would be useless in any field?!
here you have Atari software database with screenshots and some animations: http://milan.kovac.cc/atari/software/
there are software for any field!
-
Amiga ChipRam R/W 32bits (MB/s): 3.020 / 5.330 (SysSpeed (http://amiga.resource.cx/perf/sysspeed.pl?amiga=A1200&amiga=A4000&testgroup=cpu&testgroup=ram&testgroup=gfx&order=name))
No. Just no. Don't use crappy software when you want to get your facts right.
The hw bandwidth is 7MB/s read/write. Use 'bustest' to check how good your cpu is at getting the full bandwidth (I can't remember what the plain 1200 does, but it should be close to the theoretical max).
Just looking at the sysspeed page linked shows so many baloney numbers that you have to shake your head. Please don't use syssped if you want to check Amiga memory speeds.
-
I tried TOS a few times and it was the worst OS I ever used.
-
TOS on the ST was my first GUI experience and was fine for what it was. Nothing compared to AmigaOS obviously but I have to say using MacOS 9 with the Puck iMac Mouse was absolutely soul destroying, unresponsive and the mouse was the worst I've ever used. It put me off Macs for another 8 years!
-
No. Just no. Don't use crappy software when you want to get your facts right.
The hw bandwidth is 7MB/s read/write. Use 'bustest' to check how good your cpu is at getting the full bandwidth (I can't remember what the plain 1200 does, but it should be close to the theoretical max).
hm, ok. But I quote sysspeed numbers since they are better than BusTest. I am pretty sure that something is wrong with these BusTest numbers that I found on internet (can someone verify them on stock A1200?):
Falcon ST-Ram R/W 32bits (MB/s): 5.345 / 6.488 (nembench)
Amiga ChipRam R/W 32bits (MB/s): 3.900 / 3.900 (bustest (http://amigakit.leamancomputing.com/catalog/images/ACA1220_16MHZ_Bustest.JPG))
Amiga FastRAM R/W 32bits (MB/s): 10.900 / 16.4MB/s
---
Regarding price, I found info in old Usenet thread link (https://groups.google.com/d/msg/comp.sys.atari.advocacy/Z28h2Nba21M/h47YS8E-QKwJ):
Base Amiga = 300 UKP
Base Falcon = 600 UKP (for the 1 MB)
Base Falcon = 900 UKP (for the 4 MB / 65MB harddrive)
Amiga = 1/2 to 1/3 cost of Falcon
I tried TOS a few times and it was the worst OS I ever used.
I guess because of this there were many desktop replacement/extension and few different operating system for Atari (all TOS compatibile): Geneva (http://www4.pair.com/gribnif/geneva.htm), MagiC! (https://en.wikipedia.org/wiki/MagiC), MultiTOS/MiNT (https://en.wikipedia.org/wiki/MiNT). They improve every aspect of OS over TOS.
And they looks nice on modern hardware :) - http://www.atari.sk/download/PICs/zview_beta7.jpe
-
GEMDOS / TOS from start have support for Accessories.
Accessories are programs that are available from main application so TOS from start could run few !
If it had been anywhere near real multitasking, it would have been multitasking.
Sound more like the cooperative "multitasking" seen in some versions of 68k MacOS and just like MacOS at that time TOS was heading for a deadend.
Sure solutions could be found, but they allways were compromises between performance, reliability and compability.
AmigaOS3.x on the other hand was pretty much state of the art when it came to consumer OSes and could hold it's water against Win3.x and only got left behind a few years later.
None of the 3 multitasking OSes you mentioned were available/fully useable at the time the Falcon was released.
-
AmigaOS3.x on the other hand was pretty much state of the art when it came to consumer OSes and could hold it's water against Win3.x and only got left behind a few years later.
Windows 3.x was 16 bit and not even an OS yet (GUI for MS-DOS). Universal mouse support and drag and drop were new features. IMO, Windows didn't catch up until about Windows 2000. Windows may have looked better on paper by features before that where the AmigaOS 3.x was underrated because of lack of certain features but was more responsive, more flexible and more user friendly.
-
Win3.x was what was available when the A1200 and Falcon came out, making it the right measure.
TOS failed even by that measure, AOS.3.0 was "state of the art", wether it was just as good or better than Win does not really matter.
-
Amiga ChipRam R/W 32bits (MB/s): 3.900 / 3.900
Amiga FastRAM R/W 32bits (MB/s): 10.900 / 16.4MB/s
I think 0.19 is the most up to date version of bustest btw.
Never had a 1200 so I only remember my 4000/CS060 which did near as much 7MB/s and something like 40-50MB/s to fastram. (Been a while since it was alive so I can't say for sure about the fastram.)
A number of cpu cards handled write-combining and pipelining/delayed writes pretty bad and had worse numbers than a stock machine...
-
GEMDOS / TOS from start have support for Accessories.
Accessories are programs that are available from main application so TOS from start could run few programs at same time. Basic for multitasking was there from start but roas was long!
Eventually Atari got three major multitasking OSs: MagiC!, Geneva and MultiTOS (MiNT). Latest have even memory protection.
More about evolution and insides about TOS: http://www.fultonsoft.com/category/atari-st/revisiting-gem-for-the-atari-st/
not sure why Falcon would be useless in any field?!
here you have Atari software database with screenshots and some animations: http://milan.kovac.cc/atari/software/
there are software for any field!
Although not TOS compatible, there is also a port of Microware's OS9 multi-tasking/multi-user operating system for early Atari 68K based machines.
Although it is doubtful that that would run on a Falcon.
-
Win3.x was what was available when the A1200 and Falcon came out, making it the right measure.
TOS failed even by that measure, AOS.3.0 was "state of the art", wether it was just as good or better than Win does not really matter.
Actually quite true.
I was working with 68K systems when Win 3.0 was in its beta stages (we had early copies provided by IBM).
Although primitive, Win3.1 did give PCs some parity with the Amiga once the MPC standard was more widely adopted.
And the ST couldn't match either system.
You guys have to remember that while somewhat inferior, MPC based systems had many of the features that the Amiga processed (and a few they didn't).
Its not any big surprise to me that one of the big backers of the MPC standard was Tandy.
They also produced some of the first MPC systems.
Remember, they had a good working relationship with Motorola (so good in fact that they requested and got a specific upgrade to the 6847 VDG used in the Color Computer - the 6847T1, note the "T" designation, AND they had some help in designing the custom chip that replaced that in the Color Computer 3).
How do I know that?
I had phone access to their engineers directly (remember, we were developing 68K based systems) and they were pretty free with information.
So...in case you don't know where this is going...one of the early backers of the Windows based MPC standard was quite familiar with Motorola based systems and you can be assured that they were aware of the Amiga.
In fact, the Amiga was the standard I think they were aiming at, not other PCs.
-
...Eventually Atari got three major multitasking OSs: MagiC!, Geneva and MultiTOS (MiNT)...
As well as operating systems originally intended for the Sinclair QL and its descendants like SMS2.
-
As well as operating systems originally intended for the Sinclair QL and its descendants like SMS2.
I'm surprised nobody ever ported amigaos to it.
-
Today, a DSP would never be put in a general purpose computer as they are limited in what they can do and difficult to program.
That depends on what you mean by general purpose. A lot of home users use their computers for multimedia or games, and some have sound cards in their computers for this purpose, and these will usually contain a DSP. For example, if you had a Sound blaster from the late 90's or later you had a DSP in your computer, and some modern graphics cards have them. The Playstation 4 has one, although it's a quite a stretch to call that a general purpose computer.
Of course, the very name "DSP" implies a non-general purpose, but practically speaking, sound and graphics processing should be covered.
-
I'm surprised nobody ever ported amigaos to it.
Well. that is not an open OS...
Come to think of it, I can't think of anything other than an Amiga that will run AmigaOS (outside of emulation).
-
There is the Draco (and even some early version of the Casablanca )
Problem is that AmigaOS needs atleast the CIAs to work at all and either the rest of the chipset or massive patching (CGX/P96) to be really usefull.
Quite a difference to running MacOS or TOS on an Amiga where all you needed was the ROM-images and a few mild patches.
-
That depends on what you mean by general purpose. A lot of home users use their computers for multimedia or games, and some have sound cards in their computers for this purpose, and these will usually contain a DSP. For example, if you had a Sound blaster from the late 90's or later you had a DSP in your computer, and some modern graphics cards have them. The Playstation 4 has one, although it's a quite a stretch to call that a general purpose computer.
A DSP on sound cards was more common before modern CPUs which have more processing power and DSP like functionality, especially SIMDs. There still is the possibility of off-loading the CPU but most audio processing only uses a few percent of a modern medium to high performance CPU's processing power. An SIMD often has more performance than a DSP albeit at a higher cost. Perhaps the PS4 is trying to save electricity as consoles need to be more power efficient and a DSP wins over an SIMD here. However, the PS4 likely has multiple SIMD units which could do more complex audio processing faster and the overall hardware cost could be reduced by a small amount. It would be interesting to hear why the PS4 engineers included the DSP. I wouldn't be surprised if it was because they used an existing off the shelf mature audio design which included a DSP. There are probably DSPs on motherboard integrated PC audio hardware for this reason. Newly designed general purpose computers are unlikely to get an off the shelf DSP like the Falcon or A3000+. Many people seem to still want a DSP when they would be much better off with a faster CPU and/or SIMD. That was my point even if my use of "never" was perhaps too harsh.
Of course, the very name "DSP" implies a non-general purpose, but practically speaking, sound and graphics processing should be covered.
A DSP can be used for some graphics processing and they are efficient but most are not particularly flexible or fast. An SIMD is a better choice for graphics as it is more powerful, more flexible and generally easier to program. We have modern graphics processors and hardware which are more powerful yet. Audio processing is mostly simpler, more repetitive and less processor intensive which is perfect for a DSP. Still, there are many reasons why a DSP processor chip would not be used in a completely new design.
-
interesting topics: SIMD and DSP in modern days...
I read that AmigaOne X1000 have XMOS xCORE chip (and XMOS have some roots in Transputer era :)
is there any practical use of this chip or any demo that would show possibilities of xCORE?
btw IBM Cell looks like DSP: SPE's has it's own memory, no cache...
-
There have been many interesting ideas but from what I understand that all tools have not been ported yet.
Some ideas: Telescope controller/viewer, Home automation controller and others. There are some threads in the forums with many of them.
-
interesting topics: SIMD and DSP in modern days...
I read that AmigaOne X1000 have XMOS xCORE chip (and XMOS have some roots in Transputer era :)
is there any practical use of this chip or any demo that would show possibilities of xCORE?
Define practical. More processing power using extra processors can always be used but is it worthwhile? Weak processors with strong processors are generally not worth messing with for most programmers. Take the classic Amiga blitter for example, it is mostly unused on accelerated Amigas even though a few programs have used the blitter and fast CPU in parallel. Processors which are distant from the CPU like the Amiga blitter, a DSP, an xCORE or a GPU unit take a long time to setup as compared to a CPU co-processor like an SIMD, MMU or FPU. This is one of the reasons why DSP like instructions (ARM) and SIMDs are added to modern CPUs. OS software libraries can make it easier to use CPU external processors like the graphics.library does for the blitter but the use is generally narrow. Many weak processors can be used in parallel making them worthwhile. Some weak processors can offload I/O processing and have external hardware connections. The xCORE processors looks like they fit in this category but were not added for a particular purpose. I have only done basic reading about the xCore but it looks to me like a descent sized user usable FPGA would have been a better choice with more flexibility and processing power. The SAM Lattice FPGA is the right idea but it is too small for all but the simplest uses (GeekPort for embedded uses). Add a more popular FPGA big enough to handle retro chipsets and CPUs with open documentation and development tools and then you would have something retro and embedded markets would be attracted to.
btw IBM Cell looks like DSP: SPE's has it's own memory, no cache...
There are some similarities including the difficulty of programming which is what has killed Cell. The theoretical peak processing power of Cell is very good if the code is perfect. The PPC processor in Cell also looks good on paper but it is bubblicious. Cell SPEs may use less electricity than SMP but the difficulty of coding killed it. Ease of programming and compiler construction keeps getting ignored for theoretical performance and processor simplification (compilers often can't handle the complexity as well as the hardware processor).
-
A DSP on sound cards was more common before modern CPUs which have more processing power and DSP like functionality, especially SIMDs. There still is the possibility of off-loading the CPU but most audio processing only uses a few percent of a modern medium to high performance CPU's processing power. An SIMD often has more performance than a DSP albeit at a higher cost. Perhaps the PS4 is trying to save electricity as consoles need to be more power efficient and a DSP wins over an SIMD here. However, the PS4 likely has multiple SIMD units which could do more complex audio processing faster and the overall hardware cost could be reduced by a small amount. It would be interesting to hear why the PS4 engineers included the DSP.
As a guy who has done nothing but audio coding my whole career I can agree that being able to move everything off of DSPs and on to the high performance main CPUs that came along is a great thing, at least for us writing audio code. Auxiliary processors can box you in and are generally just a PITA to deal with. They are also often closed off, only presenting some high level API to interact with.
I might be cautious in saying the PS4 has a "DSP". It is not targeted for general purpose audio processing, the main CPU is what does all that. The "DSP" mostly deals with encoding and decoding. Is that a DSP, or is it just another auxiliary processor that is cost effective at that task? There might be some other DSPs on the machine I am unaware of, there are a number of them that perform various tasks, but none that synthesize the game audio like some might think.
I'm also not sure if you'd say the Cell is super DSP-like. Its instruction set is quite general purpose, but designed for modern SIMD computing (lean and mean). Its memory architecture is not at all general purpose, and more DSP like, I guess. It's interesting, to say the least.
Reading this thread I now kind of wish I had a Falcon way back when it came out. It would have been a great machine to code the audio software I wanted to develop with its 16-bit capability, and back then the DSP was needed with the main CPU being too weak for anything interesting with audio. I guess I had moved on to PCs and sound cards by that time though.
Relating this back to the Amiga I wouldn't actually mind if someone came out with a new sound card for it that had a bitchin' DSP-like processor on it. Given the Amiga tops out at 68060 you kind of need it if you want to do lots of crazy stuff. Either that or some of these new accelerators that could move us way past 060 performance, that would be even better.
-
Problem is that AmigaOS needs atleast the CIAs to work at all and either the rest of the chipset or massive patching (CGX/P96) to be really usefull.
Yes, essentially you need to write a new bootstrap, timer device, new trackdisk device, a cgx/p96 driver, AHI device and an scsi/ide device.
Quite a difference to running MacOS or TOS on an Amiga where all you needed was the ROM-images and a few mild patches.
It's unlikely to be much different to be honest. At least the amount of Mac patches wasn't trivial.
I might be cautious in saying the PS4 has a "DSP". It is not targeted for general purpose audio processing, the main CPU is what does all that. The "DSP" mostly deals with encoding and decoding. Is that a DSP, or is it just another auxiliary processor that is cost effective at that task?
It's mostly a marketing term. A DSP is an embedded cpu that has certain attributes. It can be manufactured to purely run unchangeable software from an internal ROM to perform a single function and still be a DSP. Old modems for instance used DSP's, flash came along eventually and some of them could be upgraded. The first DSP was the TMS5100 used in the speak and spell.
http://ethw.org/Milestones:Speak_%26_Spell,_the_First_Use_of_a_Digital_Signal_Processing_IC_for_Speech_Generation,_1978
Within recent years a tms5100 chip has been decapped and the mask rom dumped.
-
As a guy who has done nothing but audio coding my whole career I can agree that being able to move everything off of DSPs and on to the high performance main CPUs that came along is a great thing, at least for us writing audio code. Auxiliary processors can box you in and are generally just a PITA to deal with. They are also often closed off, only presenting some high level API to interact with.
It's nice to have someone from the industry give their insights.
I might be cautious in saying the PS4 has a "DSP". It is not targeted for general purpose audio processing, the main CPU is what does all that. The "DSP" mostly deals with encoding and decoding. Is that a DSP, or is it just another auxiliary processor that is cost effective at that task? There might be some other DSPs on the machine I am unaware of, there are a number of them that perform various tasks, but none that synthesize the game audio like some might think.
Yes, the lines are blurred between what is and isn't a DSP processor. FPGAs can be programmed to handle repetitive little data processing tasks or an FPGA DSP processor can be used but stripped of any functionality it doesn't need (before the development FPGA become a hard chip). Specialized and customized processors are very efficient and much easier to create today.
I'm also not sure if you'd say the Cell is super DSP-like. Its instruction set is quite general purpose, but designed for modern SIMD computing (lean and mean). Its memory architecture is not at all general purpose, and more DSP like, I guess. It's interesting, to say the least.
The SPEs are probably more like modern GPU shading units (which have also become more general purpose) while they are all data stream processors. I wonder if the Cell SPEs could have been supported with OpenCL and an OS library for OpenCL. Sometimes how the hardware is offered and supported makes all the difference. An FPGA can also be configured as parallel OpenCL units and an opencl.library could support available GPU shading units, FPGA parallel processing units and/or the main CPU/FPU/SIMD for parallel tasks.
Relating this back to the Amiga I wouldn't actually mind if someone came out with a new sound card for it that had a bitchin' DSP-like processor on it. Given the Amiga tops out at 68060 you kind of need it if you want to do lots of crazy stuff. Either that or some of these new accelerators that could move us way past 060 performance, that would be even better.
A DSP on a sound card still makes some sense as the CPU can vary and is far away across a shared bus introducing the possibilities of unacceptable latencies. Fast modern buses and large buffers probably help but most customers aren't going to tolerate a sound card which glitches and pauses.
Too bad A-Eon didn't realize there was an audio and DSP expert on their forum. The Prisma Megamix could have become a real sound card ;).
Gunnar was playing with an SIMD unit for the Apollo 68k FPGA core. It is Altivec like but with some simple CISC 68k style addressing modes added and some memory alignment restrictions dropped which would make it much easier to use while remaining mostly compatible with Altivec at the instruction level. Integer support is fairly cheap to implement in an SIMD but floating point is expensive. My advise was to support the 68k FPU (like 68060 FPU but with a few enhancements) for compatibility and forget about trying to add an SIMD unit until there is room (and probably single precision FP support as well). As I have thought about it more recently though, I would like to explore adding 2x 16 bit operations in a 32 bit register to the 68k integer units. Encoding space is limited so I would leave out 4x 8 bit operations. RGBA type byte color component operations could still be done in 1 cycle with superscalar instruction interleaving and there was already support for a PERM instruction which could load bytes into registers in any byte order wanted. It would have been helpful to have someone with industry experience in audio processing as part of our group. Maybe Gunnar would have even listened to you but probably not. He creates these nice "teams" which work to pull in ideas and build a consensus and then he ignores the consensus and does everything his own way with little explanation. The technology is real though and 68060 performance is being surpassed in an affordable FPGA. Perhaps up to double the performance of the 68060 is possible in the next few years in an affordable FPGA but "way past 060 performance" would probably require an ASIC. An ASIC would probably be required to fit a good SIMD unit with single precision FP also.
-
interesting topics: SIMD and DSP in modern days...
I read that AmigaOne X1000 have XMOS xCORE chip (and XMOS have some roots in Transputer era :)
is there any practical use of this chip or any demo that would show possibilities of xCORE?
btw IBM Cell looks like DSP: SPE's has it's own memory, no cache...
Although I did see a cool scrolling Super Mario demo the xCORE is aimed more at I/O operations.
One of the best projects I've seen, and experienced, was Randon International's Rain Room. It's used sensors and valves controlled by an xCORE processor to create a corridor of torrential rain that parted around you as you walked through it.
With the X1000 and X5000, it is only really likely to come of use to people with the skills and need/desire to create something with it, and whether or not that would result in something that would be of interest to the average user is another matter.
I can imagine that if the guys who used Amigas at NASA were still there it might be something they'd be interested in.
The main thing going against it is that the small userbase of A-EON hardware limits the chances of someone creating something really interesting with it. The other problem is that so far it hasn't been demonstrated that an xCORE chip connected to the CPU local bus has any advantage over one that isn't.
-
I played with an Atari ST the other day. It hasn't aged well, it's pretty crappy!
-
I played with an Atari ST the other day. It hasn't aged well, it's pretty crappy!
True, but then most of us wear rose colored glasses when waxing poetic about the Amiga.
Fact is, at least with games, it doesn't seem to perform much better than a Sega Genesis.
-
I played with an Atari ST the other day. It hasn't aged well, it's pretty crappy!
I thought you might be trolling but then noticed you are a moderator. An original unexpanded 500, 1000 or 2000 isn't much better.
Fact is, at least with games, it doesn't seem to perform much better than a Sega Genesis.
The Sega Genesis came out some 3 years later than the Atari ST with more powerful hardware (plus its a computer vs console). That would be almost as bad as comparing the Genesis to the Atari Falcon which came out several years after the Genesis. The Amiga 1200 vs Atari Falcon and Amiga 1000 vs Atari ST are much fairer comparisons.
-
An original unexpanded 500, 1000 or 2000 isn't much better.
There are some ok Atari ST games and some ST games were ported to the Amiga without making many changes. The top Amiga games were much better on the Amiga though (for example https://www.youtube.com/watch?v=yNGXuQKSjhI)
The Sega Genesis came out some 3 years later than the Atari ST with more powerful hardware (plus its a computer vs console).
The major difference is it used a sprite and tile engine rather than bitmaps. Tiles work great for games, not so well for word processors.
-
There are some ok Atari ST games and some ST games were ported to the Amiga without making many changes. The top Amiga games were much better on the Amiga though (for example https://www.youtube.com/watch?v=yNGXuQKSjhI)
I was talking about the general utility of the Amiga 500, 1000 and 2000 systems and not just games. In their minimum configurations, they were bottled up so that the advantages of the custom chips and multitasking did not shine. Early AmigaOS 1.x wasn't that great either. Add fast memory, an HD and a newer version of AmigaOS and they start to shine. The Amiga definitely had more potential than the Atari ST but cost more too. Shadow of the Beast probably needs 1MB of memory and maybe a fat Agnus where most early Amigas came with 512kB and the old Agnus.
-
Shadow of the Beast probably needs 1MB of memory and maybe a fat Agnus where most early Amigas came with 512kB and the old Agnus.
Some people online claim they were able to run shadow of the beast on an a1000. Whether it was unexpanded or not is another matter.
Graphically it should be doable on an unexpanded amiga as it's mostly just a copper list, but I don't believe you could match it on an ST no matter what you did to it.
-
Some people online claim they were able to run shadow of the beast on an a1000. Whether it was unexpanded or not is another matter.
There was a cracked 512K only version too.
-
What does the Atari ST have to do with any of this?