Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: HenryCase on December 07, 2007, 01:31:42 AM
-
First of all, let's start by introducing the device we are discussing:
http://www.amiga-hardware.com/showhardware.cgi?HARDID=239
According to amiga-hardware.com, the PowerPC developer board was designed to plug in to the 060 socket of the Cyberstorm Mk-II, but I know it works in the Blizzard 2060 060 socket too.
Seems to me that the only link the PPC developers board has to the main accelerator is the 060 socket. My question is this: If new drivers were written, could the Phase5 PPC developers board work in any Amiga 060 accelerator without hardware modification?
This is just a technical question. Please leave the reasons I am asking this aside for a moment. Thank you in advance.
-
Maybe the answer is yes, but where you hope to find one? That is an Über-rare prototype board!
Rarer than hen's teeths!
-
rkauer wrote:
Maybe the answer is yes, but where you hope to find one? That is an Über-rare prototype board!
Maybe I don't intend to find one, merely create a device like it. As I said before, this is a technical question, please focus on that aspect. Thank you.
-
I smell secret plans to acquire world domination =).
-
I tried to read between the lines, but all I could see was white bits :lol:
Sounds like HenryCase has a cunning plan :idea:
-
Does indeed sound cunning.
I don't have one, so I cannot tell you for sure, but the fact that it not only works on a CS MK2 but also on a 2060 is indeed hopeful.
It really depends on if you have to flash the CS MK2/2060 to work with the PPC board. If you do it will be a no-no, if you don't you might have a chance.
It would be very interesting to see if OS4 would run on the 2060/PPC combo, or is the A2000 just too different...
-
[reading mind]
Well there are 601 / 603e accelerators for old 68k MACs but don`t get you hopes up. There is a flash / ROM, whatever, with special amigish content in it.
[/reading mind]
Back to your basic question: Yes, if there are drivers (in ROM and / or on HD) it should be possible to make such a thing work. But getting drivers done is usually the hardest part, otherwhise we would already have Radeon9800s for G-Rex/Mediator/A1....
-
.
-
I theory yes. Actually a good basis to work from, if you had the schematics for it.
-
Thanks for all of the responses.
My query actually comes out of desperation, I had a much better plan that now seems unlikely to come to anything. Let me explain...
After reading a topic on amiga.org about PPC hardware for Minimig, I decided to contact DCE about acquiring the rights to the Blizzard 2060, Blizzard 2604 and PPC developers board. This would have given us a great head start in putting a PPC Minimig together, as we would have had the hardware designs, the driver source code, all documentation, etc...
However, despite a long discussion that lasted over a month, the negotiations are becoming increasingly difficult, due to DCE's reluctance to do any Amiga business at all. I did find out that we would need Specctra-compatible router software to view the PCB designs, but that's all the technical information I extracted.
The deal would have been good for DCE as it would have given them money for technology they no longer have a use for. However, at the moment I am looking to change my approach to obtaining more PPC hardware for Amiga by doing the following:
1. Seeing if DCE will sell all remaining stock they have for OS4 compatible A1200/A3000/A4000 accelerators.
2. Looking into reverse engineering the PPC developer board, to start PPC Minimig developments going.
I am really not knowledgeable enough at the moment to do all of the reverse engineering work myself, but I'm hoping if I start the work others will be willing to help me where they can. Here's how I see the reverse engineering being possible:
1. Develop a board that mimicked the original developer board (603PPC, 060 processor) with one major difference, an FPGA replacing the control chips.
2. Decompiling the Developer Board drivers to gain a better understanding of how the control chips work.
3. Using the knowledge gained from decompiling drivers, and the API information Framiga mentioned, program the FPGA so that the new device works with the existing drivers.
This would just be the first step towards PPC Minimigs, but an important step IMO. Downix, your Minimig accelerator slot project is also key here, how have things been going?
Tahoe wrote:
It would be very interesting to see if OS4 would run on the 2060/PPC combo, or is the A2000 just too different...
I have been wondering this myself. If anyone does have a suitable A2000 setup to try OS4 on, please look into this.
Incidentally, this is the video that really sold me on the idea of the possibilities of the Blizzard 2060 and PPC developers board combo, it shows the device being installed on an A2000 along with OS3.9:
http://8bit.dk/pepe/a2000ppc.html
Any thoughts?
-
603PPC, 060 processor
Are those CPUs available anymore?
-
HenryCase wrote:
This would just be the first step towards PPC Minimigs, but an important step IMO. Downix, your Minimig accelerator slot project is also key here, how have things been going?
Schematics are done, it's the pcb that I am slow at. (never been fast at pcb's sad to say) Would you like me to post the schematics as/is so as to see what I am attempting to do?
-
HenryCase wrote:
After reading a topic on amiga.org about PPC hardware for Minimig, I decided to contact DCE about acquiring the rights to the Blizzard 2060, Blizzard 2604 and PPC developers board. This would have given us a great head start in putting a PPC Minimig together, as we would have had the hardware designs, the driver source code, all documentation, etc...
However, despite a long discussion that lasted over a month, the negotiations are becoming increasingly difficult, due to DCE's reluctance to do any Amiga business at all. I did find out that we would need Specctra-compatible router software to view the PCB designs, but that's all the technical information I extracted.
Actually, tha fact that you have been negotiating for over a month in itself is pretty revolutionary, usually DCE bins everything even partly relating to Amiga.
I still wonder why they are so adiment in not wanting to do business with us.
All I can say is don't give up, get them to read forums or postings there. I think there are quite a few people around that want to do business with DCE. Either by buying them out , or even wanting them do do a new production run on newly designed hardware. I have even heard people try to persuade DCE to build MiniMigs.
In the end I suspect it is a money issue, I can't image them rather having it gather dust then making a last profit on selling the technology.
-
I dont get it. I also dont get the point of UAE with PPC emulation to run OS4 - why emulate the old Amiga hardware when OS4 was made to not need it? Why bother with PPC on the MiniMig, when OS4 doesnt need it? (Hohum, does it even support ECS directly, or does it rely on the CVPPC?)
It should really be alot easier to hack OS4 to boot on PearPC and old Macs than to build new "old" PPC acc-boards. I'd guess it's even easier to build new proper PPC boards using modern chips, than to build old acc-boards.
-
itix wrote:
603PPC, 060 processor
Are those CPUs available anymore?
Yes, 060's come at up to 100Mhz now, and 603e's can reach 300Mhz.
Incidentally, I have a bus interface module designed for the 750-series as well. Was going to use it in Eddas back when I flirted with using a PPC (too many headaches for me), but glad to donate it if you'd like.
-
Yes, 060's come at up to 100Mhz now
You have reference for this, other than slightly vapored CT60?
Freescale certainly has no clue about this, they sell full 060 at 50MHz and 60MHz, and have LC (broken FPU) and EC (broken FPU and MMU) available up to 75MHz.
http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MC68060
-
It'd be nice if you could put up a page, or ask someone to host a page, about the schematics/pcbs/modules etc that you've got. It might come to nothing but it could also give people a head start with their own interests.
I'd be willing to donate the hosting for them. PM me if you're interested.
Andy
-
kolla wrote:
Yes, 060's come at up to 100Mhz now
You have reference for this, other than slightly vapored CT60?
Freescale certainly has no clue about this, they sell full 060 at 50MHz and 60MHz, and have LC (broken FPU) and EC (broken FPU and MMU) available up to 75MHz.
http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MC68060
Have a dishwasher with a 68EC060 that runs at 100Mhz, didn't look to see if a) it's overclocked or b) if it was available in normal 060's. My appologies, what I said, retracted.
-
@HenryCase
I take it that you've basically said to DCE: "I just want to produce more of these cards and give you money, I'll take care of all support, manufacturing and other issue, where is the part of that deal you don't like?" :-D
Andy
-
downix wrote:
Schematics are done, it's the pcb that I am slow at. (never been fast at pcb's sad to say) Would you like me to post the schematics as/is so as to see what I am attempting to do?
Downix, that would be useful.
Tahoe wrote:
Actually, tha fact that you have been negotiating for over a month in itself is pretty revolutionary, usually DCE bins everything even partly relating to Amiga.
I still wonder why they are so adiment in not wanting to do business with us.
Tahoe, yes it is unusual that DCE haven't wanted to do business with us. The only explanation I was given was that DCE lost a lot of money in the Amiga market (most likely from when the market was collapsing), but surely that's even more reason to sell me/us the rights to the Amiga devices they own? Anyway, I haven't given up on DCE just yet, I'm just trying to look at all our options.
kolla wrote:
Why bother with PPC on the MiniMig, when OS4 doesnt need it?
Kolla, OS4 does need PPC to run.
kolla wrote:
I'd guess it's even easier to build new proper PPC boards using modern chips, than to build old acc-boards.
Recreating the Phase5 PPC developer board is just the beginning, the ultimate plan for this tech is using it to develop new PPC accelerators for classic hardware, PPC Minimigs, etc...
downix wrote:
Incidentally, I have a bus interface module designed for the 750-series as well. Was going to use it in Eddas back when I flirted with using a PPC (too many headaches for me), but glad to donate it if you'd like.
Should be useful for future developments, and any PPC information is very welcome, I have a lot of learning ahead of me. Thanks downix.
Eddas, what's that?
AJCopland wrote:
It'd be nice if you could put up a page, or ask someone to host a page, about the schematics/pcbs/modules etc that you've got.
AJCopland, I'm assuming you're talking to downix here, as I don't have any schematics yet. However, when I do get this project up and running I would be looking for somewhere to host the information, so I will PM you at this point to see if you can help me. Hope that's alright.
AJCopland wrote:
@HenryCase
I take it that you've basically said to DCE: "I just want to produce more of these cards and give you money, I'll take care of all support, manufacturing and other issue, where is the part of that deal you don't like?" :-D
Andy
Not too far off AJCopland. :-D
Basically I said I'd like to buy the rights to the Blizzard 2060/2604 from DCE (I'd discuss the PPC developers board once I could ensure that the other devices were on the table). I offered to do the vast majority of the legal work, travel to Germany to pick up the kit, etc... I even offered them €50 just to have some information on the devices (name of software used to write drivers, volume of stock available, etc...). Not interested (so far).
Manufacturing was a different issue. I always intended to recreate the tech using FPGAs (Virtex-4s maybe) so that we could be free from legacy hardware speeds and non-RoHS compliant parts. DCE know this is my plan. I have had encouragement from them to build a whole new PPC accelerator, but that's a lot more extra work and would remove the possible OS4 benefits (if OS4 can run on Blizzard 2060+PPC dev board that is!).
-
HenryCase wrote:
downix wrote:
Incidentally, I have a bus interface module designed for the 750-series as well. Was going to use it in Eddas back when I flirted with using a PPC (too many headaches for me), but glad to donate it if you'd like.
Should be useful for future developments, and any PPC information is very welcome, I have a lot of learning ahead of me. Thanks downix.
I'll dig it out It is based on the Xilinx reference design for the Virtex 2, so it will likely need to be adapted to a new source
Eddas, what's that?
Enhanced Direct Data Acceleration System, a chipset I've been tinkering with since '97. By tinkering I mean writing specs, doing some coding, finding that I made a judgement error, go back to specs, recode, retest... At present it's a SPARC based IO Processor w/ DSP accelerator tied to a a video display system and audio engine. Never likely will finish it, no time to finish development, and the target keeps moving.
-
About DCE:
They are afraid that once they sold anything to anyone they will be buged with questions: "How does this work, where is the missing page 89 of the documentation, do you think it will be OK exchange nomoremade part A by yetavailable part b"
I`m afraid unless you can garantee them a deal that goes like this: "I give you the money, you give me the stuff and after that we will never talk to each other again" your chances of success are slim.
About OS4 on Blizz2060 + dev board:
I doubt that it will work out of the box. For WarpOS for example a different version of powerpc.library is needed for the dev board (you can choose between BlizzardPPC, CyberstormPPC and DeveloperbBoard during WarpOS installation).
Though it might only be minor changes like how to initialise the PPC or where in the system it can be found, but they would require changes in OS4. With about 15-20 of those dev boards in existance it is (yet) doubtfull that Hyperion will put any effort in making the required changes.
-
I`m afraid unless you can garantee them a deal that goes like this: "I give you the money, you give me the stuff and after that we will never talk to each other again" your chances of success are slim.
I agree, or make it clear that anything after the fact would be on a paid consultation basis.
-
Regarding PPC capability. Isn't it more efficient to use a Xilinx Virtex-4 FX or similar with builtin PPC ..?
(which can run ordinary m68k minimig on the standard logic matrix)
V-4FX
http://www.xilinx.com/products/silicon_solutions/fpgas/virtex/virtex4/capabilities/index.htm
V-II Pro
http://www.xilinx.com/products/silicon_solutions/fpgas/virtex/virtex_ii_pro_fpgas/capabilities/index.htm
Is there anything in particular that makes it worthwhile to persue DCE rather than reverse engineer existing hw/sw ..?
Is there a standard spec for Amiga PPC accelerators, or a defacto one?, otherwise maybe one have more freedom in choosing the functionality.
-
kolla wrote:
Why bother with PPC on the MiniMig, when OS4 doesnt need it?
Kolla, OS4 does need PPC to run.
What I meant is that OS4 doesnt need MiniMig :-)
-
downix wrote:
I'll dig it out It is based on the Xilinx reference design for the Virtex 2, so it will likely need to be adapted to a new source
Thanks downix. Please contact AJCopland if you need help with hosting the information.
downix wrote:
Enhanced Direct Data Acceleration System, a chipset I've been tinkering with since '97.
Sounds interesting. From your description it sounds like a configurable input device platform which also handles video and audio processing. Is that right? How do you envision the chipset being used, what improvements would it bring to a system (if it was finished)?
Lemmink wrote:
About DCE:
They are afraid that once they sold anything to anyone they will be buged with questions: "How does this work, where is the missing page 89 of the documentation, do you think it will be OK exchange nomoremade part A by yetavailable part b"
I`m afraid unless you can garantee them a deal that goes like this: "I give you the money, you give me the stuff and after that we will never talk to each other again" your chances of success are slim.
That is a very good point, I never assumed that this is why DCE would be reluctant to do business with me. I certainly will not be bugging them with questions should I ever purchase the board. I will make this clear in my next e-mail to them.
Lemmink wrote:
About OS4 on Blizz2060 + dev board:
I doubt that it will work out of the box. For WarpOS for example a different version of powerpc.library is needed for the dev board (you can choose between BlizzardPPC, CyberstormPPC and DeveloperbBoard during WarpOS installation).
Though it might only be minor changes like how to initialise the PPC or where in the system it can be found, but they would require changes in OS4. With about 15-20 of those dev boards in existance it is (yet) doubtfull that Hyperion will put any effort in making the required changes.
These are definitely concerns, but I would like to see if OS4 can use the developer board option, as it could be an undocumented feature. Preferably I'd like someone to test OS4 on a Blizzard 2060 + PPC dev board. I know the amiga.org user hakker has such a device, hopefully I can get hold of him soon. Does anyone else here own a Blizzard 2060 + PPC dev board? I'm prepared to buy a copy of OS4 for whomever is willing to test out compatibility for us.
If OS4 compatibility isn't offered straight out of the box, it might be possible to create a small workaround without breaking OS4. For instance if OS4 currently needs a certain library file to be present at boot time, we could create a dummy file with the same name and use it to pass calls through to the developer board drivers.
freqmax wrote:
Regarding PPC capability. Isn't it more efficient to use a Xilinx Virtex-4 FX or similar with builtin PPC ..?
Freqmax, I've been thinking about the Virtex-4 FPGAs for a while now. The fact that chips like the Virtex-4 exist is one of the major reasons my project is even slightly feasible.
freqmax wrote:
Is there anything in particular that makes it worthwhile to persue DCE rather than reverse engineer existing hw/sw ..?
The time taken to develop a brand new device vs the time taken to adapt an existing device to our needs still makes DCE worth pursuing IMO. Reverse engineering would give us the results we want but we could get to where we wanted to be much faster if we had access to the designs.
freqmax wrote:
Is there a standard spec for Amiga PPC accelerators, or a defacto one?, otherwise maybe one have more freedom in choosing the functionality.
As far as I knew Phase5/DCE were the only companies to ever release PPC accelerators for the Amiga, so I suppose their devices are the de facto ones. Did any other companies develop PPC hardware for Amiga?
kolla wrote:
What I meant is that OS4 doesnt need MiniMig :-)
Kolla, OS4 definitely needs more hardware capable of running it. Minimig could be one of those solutions. I don't expect any new OS4 hardware from AInc (or OS4 support for existing hardware from Hyperion) until the court case is settled. In fact, new OS4 hardware/support may not even come along at all.
-
kolla wrote:
What I meant is that OS4 doesnt need MiniMig :-)
Kolla, OS4 definitely needs more hardware capable of running it. Minimig could be one of those solutions.
OK?
Let's say you've made a MiniMig with PPC - now what?
-
kolla wrote:
kolla wrote:
What I meant is that OS4 doesnt need MiniMig :-)
Kolla, OS4 definitely needs more hardware capable of running it. Minimig could be one of those solutions.
OK?
Let's say you've made a MiniMig with PPC - now what?
After a PPC Minimig was created, we'd probably need to do some work to get OS4 working on it. After this, the OS4 h/w problem is solved.
More OS4 h/w means more users of OS4, which means more OS4 development. How hard is that to understand?
-
HenryCase wrote:
downix wrote:
I'll dig it out It is based on the Xilinx reference design for the Virtex 2, so it will likely need to be adapted to a new source
Thanks downix. Please contact AJCopland if you need help with hosting the information.
Have my own host, thankfully, just need to get off my butt and get it all together
downix wrote:
Enhanced Direct Data Acceleration System, a chipset I've been tinkering with since '97.
Sounds interesting. From your description it sounds like a configurable input device platform which also handles video and audio processing. Is that right? How do you envision the chipset being used, what improvements would it bring to a system (if it was finished)?
Rather than stcking to old standards, which eat up silicon and slow down the system, it is a whole-cloth design change. The video display system, for instance, can generate a windowing GUI in a fraction of the RAM than a vga-based display system, and context switching or even resolution switching is also much simpler.
The basic concept is, rather than add modern features such as memory control, DMA, and bus mastering to an older (XT) based system, you build it from the ground up around these concepts. You get a performance boost for general functionality and can grow the system with less conflicts as well. In addition, it is flexible in nature, so rather than having a massve GPU that sits idle while you crunch spreadsheets, you have a massive DSP which can run your spreadsheet, or as a GPU for when you run that must-have game. I'm in the middle of re-designing this DSP to avoid several patents I ran across.
-
downix wrote:
Have my own host, thankfully, just need to get off my butt and get it all together
Cool beans.
downix wrote:
Rather than stcking to old standards, which eat up silicon and slow down the system, it is a whole-cloth design change. The video display system, for instance, can generate a windowing GUI in a fraction of the RAM than a vga-based display system, and context switching or even resolution switching is also much simpler.
I have no idea how you are achieving this, but well done anyway!
downix wrote:
The basic concept is, rather than add modern features such as memory control, DMA, and bus mastering to an older (XT) based system, you build it from the ground up around these concepts. You get a performance boost for general functionality and can grow the system with less conflicts as well. In addition, it is flexible in nature, so rather than having a massve GPU that sits idle while you crunch spreadsheets, you have a massive DSP which can run your spreadsheet, or as a GPU for when you run that must-have game. I'm in the middle of re-designing this DSP to avoid several patents I ran across.
Could you achieve a similar result by designing a hardware template that included a Tesla GPU as standard, running an OS built using CUDA? More info here:
http://www.nvidia.com/object/tesla_computing_solutions.html
-
HenryCase wrote:
downix wrote:
Rather than stcking to old standards, which eat up silicon and slow down the system, it is a whole-cloth design change. The video display system, for instance, can generate a windowing GUI in a fraction of the RAM than a vga-based display system, and context switching or even resolution switching is also much simpler.
I have no idea how you are achieving this, but well done anyway!
Well, for example, the classic RAMDAC model, where your GPU, once you've finished rendering the 3D scene, NOW has to put the frame together. I split this function away from the 3D section, slimming both cores while being more flexible. Also I focused a lot more attention on memory control and bus utilization, two areas that really lack in normal PC's but are focused on heavily in the heavy duty workstation market.
downix wrote:
The basic concept is, rather than add modern features such as memory control, DMA, and bus mastering to an older (XT) based system, you build it from the ground up around these concepts. You get a performance boost for general functionality and can grow the system with less conflicts as well. In addition, it is flexible in nature, so rather than having a massve GPU that sits idle while you crunch spreadsheets, you have a massive DSP which can run your spreadsheet, or as a GPU for when you run that must-have game. I'm in the middle of re-designing this DSP to avoid several patents I ran across.
Could you achieve a similar result by designing a hardware template that included a Tesla GPU as standard, running an OS built using CUDA? More info here:
http://www.nvidia.com/object/tesla_computing_solutions.html
Tesla is unable to be purchased without the display engine nor the memory control system, so would be wasting a lot of silicon for no gain. Conversely, I could utilize an Ageia Phys-X as a GPU much more efficiently, for this reason.
-
@Downix
That reminds me, a while ago I found this analysis of the Physx patents (http://www.blachford.info/computer/articles/PhysX1.html).
Worth noting that the first page is meaningless introductory drivel but the second and third pages contains the goods! Erm, enjoy!
Andy
-
HenryCase wrote:
kolla wrote:
kolla wrote:
What I meant is that OS4 doesnt need MiniMig :-)
Kolla, OS4 definitely needs more hardware capable of running it. Minimig could be one of those solutions.
OK?
Let's say you've made a MiniMig with PPC - now what?
After a PPC Minimig was created, we'd probably need to do some work to get OS4 working on it. After this, the OS4 h/w problem is solved.
So, let's ignore that there is no PCI or Zorro bus for gfx cards, no networking, no disk controller, and only glorious OCS for OS4 to display its 24bit icons on. Yay!
I'd say you have _a lot_ of work to do still. And after that again you have to make OS4 work on it. Why not just skip directly to "make OS4 work on it" and replace the hardware with something that exists already?
Grab an efika or something, it's a heck lot better starting point than the MiniMig!
More OS4 h/w means more users of OS4, which means more OS4 development. How hard is that to understand?
The hard bit is that I dont see the point in wasting time with OS4 for MiniMig. The point of MiniMig is to be able to run old amiga software. The point of OS4 was to move to PPC and get away from legacy hardware!
OS4 for classic is a nice treat for those of us with capable PPC cards in our amigas, but that's really all it is!
-
kolla wrote:
So, let's ignore that there is no PCI or Zorro bus for gfx cards, no networking, no disk controller, and only glorious OCS for OS4 to display its 24bit icons on. Yay!
I'm working on an AGA solution for Minimig too, and I'm not the only one. OS4 will run with AGA controlling the display stuff.
kolla wrote:
Grab an efika or something, it's a heck lot better starting point than the MiniMig!
Efika is closed source. Minimig is open source. Open source is the only way Amiga can continue to develop and thrive, as it doesn't rely on other companies that may fail us in the future.
kolla wrote:
The hard bit is that I dont see the point in wasting time with OS4 for MiniMig. The point of MiniMig is to be able to run old amiga software. The point of OS4 was to move to PPC and get away from legacy hardware!
Minimig = new Amiga hardware. As far as I can see, you see it as near enough the finished article. I see it as a development platform. Got to have some ambitions otherwise you never get anywhere, even if those ambitions require a lot of work.